Why Your Critical Path Changes When Production Changes

1
00:00:00,000 --> 00:00:02,360
Here's the problem most manufacturers don't talk about.

2
00:00:02,360 --> 00:00:04,720
Your planned completion date looks perfectly reasonable

3
00:00:04,720 --> 00:00:06,240
until production actually starts.

4
00:00:06,240 --> 00:00:09,200
Then one operation runs late, a material release slips,

5
00:00:09,200 --> 00:00:11,240
or a resource becomes unavailable,

6
00:00:11,240 --> 00:00:13,240
and every promised downstream moves with it.

7
00:00:13,240 --> 00:00:15,200
That's where critical path method becomes useful

8
00:00:15,200 --> 00:00:17,160
because it helps you find the dependency chain

9
00:00:17,160 --> 00:00:18,880
that controls delivery right now,

10
00:00:18,880 --> 00:00:20,400
not the chain that looked sensible

11
00:00:20,400 --> 00:00:22,520
when someone first released the order.

12
00:00:22,520 --> 00:00:24,200
Let's follow one order through a disruption

13
00:00:24,200 --> 00:00:26,160
that looks completely normal at first.

14
00:00:26,160 --> 00:00:28,720
A factory order isn't a project plan until it is.

15
00:00:28,720 --> 00:00:31,000
Most people first meet CPM in project management,

16
00:00:31,000 --> 00:00:33,160
but let's zoom out and think about a project

17
00:00:33,160 --> 00:00:35,200
with tasks where each takes time.

18
00:00:35,200 --> 00:00:36,680
Some tasks can start at once

19
00:00:36,680 --> 00:00:38,840
while others must wait for earlier work to finish.

20
00:00:38,840 --> 00:00:41,360
Put those dependencies together, estimate duration,

21
00:00:41,360 --> 00:00:44,000
and you can calculate the earliest possible finish date.

22
00:00:44,000 --> 00:00:46,240
The critical path is the chain of dependent tasks

23
00:00:46,240 --> 00:00:47,680
that controls that date.

24
00:00:47,680 --> 00:00:50,560
If a task on that path slips, the whole project slips.

25
00:00:50,560 --> 00:00:52,640
If a task outside that path slips,

26
00:00:52,640 --> 00:00:53,960
it may not affect the end date

27
00:00:53,960 --> 00:00:56,400
because it has spare time, which project managers call

28
00:00:56,400 --> 00:00:57,640
float or slack.

29
00:00:57,640 --> 00:00:59,360
We'll come back to that because it matters a lot

30
00:00:59,360 --> 00:01:00,200
on the shop floor.

31
00:01:00,200 --> 00:01:02,040
At first, this sounds like a project problem,

32
00:01:02,040 --> 00:01:03,240
not a production problem

33
00:01:03,240 --> 00:01:05,560
because a factory runs the same products again and again

34
00:01:05,560 --> 00:01:08,880
with routings, standard cycle times, work centers,

35
00:01:08,880 --> 00:01:11,400
shift patterns, and planning rules.

36
00:01:11,400 --> 00:01:12,600
A project might build a plant,

37
00:01:12,600 --> 00:01:15,000
install a new line or design a one off machine.

38
00:01:15,000 --> 00:01:16,440
Production repeats known work.

39
00:01:16,440 --> 00:01:18,320
But here's the thing, a released production order

40
00:01:18,320 --> 00:01:20,520
has a lot in common with a project plan.

41
00:01:20,520 --> 00:01:23,160
It has operations that must happen in a certain order

42
00:01:23,160 --> 00:01:25,920
and it needs material before some can start.

43
00:01:25,920 --> 00:01:28,680
It may need a particular machine, a fixture, a cutting tool,

44
00:01:28,680 --> 00:01:30,880
a qualified operator, an inspection result,

45
00:01:30,880 --> 00:01:31,960
and a quality release.

46
00:01:31,960 --> 00:01:34,320
It ends only when the product reaches the state

47
00:01:34,320 --> 00:01:36,000
the customer actually expects.

48
00:01:36,000 --> 00:01:37,400
That is a dependency network.

49
00:01:37,400 --> 00:01:39,920
Think about a work order for a machine assembly.

50
00:01:39,920 --> 00:01:41,480
Engineering defines the product,

51
00:01:41,480 --> 00:01:43,480
the routing defines the intended process,

52
00:01:43,480 --> 00:01:45,240
planning releases the order,

53
00:01:45,240 --> 00:01:47,360
a machine processes the first operation,

54
00:01:47,360 --> 00:01:49,440
material moves to the next work center,

55
00:01:49,440 --> 00:01:50,680
someone inspects the part

56
00:01:50,680 --> 00:01:52,280
and another operation can begin only

57
00:01:52,280 --> 00:01:54,120
when the inspection result clears it.

58
00:01:54,120 --> 00:01:55,680
Each of those links carries a condition.

59
00:01:55,680 --> 00:01:57,320
The next operation doesn't start simply

60
00:01:57,320 --> 00:01:59,960
because the previous operation has a planned end time.

61
00:01:59,960 --> 00:02:01,920
It starts when the prior work finishes,

62
00:02:01,920 --> 00:02:05,400
the quantity is available, the material status is acceptable,

63
00:02:05,400 --> 00:02:08,400
the resource is ready, and the schedule has a slot for it.

64
00:02:08,400 --> 00:02:09,560
That's a very different thing.

65
00:02:09,560 --> 00:02:11,560
In project planning, people often draw a network

66
00:02:11,560 --> 00:02:14,720
around one project with its own tasks, path, and team.

67
00:02:14,720 --> 00:02:16,080
Production doesn't give you that luxury.

68
00:02:16,080 --> 00:02:17,920
Your order shares machines with other orders

69
00:02:17,920 --> 00:02:19,840
and operators with other work centers.

70
00:02:19,840 --> 00:02:21,720
It competes for tools, test equipment,

71
00:02:21,720 --> 00:02:24,720
forklifts, heat treatment capacity, inspection staff,

72
00:02:24,720 --> 00:02:28,080
and sometimes even physical space to stage work in progress.

73
00:02:28,080 --> 00:02:30,640
One local decision can move several delivery dates.

74
00:02:30,640 --> 00:02:33,000
So a production order isn't an isolated project plan.

75
00:02:33,000 --> 00:02:36,160
It's one moving part inside a wider network of orders

76
00:02:36,160 --> 00:02:38,240
and constraints and that wider network changes

77
00:02:38,240 --> 00:02:39,440
throughout the day.

78
00:02:39,440 --> 00:02:40,840
A planner can release 10 orders

79
00:02:40,840 --> 00:02:42,800
that each look feasible on their own,

80
00:02:42,800 --> 00:02:45,640
then discover that all 10 need the same constrained resource

81
00:02:45,640 --> 00:02:46,840
during the same shift.

82
00:02:46,840 --> 00:02:49,280
The routing didn't change, the dependency network did.

83
00:02:49,280 --> 00:02:51,000
This is why a fixed production plan

84
00:02:51,000 --> 00:02:52,720
often starts to lose accuracy,

85
00:02:52,720 --> 00:02:54,240
the moment it meets the factory.

86
00:02:54,240 --> 00:02:56,480
The plan carries assumptions about duration,

87
00:02:56,480 --> 00:02:58,400
capacity, availability, and sequence,

88
00:02:58,400 --> 00:03:00,240
some of which hold and some don't.

89
00:03:00,240 --> 00:03:02,360
The machine may run slower than planned,

90
00:03:02,360 --> 00:03:04,120
a tool may reach its life limit,

91
00:03:04,120 --> 00:03:06,880
a skilled operator may move to another urgent job,

92
00:03:06,880 --> 00:03:09,280
or a material lot may wait for inspection.

93
00:03:09,280 --> 00:03:11,040
None of this means planning failed.

94
00:03:11,040 --> 00:03:12,960
It means manufacturing deals with physical work,

95
00:03:12,960 --> 00:03:14,720
shared resources, and changing conditions.

96
00:03:14,720 --> 00:03:16,680
You can't restart a production day

97
00:03:16,680 --> 00:03:18,240
the way you restart a cloud service.

98
00:03:18,240 --> 00:03:19,840
Parts are already in process.

99
00:03:19,840 --> 00:03:21,560
People have already started setups

100
00:03:21,560 --> 00:03:23,880
and material may already sit in a furnace

101
00:03:23,880 --> 00:03:25,840
or on a pallet between departments.

102
00:03:25,840 --> 00:03:27,720
That's why project management thinking helps.

103
00:03:27,720 --> 00:03:28,680
As long as we adapt it,

104
00:03:28,680 --> 00:03:31,280
we don't need to pretend a factory is one giant project

105
00:03:31,280 --> 00:03:34,200
that would turn into a planning exercise nobody can maintain.

106
00:03:34,200 --> 00:03:36,600
But we do need the discipline behind CPM,

107
00:03:36,600 --> 00:03:39,200
make dependencies explicit, track the conditions

108
00:03:39,200 --> 00:03:41,240
that change them, and calculate which chain

109
00:03:41,240 --> 00:03:43,040
now controls the promised date.

110
00:03:43,040 --> 00:03:44,560
The plan gives you an initial answer

111
00:03:44,560 --> 00:03:47,320
but life production keeps asking whether that answer still holds.

112
00:03:47,320 --> 00:03:49,840
Let's make this concrete with one order, one normal route,

113
00:03:49,840 --> 00:03:52,840
and one machine problem that exposes the whole chain.

114
00:03:52,840 --> 00:03:55,000
One-late operation and the delivery date.

115
00:03:55,000 --> 00:03:57,560
Let's walk through a production order for a machine assembly

116
00:03:57,560 --> 00:03:59,800
that really needs to ship by end of week.

117
00:03:59,800 --> 00:04:01,760
The route looks standard, machining first,

118
00:04:01,760 --> 00:04:03,600
then heat treatment, surface finishing,

119
00:04:03,600 --> 00:04:06,280
final inspection, packing, and dispatch.

120
00:04:06,280 --> 00:04:09,200
Planning placed each operation in a sensible time slot.

121
00:04:09,200 --> 00:04:12,280
The machining center kicks off early enough to feed heat treatment,

122
00:04:12,280 --> 00:04:13,960
which releases the batch to finishing,

123
00:04:13,960 --> 00:04:15,920
finishing hands parts off to inspection,

124
00:04:15,920 --> 00:04:18,040
and inspection clears them for packing.

125
00:04:18,040 --> 00:04:20,680
On paper, every handoff has time built in.

126
00:04:20,680 --> 00:04:23,920
Then, during machining, the constrained machining center stops.

127
00:04:23,920 --> 00:04:25,840
It doesn't have to be a dramatic failure.

128
00:04:25,840 --> 00:04:28,480
Maybe a spindle alarm triggers and needs maintenance.

129
00:04:28,480 --> 00:04:30,000
The repair might take a few hours.

130
00:04:30,000 --> 00:04:31,240
The work order has started,

131
00:04:31,240 --> 00:04:33,680
but only part of the required quantity is done.

132
00:04:33,680 --> 00:04:35,640
And the remaining parts can't move to heat treatment

133
00:04:35,640 --> 00:04:37,520
because the batch needs the full quantity,

134
00:04:37,520 --> 00:04:40,840
or the process rules demand matched pieces from the same lot.

135
00:04:40,840 --> 00:04:42,640
The planned handoff just evaporated.

136
00:04:42,640 --> 00:04:44,840
A common reaction is to look at the next apartment.

137
00:04:44,840 --> 00:04:47,600
Heat treatment might show open capacity later that day.

138
00:04:47,600 --> 00:04:49,000
Finishing may have no queue.

139
00:04:49,000 --> 00:04:51,680
Inspection could even have people waiting for work.

140
00:04:51,680 --> 00:04:53,840
That creates a false sense of safety.

141
00:04:53,840 --> 00:04:56,640
Each department looks available when you view it in isolation,

142
00:04:56,640 --> 00:04:58,160
but open capacity downstream

143
00:04:58,160 --> 00:05:00,520
doesn't pull material through an unfinished operation.

144
00:05:00,520 --> 00:05:02,200
Heat treatment can't process parts

145
00:05:02,200 --> 00:05:03,480
that haven't been machined.

146
00:05:03,480 --> 00:05:06,120
Finishing can't start until heat treatment releases the batch.

147
00:05:06,120 --> 00:05:07,560
Inspection has nothing to inspect.

148
00:05:07,560 --> 00:05:09,000
Packing can't prepare a product

149
00:05:09,000 --> 00:05:11,880
that's still sitting partly completed at the first work center.

150
00:05:11,880 --> 00:05:14,040
The outage creates a delay at machining,

151
00:05:14,040 --> 00:05:16,920
and then that delay travels through every dependent handoff.

152
00:05:16,920 --> 00:05:19,680
Now add the detail planners deal with every single day.

153
00:05:19,680 --> 00:05:21,520
The furnace may only load at set times.

154
00:05:21,520 --> 00:05:23,480
If machining misses the next load window,

155
00:05:23,480 --> 00:05:25,720
the order doesn't just lose the repair time.

156
00:05:25,720 --> 00:05:27,960
It waits until the next usable furnace slot,

157
00:05:27,960 --> 00:05:30,560
and that later release may miss the finishing shift.

158
00:05:30,560 --> 00:05:31,840
A few hours at the machining center

159
00:05:31,840 --> 00:05:33,880
can turn into a full day of delivery risk.

160
00:05:33,880 --> 00:05:35,680
That's how a due date becomes exposed.

161
00:05:35,680 --> 00:05:38,680
The customer doesn't care which operation caused the delay.

162
00:05:38,680 --> 00:05:40,800
They care whether the order ships went promised.

163
00:05:40,800 --> 00:05:43,080
Yet internally, different people see different parts

164
00:05:43,080 --> 00:05:44,160
of the same issue.

165
00:05:44,160 --> 00:05:46,240
Maintenance sees an equipment problem.

166
00:05:46,240 --> 00:05:47,840
The machining sees an interrupted job.

167
00:05:47,840 --> 00:05:49,640
The heat treatment team sees a missing batch.

168
00:05:49,640 --> 00:05:52,040
Customer service sees an order that still appears on track

169
00:05:52,040 --> 00:05:53,800
until someone updates the date.

170
00:05:53,800 --> 00:05:56,560
Each view is correct, but none tells the whole story.

171
00:05:56,560 --> 00:05:58,080
Let's make the timing more precise

172
00:05:58,080 --> 00:06:00,000
without turning this into a math lesson.

173
00:06:00,000 --> 00:06:02,120
Suppose machining should finish early Tuesday.

174
00:06:02,120 --> 00:06:04,240
The next furnace load accepts work at noon.

175
00:06:04,240 --> 00:06:05,960
Finishing runs Tuesday afternoon.

176
00:06:05,960 --> 00:06:07,760
Inspection happens Wednesday morning

177
00:06:07,760 --> 00:06:10,440
and packing leaves enough time for the carrier collection.

178
00:06:10,440 --> 00:06:11,720
When the machining center stops

179
00:06:11,720 --> 00:06:13,960
and work finishes after the noon furnace load,

180
00:06:13,960 --> 00:06:15,560
the order joins the next load instead.

181
00:06:15,560 --> 00:06:17,200
Maybe that load runs Wednesday.

182
00:06:17,200 --> 00:06:19,160
The promised shipment now depends on whether finishing

183
00:06:19,160 --> 00:06:21,280
can take the work immediately after release.

184
00:06:21,280 --> 00:06:22,760
Whether inspection can clear it

185
00:06:22,760 --> 00:06:25,880
and whether the dispatch cutoff still allows it to leave.

186
00:06:25,880 --> 00:06:27,640
A single-late operation changed the chain

187
00:06:27,640 --> 00:06:28,760
that controls delivery.

188
00:06:28,760 --> 00:06:31,600
It may also change what production should focus on.

189
00:06:31,600 --> 00:06:33,880
Before the outage, the machining operation

190
00:06:33,880 --> 00:06:36,640
carried the most pressure because it fed everything else.

191
00:06:36,640 --> 00:06:38,240
After machining catches up,

192
00:06:38,240 --> 00:06:41,080
the furnace load may become the limiting point.

193
00:06:41,080 --> 00:06:43,280
Once the furnace releases, final inspection

194
00:06:43,280 --> 00:06:44,280
may carry the pressure

195
00:06:44,280 --> 00:06:46,000
because there is no time left before packing

196
00:06:46,000 --> 00:06:47,200
and carrier collection.

197
00:06:47,200 --> 00:06:49,520
The operational constraint moves.

198
00:06:49,520 --> 00:06:51,520
That distinction matters because people often ask

199
00:06:51,520 --> 00:06:53,160
which operation is late.

200
00:06:53,160 --> 00:06:55,080
Fair question, but it isn't enough.

201
00:06:55,080 --> 00:06:57,480
The better question is, which remaining dependency

202
00:06:57,480 --> 00:06:59,400
now controls whether this orderships?

203
00:06:59,400 --> 00:07:01,520
Those questions can produce different answers.

204
00:07:01,520 --> 00:07:03,120
An operation can report late

205
00:07:03,120 --> 00:07:04,760
and still not threaten the due date

206
00:07:04,760 --> 00:07:06,440
if there's time later in the route.

207
00:07:06,440 --> 00:07:08,920
Another operation can finish exactly on its planned date

208
00:07:08,920 --> 00:07:09,960
and still control delivery

209
00:07:09,960 --> 00:07:11,880
because every remaining step depends on it

210
00:07:11,880 --> 00:07:13,000
with no room left.

211
00:07:13,000 --> 00:07:14,360
Planners know this instinctively.

212
00:07:14,360 --> 00:07:16,440
They chase the work that can still change the outcome.

213
00:07:16,440 --> 00:07:18,600
Not just the work that looks worse in a status report.

214
00:07:18,600 --> 00:07:20,480
And there's also a human side to this.

215
00:07:20,480 --> 00:07:22,840
When everyone sees a late machine operation,

216
00:07:22,840 --> 00:07:25,560
the natural response is to expedite it, push the repair,

217
00:07:25,560 --> 00:07:27,480
move the job forward, ask for overtime.

218
00:07:27,480 --> 00:07:30,320
Sometimes that's the right choice, sometimes it isn't.

219
00:07:30,320 --> 00:07:32,440
If the order has already missed the only furnace slot

220
00:07:32,440 --> 00:07:34,440
that protects the ship date, extra speed

221
00:07:34,440 --> 00:07:35,960
at machining may not recover anything

222
00:07:35,960 --> 00:07:37,840
unless the furnace plan changes, too.

223
00:07:37,840 --> 00:07:39,120
You need to see the linked chain

224
00:07:39,120 --> 00:07:41,920
before you spend effort on the wrong recovery action.

225
00:07:41,920 --> 00:07:44,680
Otherwise, the plant runs faster around a constraint

226
00:07:44,680 --> 00:07:46,160
that has already moved elsewhere.

227
00:07:46,160 --> 00:07:48,080
Manufacturing has plenty of those moments,

228
00:07:48,080 --> 00:07:49,600
usually followed by an urgent meeting

229
00:07:49,600 --> 00:07:51,920
and an Excel file with 17 tabs.

230
00:07:51,920 --> 00:07:54,040
So the visible delay is only the starting point.

231
00:07:54,040 --> 00:07:56,280
The real work is tracing from that delay

232
00:07:56,280 --> 00:07:59,040
through the operations, handoffs, and release conditions

233
00:07:59,040 --> 00:08:00,560
that lead to the customer promise.

234
00:08:00,560 --> 00:08:03,040
Only then can you tell whether the order is recoverable

235
00:08:03,040 --> 00:08:04,160
where intervention helps

236
00:08:04,160 --> 00:08:06,640
and which dependency controls completion now.

237
00:08:06,640 --> 00:08:09,200
What critical path actually means on the shop floor?

238
00:08:09,200 --> 00:08:11,120
Critical path sounds like a fixed label.

239
00:08:11,120 --> 00:08:12,920
In a factory, it rarely is.

240
00:08:12,920 --> 00:08:14,480
Here's what it actually means.

241
00:08:14,480 --> 00:08:16,080
The chain of remaining dependencies

242
00:08:16,080 --> 00:08:18,560
that sets the earliest possible completion date

243
00:08:18,560 --> 00:08:21,480
for a specific order, given what the factory can actually do

244
00:08:21,480 --> 00:08:24,200
right now, not what the original plan assumed

245
00:08:24,200 --> 00:08:26,240
when it released the order.

246
00:08:26,240 --> 00:08:28,760
That chain might begin with work already in progress.

247
00:08:28,760 --> 00:08:31,200
It might begin with an operation that hasn't started.

248
00:08:31,200 --> 00:08:32,720
It can even begin with a release condition

249
00:08:32,720 --> 00:08:34,520
that nobody sees as an operation,

250
00:08:34,520 --> 00:08:37,240
but which still prevents the next physical step from starting.

251
00:08:37,240 --> 00:08:39,680
Think about an order with three remaining branches of work.

252
00:08:39,680 --> 00:08:42,280
The main component needs machining and heat treatment,

253
00:08:42,280 --> 00:08:43,920
a purchased subassembly needs to arrive

254
00:08:43,920 --> 00:08:45,600
and pass incoming inspection.

255
00:08:45,600 --> 00:08:48,360
Meanwhile, a document pack needs approval before shipment.

256
00:08:48,360 --> 00:08:50,800
All three branches lead to the same customer commitment,

257
00:08:50,800 --> 00:08:53,400
whichever branch finishes last controls the delivery date.

258
00:08:53,400 --> 00:08:56,040
That is the critical path, not the most expensive operation,

259
00:08:56,040 --> 00:08:57,880
not the operation with the loudest escalation,

260
00:08:57,880 --> 00:09:00,680
not automatically the first operation in the routing.

261
00:09:00,680 --> 00:09:02,520
It's the chain that leaves no usable time

262
00:09:02,520 --> 00:09:05,000
between the work still required and the promised finish.

263
00:09:05,000 --> 00:09:07,960
Project planners describe that spare time as float or slack.

264
00:09:07,960 --> 00:09:09,840
If an operation can slip by a shift

265
00:09:09,840 --> 00:09:12,640
and the order can still ship on time, it has float.

266
00:09:12,640 --> 00:09:15,960
If moving it by one hour moves the final completion by one hour,

267
00:09:15,960 --> 00:09:17,200
it has zero float.

268
00:09:17,200 --> 00:09:19,680
That operation sits on the critical path at that moment.

269
00:09:19,680 --> 00:09:21,400
Low float deserves attention too.

270
00:09:21,400 --> 00:09:24,680
A job with only a small amount of slack may not control the date yet,

271
00:09:24,680 --> 00:09:27,320
but a routine disruption can put it there very quickly.

272
00:09:27,320 --> 00:09:30,520
A late operator handover, a slightly longer change over,

273
00:09:30,520 --> 00:09:34,000
or a short wait for inspection can consume that remaining room.

274
00:09:34,000 --> 00:09:37,120
The path changes before anyone updates the order status to red.

275
00:09:37,120 --> 00:09:40,240
This is where shop floor planning differs from a neat project network.

276
00:09:40,240 --> 00:09:43,120
In a project task, durations often sit still long enough

277
00:09:43,120 --> 00:09:44,920
for a planned path to remain useful.

278
00:09:44,920 --> 00:09:48,400
On the shop floor, durations can move, resource availability can move,

279
00:09:48,400 --> 00:09:50,800
and the sequence can move sometimes within the same shift.

280
00:09:50,800 --> 00:09:52,840
The calculation needs to deal with the current state,

281
00:09:52,840 --> 00:09:55,320
not just planned dates copied forward from Monday morning.

282
00:09:55,320 --> 00:09:56,720
Some work happens in sequence.

283
00:09:56,720 --> 00:09:58,600
You can't inspect a part before the process

284
00:09:58,600 --> 00:10:00,800
that creates the feature you need to inspect.

285
00:10:00,800 --> 00:10:02,360
Other work can happen at the same time,

286
00:10:02,360 --> 00:10:04,640
while one team completes a mechanical assembly,

287
00:10:04,640 --> 00:10:08,520
another team may prepare packaging material or assemble a separate module.

288
00:10:08,520 --> 00:10:10,680
Parallel work can shorten elapsed lead time,

289
00:10:10,680 --> 00:10:14,200
but parallel work only helps if the branch is joined before the final deadline.

290
00:10:14,200 --> 00:10:16,240
If one branch has room and the other has none,

291
00:10:16,240 --> 00:10:18,760
speeding up the branch with room doesn't change delivery.

292
00:10:18,760 --> 00:10:20,560
This sounds obvious when you say it out loud,

293
00:10:20,560 --> 00:10:22,360
yet plants spend a lot of energy,

294
00:10:22,360 --> 00:10:25,680
improving visible activity that doesn't alter the customer date.

295
00:10:25,680 --> 00:10:29,920
A critical path calculation puts a different question in front of the planner.

296
00:10:29,920 --> 00:10:32,200
Where does extra time still change the outcome?

297
00:10:32,200 --> 00:10:34,360
That question changes the daily discussion.

298
00:10:34,360 --> 00:10:36,040
Instead of asking every department to hurry,

299
00:10:36,040 --> 00:10:39,120
you can ask which operation release or handoff needs protection

300
00:10:39,120 --> 00:10:40,520
because it has no slack left.

301
00:10:40,520 --> 00:10:43,040
You can also see which work can wait without causing harm,

302
00:10:43,040 --> 00:10:45,320
even if it feels uncomfortable to leave it alone.

303
00:10:45,320 --> 00:10:47,440
Urgent status doesn't answer that question.

304
00:10:47,440 --> 00:10:50,160
An order may carry a high priority because of its customer,

305
00:10:50,160 --> 00:10:52,560
margin, contract terms or internal attention.

306
00:10:52,560 --> 00:10:54,760
Those are business reasons to care about the order.

307
00:10:54,760 --> 00:10:57,920
They don't prove that every operation on that order controls its finish date.

308
00:10:57,920 --> 00:11:01,800
Imagine an urgent order with final packing scheduled well before dispatch.

309
00:11:01,800 --> 00:11:03,520
Packing may carry two days of float.

310
00:11:03,520 --> 00:11:08,240
At the same time, a lower priority order may need a short test on the same constrained resource.

311
00:11:08,240 --> 00:11:11,600
And it has no float at all because a follow-on shipment depends on it.

312
00:11:11,600 --> 00:11:15,480
If you pull the test resource away solely because the first order has an urgent flag,

313
00:11:15,480 --> 00:11:18,480
you can create a new late order without improving the urgent one.

314
00:11:18,480 --> 00:11:19,560
Priority matters.

315
00:11:19,560 --> 00:11:22,600
Dependency timing matters too, planning needs both.

316
00:11:22,600 --> 00:11:26,320
That's why a live critical path shouldn't become another color code in a report.

317
00:11:26,320 --> 00:11:28,800
It should explain the current logic behind the date.

318
00:11:28,800 --> 00:11:30,880
This operation needs to finish by this time

319
00:11:30,880 --> 00:11:36,000
because the next operation needs a resource at that time and the next handoff has no remaining room.

320
00:11:36,000 --> 00:11:39,480
When the logic is visible, production, planning, quality and customer service

321
00:11:39,480 --> 00:11:43,280
can discuss the same problem instead of defending separate status views.

322
00:11:43,280 --> 00:11:45,560
There's one common shortcut that causes confusion.

323
00:11:45,560 --> 00:11:49,280
People look at the routing and assume its sequence already tells them the critical path.

324
00:11:49,280 --> 00:11:51,960
It tells you part of the story but only part.

325
00:11:51,960 --> 00:11:54,080
Why a routing is not a critical path?

326
00:11:54,080 --> 00:11:56,520
A routing tells you how a product should move through production.

327
00:11:56,520 --> 00:12:00,640
It might say a part goes from turning to milling, then heat treatment, then inspection,

328
00:12:00,640 --> 00:12:05,040
and it includes plan setup time, run time, the work center and sometimes an alternate work center.

329
00:12:05,040 --> 00:12:09,160
That process logic matters but it doesn't calculate the delivery risk on its own.

330
00:12:09,160 --> 00:12:11,400
Think of the routing as the recipe for the part.

331
00:12:11,400 --> 00:12:13,800
A critical path answers a different question.

332
00:12:13,800 --> 00:12:17,240
It takes the work already released, the resources available right now,

333
00:12:17,240 --> 00:12:22,000
and the plan's timing rules then figures out which chain actually controls when this order finishes.

334
00:12:22,000 --> 00:12:24,640
Those two things often overlap, they don't always match.

335
00:12:24,640 --> 00:12:27,400
Take a routing where milling comes before heat treatment.

336
00:12:27,400 --> 00:12:29,480
The routing tells you the physical order of work.

337
00:12:29,480 --> 00:12:33,240
But if milling finishes early and the next qualified furnace run is two days away,

338
00:12:33,240 --> 00:12:35,040
milling isn't controlling the date anymore.

339
00:12:35,040 --> 00:12:38,640
The furnace queue is, or look at an operation with two approved machines.

340
00:12:38,640 --> 00:12:41,920
On the routing, it's one step with one planned duration.

341
00:12:41,920 --> 00:12:44,760
In live production, those two machines don't give you the same answer.

342
00:12:44,760 --> 00:12:46,320
One might be tied up on a long job.

343
00:12:46,320 --> 00:12:50,120
The other might be technically capable, but it needs a fixture change, a proven program,

344
00:12:50,120 --> 00:12:51,960
and an operator who knows that product family.

345
00:12:51,960 --> 00:12:54,800
So the operation doesn't simply have two slots.

346
00:12:54,800 --> 00:12:58,520
It has a set of conditions that decide whether either slot is actually usable in time.

347
00:12:58,520 --> 00:13:03,480
The routing records, the allowed choices, the critical path depends on which choice remains feasible right now.

348
00:13:03,480 --> 00:13:05,200
Parallel work adds another layer.

349
00:13:05,200 --> 00:13:09,760
Some products can move in transfer batches instead of waiting for the whole order quantity to finish.

350
00:13:09,760 --> 00:13:15,200
If the first batch leaves machining and feeds the next process, downstream work starts while machining continues.

351
00:13:15,200 --> 00:13:16,840
That can shorten the order's total time.

352
00:13:16,840 --> 00:13:21,440
But only if the process rules allow the overlap, the next work center can accept the partial batch

353
00:13:21,440 --> 00:13:23,320
and the handoff doesn't create a new delay.

354
00:13:23,320 --> 00:13:29,040
If heat treatment needs a full matched batch, then partial machining completion changes nothing for that operation.

355
00:13:29,040 --> 00:13:33,240
The routing might list the same sequence in both cases, but the timing logic differs completely.

356
00:13:33,240 --> 00:13:35,560
Waiting time causes the same kind of confusion.

357
00:13:35,560 --> 00:13:41,080
A routing often contains standard labor and machine time because those figures support planning and costing.

358
00:13:41,080 --> 00:13:48,320
Yet the order might spend more elapsed time waiting between operations than it actually spends under a tool in a furnace or at an inspection bench.

359
00:13:48,320 --> 00:13:52,200
The route tells you where the part should go, it doesn't tell you when the part can actually go there.

360
00:13:52,200 --> 00:13:54,680
That gap is where planners spend most of their day.

361
00:13:54,680 --> 00:13:58,000
A short operation can control delivery more than a long one.

362
00:13:58,000 --> 00:14:05,640
Picture 10 hours of machining with a full shift of float behind it, followed by a 20 minute inspection that has to happen before a fixed shipping cut off.

363
00:14:05,640 --> 00:14:08,880
The machining looks bigger, but inspection carries the timing pressure.

364
00:14:08,880 --> 00:14:12,360
If the inspector isn't available in that window, the order misses dispatched.

365
00:14:12,360 --> 00:14:14,560
That doesn't mean the machining team can relax.

366
00:14:14,560 --> 00:14:19,080
It means the plan needs to protect the inspection slot just as carefully as the main production work.

367
00:14:19,080 --> 00:14:25,320
Without a time-based dependency view, the plant might focus on the longest operation just because it looks risky.

368
00:14:25,320 --> 00:14:27,360
Length and criticality aren't the same thing.

369
00:14:27,360 --> 00:14:32,520
There are also routing rules that look stable until a planner starts asking practical questions.

370
00:14:32,520 --> 00:14:37,720
Like whether this operation can overlap with the next one, or if the product needs to cool for a set period,

371
00:14:37,720 --> 00:14:40,440
or whether a batch can split across two machines.

372
00:14:40,440 --> 00:14:42,200
Those rules shape the possible schedule.

373
00:14:42,200 --> 00:14:46,120
A critical path calculation takes those rules and then applies the current facts.

374
00:14:46,120 --> 00:14:48,440
Actual release time, remaining work.

375
00:14:48,440 --> 00:14:51,880
Available capacity, handoffs, and any known restrictions.

376
00:14:51,880 --> 00:14:56,360
The result isn't a pretty or rooting, it's a current explanation of what really controls completion.

377
00:14:56,360 --> 00:14:59,560
For planners, this means you need two connected views of the same order.

378
00:14:59,560 --> 00:15:02,840
You need the routing because it captures product and process intent.

379
00:15:02,840 --> 00:15:08,040
It tells the system what work belongs in the order and which paths comply with engineering and quality rules.

380
00:15:08,040 --> 00:15:13,560
Then you need live constraint logic to test whether that intended route can actually finish under current factory conditions.

381
00:15:13,560 --> 00:15:15,400
One without the other leaves a blind spot.

382
00:15:15,400 --> 00:15:20,600
A routing without live constraints might look feasible while the order actually waits behind another job,

383
00:15:20,600 --> 00:15:24,040
misses a handoff, or loses access to an alternate resource.

384
00:15:24,040 --> 00:15:28,360
And a schedule without rooting logic can produce an attractive answer that ignores

385
00:15:28,360 --> 00:15:30,600
how the product must actually be built.

386
00:15:30,600 --> 00:15:34,680
Neither result helps the person standing at the machine with a work order in their hand.

387
00:15:34,680 --> 00:15:38,440
This is also why copying routing dates into a spreadsheet doesn't fix the issue.

388
00:15:38,440 --> 00:15:41,080
The spreadsheet might show every planned operation in sequence,

389
00:15:41,080 --> 00:15:44,440
but unless it includes the changing constraints in the relationships between them,

390
00:15:44,440 --> 00:15:46,360
it's just a list of intended dates.

391
00:15:46,360 --> 00:15:50,600
People repair the missing logic manually through calls, whiteboards, and experience.

392
00:15:50,600 --> 00:15:54,840
Excel keeps surviving because it captures the conversation that systems often miss.

393
00:15:54,840 --> 00:15:58,040
The job isn't to replace that conversation with a more colorful report.

394
00:15:58,040 --> 00:16:00,680
The job is to give planners a model that carries the process rules

395
00:16:00,680 --> 00:16:03,320
and updates the timing logic when production changes.

396
00:16:03,320 --> 00:16:05,720
Then they can see whether a delay matters, where it travels,

397
00:16:05,720 --> 00:16:08,200
and which intervention still protects the customer date.

398
00:16:08,200 --> 00:16:11,720
The first moving dependency often appears before any machine starts.

399
00:16:11,720 --> 00:16:12,680
Material readiness.

400
00:16:13,640 --> 00:16:16,680
Material changes the path before the first machine runs.

401
00:16:16,680 --> 00:16:21,240
Material readiness can put an order on the critical path before anyone touches a machine.

402
00:16:21,240 --> 00:16:24,280
A planner sees a confirmed purchase date in the ERP

403
00:16:24,280 --> 00:16:26,920
and assumes the material will feed production on Monday.

404
00:16:26,920 --> 00:16:29,480
But that date usually describes an expectation.

405
00:16:29,480 --> 00:16:33,400
Production needs material that has actually arrived past the required checks

406
00:16:33,400 --> 00:16:35,320
and can legally enter the process.

407
00:16:35,320 --> 00:16:38,040
Picture a purchase casting needed for the first operation.

408
00:16:38,040 --> 00:16:40,120
The supplier confirms delivery for Friday,

409
00:16:40,120 --> 00:16:42,520
so planning schedules, machining for Monday morning,

410
00:16:42,520 --> 00:16:44,280
and the rest of the route follows from there.

411
00:16:44,280 --> 00:16:45,080
It looks clean.

412
00:16:45,080 --> 00:16:47,880
Then Friday arrives and the casting shows up,

413
00:16:47,880 --> 00:16:50,120
but incoming inspection hasn't released it.

414
00:16:50,120 --> 00:16:53,080
Maybe the receiving team still needs the certificate of conformity

415
00:16:53,080 --> 00:16:54,920
or the lot waits for dimensional checks,

416
00:16:54,920 --> 00:16:57,240
or a quality rule requires sample approval

417
00:16:57,240 --> 00:16:59,720
before the material can move into usable stock.

418
00:16:59,720 --> 00:17:01,640
The casting is physically in the plant.

419
00:17:01,640 --> 00:17:04,360
For the work order, it might as well still be on the truck.

420
00:17:04,360 --> 00:17:08,280
That's a common gap between an ERP material date and a production-ready material date.

421
00:17:08,280 --> 00:17:12,200
ERP holds the demand, purchase order, supply or promise, and expected receipt.

422
00:17:12,200 --> 00:17:15,320
Those records matter, but the first operation needs a different answer.

423
00:17:15,320 --> 00:17:18,680
Whether the operator can issue the right quantity from the right lot

424
00:17:18,680 --> 00:17:21,880
under the right quality status at the time the schedule needs it,

425
00:17:21,880 --> 00:17:23,640
and that answer can change quickly.

426
00:17:23,640 --> 00:17:25,400
Raw stock might arrive short,

427
00:17:25,400 --> 00:17:28,120
or the supplier splits delivery across several receipts.

428
00:17:28,120 --> 00:17:32,440
The material exists, but only enough quantity has passed inspection to start part of the order.

429
00:17:32,440 --> 00:17:34,040
At that point, planning has choices,

430
00:17:34,040 --> 00:17:36,120
and each choice changes the path.

431
00:17:36,120 --> 00:17:39,640
You could wait until the full quantity clears and run the order as one lot.

432
00:17:39,640 --> 00:17:43,720
You could release the available quantity and split the order into smaller production batches.

433
00:17:43,720 --> 00:17:45,800
If the process and traceability rules allow it.

434
00:17:45,800 --> 00:17:50,600
Or you might protect another order that needs the same material and delay this one.

435
00:17:50,600 --> 00:17:52,920
None of those choices belongs only to purchasing.

436
00:17:52,920 --> 00:17:56,120
Each choice changes the timing of operations, transfer points,

437
00:17:56,120 --> 00:17:58,680
inspection work, and the likely shipment date.

438
00:17:58,680 --> 00:18:01,480
Splitting an order can pull the first finished units forward,

439
00:18:01,480 --> 00:18:04,120
but it can also create extra setups, more handling,

440
00:18:04,120 --> 00:18:05,640
separate quality records,

441
00:18:05,640 --> 00:18:08,040
or a mismatch with a downstream batch process.

442
00:18:08,040 --> 00:18:11,320
The material problem doesn't disappear when material becomes available.

443
00:18:11,320 --> 00:18:12,200
It changes shape.

444
00:18:12,200 --> 00:18:16,600
Substitute material creates another version of the same issue.

445
00:18:16,600 --> 00:18:18,760
Someone might suggest an approved alternative grade,

446
00:18:18,760 --> 00:18:20,040
a different supplier lot,

447
00:18:20,040 --> 00:18:23,560
or stock meant for another product family that sounds like a simple stock decision.

448
00:18:23,560 --> 00:18:26,600
In practical terms, it can change the required process parameters,

449
00:18:26,600 --> 00:18:29,400
tooling, inspection plan, or documents needed at shipment.

450
00:18:29,400 --> 00:18:31,240
An alternate material can save the date,

451
00:18:31,240 --> 00:18:33,160
but it can also create new dependencies,

452
00:18:33,160 --> 00:18:34,680
the original route never needed.

453
00:18:34,680 --> 00:18:36,920
Engineering may need to approve the substitution.

454
00:18:36,920 --> 00:18:39,800
Quality may need a deviation or extra testing.

455
00:18:39,800 --> 00:18:42,200
The machine program may need a different setting.

456
00:18:42,200 --> 00:18:45,640
A material change can move the critical path from supplier receipt

457
00:18:45,640 --> 00:18:47,240
to approval and release.

458
00:18:47,240 --> 00:18:49,080
That isn't bureaucracy for its own sake.

459
00:18:49,080 --> 00:18:52,520
Material rules exist because the part has to perform after it leaves the plant.

460
00:18:52,520 --> 00:18:54,840
A schedule that ignores those rules might look fast,

461
00:18:54,840 --> 00:18:58,600
right until it creates a quality or compliance problem nobody can explain later.

462
00:18:58,600 --> 00:19:00,040
Consider a more subtle case.

463
00:19:00,040 --> 00:19:02,760
The material arrives on time, passes incoming inspection,

464
00:19:02,760 --> 00:19:03,880
and sits in stock.

465
00:19:03,880 --> 00:19:05,560
The machine plan still looks clean,

466
00:19:05,560 --> 00:19:09,880
but the order requires a specific lot because it must stay traceable with an earlier subassembly,

467
00:19:09,880 --> 00:19:12,440
and that lot has been reserved for another customer commitment.

468
00:19:12,440 --> 00:19:14,680
Now the constraint isn't physical inventory.

469
00:19:14,680 --> 00:19:16,840
It's usable inventory with the correct identity.

470
00:19:16,840 --> 00:19:18,520
Planners run into this all the time.

471
00:19:18,520 --> 00:19:20,760
The ERP may report enough stock in total,

472
00:19:20,760 --> 00:19:24,120
but the production order can't use that stock because of lot status,

473
00:19:24,120 --> 00:19:27,480
shelf life, customer rules, test results, or reservation logic.

474
00:19:27,480 --> 00:19:29,880
Total quantity answers an accounting question.

475
00:19:29,880 --> 00:19:32,360
Production readiness answers a different one.

476
00:19:32,360 --> 00:19:33,640
For a live critical path,

477
00:19:33,640 --> 00:19:36,200
material needs more than a planned availability date.

478
00:19:36,200 --> 00:19:38,120
The model needs to know what the order consumes,

479
00:19:38,120 --> 00:19:39,400
how much it consumes,

480
00:19:39,400 --> 00:19:41,080
which lot of specification it requires,

481
00:19:41,080 --> 00:19:42,360
whether a partial release works,

482
00:19:42,360 --> 00:19:45,400
and which events turn material from expected into usable.

483
00:19:45,400 --> 00:19:48,760
Receipt inspection release, document approval, and issue to production

484
00:19:48,760 --> 00:19:50,760
can each change the earliest start time.

485
00:19:50,760 --> 00:19:53,720
This also changes how you respond to a late supplier delivery.

486
00:19:53,720 --> 00:19:55,720
The obvious response is to expedite the supplier,

487
00:19:55,720 --> 00:19:56,840
and that can help.

488
00:19:56,840 --> 00:20:01,640
But it won't protect the order if incoming inspection closes before the material arrives,

489
00:20:01,640 --> 00:20:05,320
or if the next production window falls after a weekend or a planned shutdown.

490
00:20:05,320 --> 00:20:06,600
You need the full chain.

491
00:20:06,600 --> 00:20:08,760
Material doesn't just feed the first operation,

492
00:20:08,760 --> 00:20:10,680
it can control every promise that follows.

493
00:20:10,680 --> 00:20:13,160
And once material becomes ready, it enters a factory,

494
00:20:13,160 --> 00:20:17,160
where it competes for the same finite resources as every other released order.

495
00:20:17,160 --> 00:20:20,120
Shared machines turn separate orders into one network.

496
00:20:20,120 --> 00:20:21,240
Once the material is ready,

497
00:20:21,240 --> 00:20:23,480
a different kind of dependency takes over.

498
00:20:23,480 --> 00:20:24,520
Shared capacity.

499
00:20:24,520 --> 00:20:26,600
Picture a single five-axis machining center

500
00:20:26,600 --> 00:20:28,840
that handles a family of high-value parts.

501
00:20:28,840 --> 00:20:30,680
It's not the only machine in the plant,

502
00:20:30,680 --> 00:20:32,040
but for this product family,

503
00:20:32,040 --> 00:20:34,600
it's often the only one with the right travel range,

504
00:20:34,600 --> 00:20:38,600
spindle capability, program approval, and inspection history.

505
00:20:38,600 --> 00:20:41,400
Every order that needs that machine joins the same queue.

506
00:20:41,400 --> 00:20:44,200
On their own, those orders can all look feasible.

507
00:20:44,200 --> 00:20:46,520
Each routing contains a machining operation.

508
00:20:46,520 --> 00:20:48,920
Each work order has a planned duration,

509
00:20:48,920 --> 00:20:51,320
and each customer order has a due date.

510
00:20:51,320 --> 00:20:53,880
But the machine can only process one job at a time,

511
00:20:53,880 --> 00:20:56,040
and the sequence you choose at that work center

512
00:20:56,040 --> 00:20:58,440
changes the completion path for every order behind it

513
00:20:58,440 --> 00:21:00,520
that turns separate orders into one network.

514
00:21:00,520 --> 00:21:02,680
Say you're the planner with four jobs waiting.

515
00:21:02,680 --> 00:21:04,280
One needs a short machining operation

516
00:21:04,280 --> 00:21:06,360
then moves into a long downstream process.

517
00:21:06,360 --> 00:21:08,200
Another needs most of the shift on the machine,

518
00:21:08,200 --> 00:21:10,120
but ships later in the week.

519
00:21:10,120 --> 00:21:12,920
A third looks urgent because of its customer priority,

520
00:21:12,920 --> 00:21:14,920
and the fourth needs a setup that matches

521
00:21:14,920 --> 00:21:16,600
the machine's current configuration.

522
00:21:16,600 --> 00:21:18,360
There isn't an automatic right answer.

523
00:21:18,360 --> 00:21:19,880
If you run the short job first,

524
00:21:19,880 --> 00:21:21,480
you may release it into the next process

525
00:21:21,480 --> 00:21:23,320
before a time-sensitive hand off closes.

526
00:21:23,320 --> 00:21:24,760
If you run the long job first,

527
00:21:24,760 --> 00:21:26,440
you might protect a later delivery date

528
00:21:26,440 --> 00:21:28,360
but block two orders that have less slack.

529
00:21:28,360 --> 00:21:29,960
If you preserve the current setup,

530
00:21:29,960 --> 00:21:31,480
you save change over time,

531
00:21:31,480 --> 00:21:33,560
but you may hold an order whose next operation

532
00:21:33,560 --> 00:21:34,760
has a narrow window.

533
00:21:34,760 --> 00:21:37,720
A local sequencing choice becomes a delivery decision

534
00:21:37,720 --> 00:21:39,560
across several customer orders.

535
00:21:39,560 --> 00:21:42,600
This is where Project CPM needs an adjustment for manufacturing.

536
00:21:42,600 --> 00:21:44,040
A project network usually assumes

537
00:21:44,040 --> 00:21:46,280
a task has access to its planned resource,

538
00:21:46,280 --> 00:21:49,320
but a factory schedule needs to account for finite capacity.

539
00:21:49,320 --> 00:21:51,080
It has to recognize that the same machine

540
00:21:51,080 --> 00:21:53,960
links work orders that have no direct product relationship at all.

541
00:21:53,960 --> 00:21:57,160
They become connected because they need the same constrained resource.

542
00:21:57,160 --> 00:21:59,240
The operation duration can even mislead you.

543
00:21:59,240 --> 00:22:01,800
A 10-minute job can block an order for a day.

544
00:22:01,800 --> 00:22:03,560
Not because the operation takes a day,

545
00:22:03,560 --> 00:22:05,240
but because it's placed in the queue

546
00:22:05,240 --> 00:22:07,720
prevents another order from reaching a process window.

547
00:22:07,720 --> 00:22:09,800
If that short job sits ahead of an order

548
00:22:09,800 --> 00:22:11,400
waiting for a furnace load,

549
00:22:11,400 --> 00:22:13,640
its sequence can decide whether the second order

550
00:22:13,640 --> 00:22:16,040
catches that load or waits for the next one.

551
00:22:16,040 --> 00:22:19,080
The short job now belongs to the other orders' critical path.

552
00:22:19,080 --> 00:22:22,200
That feels strange until you think about the factory as it runs.

553
00:22:22,200 --> 00:22:24,440
A part doesn't only depend on its own route,

554
00:22:24,440 --> 00:22:27,800
it depends on the other work competing for the resources along that route,

555
00:22:27,800 --> 00:22:29,640
and the more constrained the resource,

556
00:22:29,640 --> 00:22:32,440
the more those dependencies spread across the order book.

557
00:22:32,440 --> 00:22:35,000
Setup families add more details.

558
00:22:35,000 --> 00:22:37,400
Suppose several work orders use the same fixture,

559
00:22:37,400 --> 00:22:40,680
tool package, material grade, or machining program family.

560
00:22:40,680 --> 00:22:42,280
Grouping them can reduce setup work

561
00:22:42,280 --> 00:22:44,440
and lower the risk of configuration mistakes

562
00:22:44,440 --> 00:22:46,440
that often makes good operational sense,

563
00:22:46,440 --> 00:22:49,000
but grouping only by setup family can push

564
00:22:49,000 --> 00:22:50,920
a due date sensitive order behind work

565
00:22:50,920 --> 00:22:52,920
that happens to fit the current machine state.

566
00:22:52,920 --> 00:22:54,600
Setup efficiency and delivery protection

567
00:22:54,600 --> 00:22:56,200
can point in different directions.

568
00:22:56,200 --> 00:22:58,520
Machine qualification adds another condition.

569
00:22:58,520 --> 00:23:01,000
Two machines may appear as approved alternatives

570
00:23:01,000 --> 00:23:02,120
in a planning system,

571
00:23:02,120 --> 00:23:04,520
while only one can process a certain revision,

572
00:23:04,520 --> 00:23:05,960
user required fixture,

573
00:23:05,960 --> 00:23:07,320
or meet the measurement tolerance

574
00:23:07,320 --> 00:23:09,400
for a specific customer requirement.

575
00:23:09,400 --> 00:23:10,440
The second machine exists,

576
00:23:10,440 --> 00:23:13,240
but it doesn't create usable capacity for that operation.

577
00:23:13,240 --> 00:23:14,520
A planner needs to know the difference

578
00:23:14,520 --> 00:23:16,920
between available capacity and qualified capacity.

579
00:23:16,920 --> 00:23:17,960
Without that detail,

580
00:23:17,960 --> 00:23:19,880
a schedule can look feasible right up to the point

581
00:23:19,880 --> 00:23:21,640
where someone tries to release the job

582
00:23:21,640 --> 00:23:22,760
and then the team discovers

583
00:23:22,760 --> 00:23:24,520
that the tool sits on another machine,

584
00:23:24,520 --> 00:23:25,960
the program needs approval

585
00:23:25,960 --> 00:23:28,120
or the alternate resource cannot run the part

586
00:23:28,120 --> 00:23:30,120
under the required process rules.

587
00:23:30,120 --> 00:23:32,200
The system has counted hours that production

588
00:23:32,200 --> 00:23:33,240
can't actually use.

589
00:23:33,240 --> 00:23:34,680
That isn't a small data issue.

590
00:23:34,680 --> 00:23:36,680
It changes the dependency network.

591
00:23:36,680 --> 00:23:38,200
For a live critical path,

592
00:23:38,200 --> 00:23:40,360
each constrained machine needs a current queue,

593
00:23:40,360 --> 00:23:41,800
real remaining durations,

594
00:23:41,800 --> 00:23:43,080
allowed operations,

595
00:23:43,080 --> 00:23:46,120
and the conditions that make an operation eligible to run.

596
00:23:46,120 --> 00:23:49,080
Then the model can trace more than the root within one order.

597
00:23:49,080 --> 00:23:52,360
It can trace which other order occupies the resource first,

598
00:23:52,360 --> 00:23:54,200
which sequence creates delay,

599
00:23:54,200 --> 00:23:56,600
and which change might recover a delivery date

600
00:23:56,600 --> 00:23:58,760
without quietly damaging another one.

601
00:23:58,760 --> 00:24:01,000
This doesn't mean planners should hand every sequence

602
00:24:01,000 --> 00:24:03,000
to an algorithm and walk away.

603
00:24:03,000 --> 00:24:04,680
Local knowledge still matters.

604
00:24:04,680 --> 00:24:07,560
A supervisor may know a tool has reached its wear limit.

605
00:24:07,560 --> 00:24:10,200
An operator may know that a program runs poorly

606
00:24:10,200 --> 00:24:12,520
after a certain setup and a maintenance team

607
00:24:12,520 --> 00:24:14,040
may know that the machine can run

608
00:24:14,040 --> 00:24:16,520
but shouldn't take a long unattended cycle.

609
00:24:16,520 --> 00:24:18,680
The point is to put that knowledge into the decision,

610
00:24:18,680 --> 00:24:20,680
rather than leaving the dependency hidden.

611
00:24:20,680 --> 00:24:22,280
A bottleneck machine connects orders

612
00:24:22,280 --> 00:24:24,200
and its queue connects customer promises.

613
00:24:24,200 --> 00:24:25,560
Even a perfect machine schedule

614
00:24:25,560 --> 00:24:28,200
still leaves another constraint that can change the path.

615
00:24:28,200 --> 00:24:30,840
The people who can perform and approve the work.

616
00:24:30,840 --> 00:24:33,240
People's skills and shift calendars change the schedule.

617
00:24:33,240 --> 00:24:34,760
A machine can sit open on the plan

618
00:24:34,760 --> 00:24:36,520
and still provide no usable capacity.

619
00:24:36,520 --> 00:24:39,080
That happens when the operation needs more than machine time,

620
00:24:39,080 --> 00:24:40,760
which is common in complex production,

621
00:24:40,760 --> 00:24:42,200
where a qualified operator,

622
00:24:42,200 --> 00:24:43,320
a first piece approval,

623
00:24:43,320 --> 00:24:45,400
or a specialist sign off decides where the work

624
00:24:45,400 --> 00:24:47,640
can actually start during a given window.

625
00:24:47,640 --> 00:24:51,480
Planning systems often represent labor as a pool of available hours

626
00:24:51,480 --> 00:24:53,480
and that works for rough capacity planning.

627
00:24:53,480 --> 00:24:56,120
But on the floor, the question is usually more specific.

628
00:24:56,120 --> 00:24:58,120
Who can run this machine, this product revision,

629
00:24:58,120 --> 00:25:00,760
this toolset and this process under the required rules?

630
00:25:00,760 --> 00:25:02,760
Generic labor capacity doesn't answer that.

631
00:25:02,760 --> 00:25:05,000
Imagine an operation that needs a trained operator

632
00:25:05,000 --> 00:25:06,920
because the process includes a controlled setup

633
00:25:06,920 --> 00:25:08,520
and a first piece measurement.

634
00:25:08,520 --> 00:25:10,600
The machine becomes free at two in the afternoon,

635
00:25:10,600 --> 00:25:12,200
material weights at the work centre

636
00:25:12,200 --> 00:25:14,040
and the order carries little slack.

637
00:25:14,040 --> 00:25:16,120
But the qualified operator works the early shift

638
00:25:16,120 --> 00:25:17,320
and has already gone home.

639
00:25:17,480 --> 00:25:19,080
A colleague may know the machine well enough

640
00:25:19,080 --> 00:25:20,680
to keep a familiar job running,

641
00:25:20,680 --> 00:25:22,760
but that doesn't mean they can start this job,

642
00:25:22,760 --> 00:25:23,960
make the setup decision,

643
00:25:23,960 --> 00:25:25,800
or sign the first piece record.

644
00:25:25,800 --> 00:25:28,520
The work waits until the next qualified person is available,

645
00:25:28,520 --> 00:25:31,240
even though the machine Canada shows empty capacity.

646
00:25:31,240 --> 00:25:33,160
That waiting time belongs in the schedule.

647
00:25:33,160 --> 00:25:34,600
People also move through the factory

648
00:25:34,600 --> 00:25:36,760
in ways of standard routing rarely captures.

649
00:25:36,760 --> 00:25:38,760
A process engineer may need to approve a setup

650
00:25:38,760 --> 00:25:40,120
after an engineering change,

651
00:25:40,120 --> 00:25:42,680
a quality technician may need to release a first article,

652
00:25:42,680 --> 00:25:44,200
or a maintenance electrician may need

653
00:25:44,200 --> 00:25:45,560
to reset a protected condition

654
00:25:45,560 --> 00:25:47,560
before the machine can return to service.

655
00:25:47,560 --> 00:25:49,160
Those people become dependencies too.

656
00:25:49,160 --> 00:25:50,600
This doesn't mean every specialist needs

657
00:25:50,600 --> 00:25:52,920
to appear as a separate work centre in every plan,

658
00:25:52,920 --> 00:25:54,840
that would create a model nobody wants to maintain.

659
00:25:54,840 --> 00:25:56,920
It means the model needs to represent the conditions

660
00:25:56,920 --> 00:25:59,480
that matter for the flow you are trying to protect.

661
00:25:59,480 --> 00:26:01,880
If first piece approval blocks a constrained operation,

662
00:26:01,880 --> 00:26:03,400
the schedule needs to know that.

663
00:26:03,400 --> 00:26:04,840
Shift calendars add another layer.

664
00:26:04,840 --> 00:26:07,400
An operation may require four hours of touch time,

665
00:26:07,400 --> 00:26:09,880
but four hours of work don't always equal four hours

666
00:26:09,880 --> 00:26:11,080
of elapsed time.

667
00:26:11,080 --> 00:26:12,840
Started late in an afternoon shift

668
00:26:12,840 --> 00:26:14,360
and the job may pause at handover,

669
00:26:14,360 --> 00:26:15,560
or started before a weekend

670
00:26:15,560 --> 00:26:18,520
and the next required skill may not return until Monday.

671
00:26:18,520 --> 00:26:20,600
The calendar changes the path, take an order

672
00:26:20,600 --> 00:26:22,600
that reaches a test cell at five o'clock.

673
00:26:22,600 --> 00:26:24,360
The test itself takes less than an hour

674
00:26:24,360 --> 00:26:25,720
so on paper it looks harmless,

675
00:26:25,720 --> 00:26:29,080
but the person authorised to review the result works until 430

676
00:26:29,080 --> 00:26:31,800
and no one else can release the unit for the next process.

677
00:26:31,800 --> 00:26:33,320
The job doesn't move that evening.

678
00:26:33,320 --> 00:26:36,200
A simple duration estimate would place the operation at one hour,

679
00:26:36,200 --> 00:26:37,880
but the actual lead time becomes longer

680
00:26:37,880 --> 00:26:39,880
because the task includes a review gate

681
00:26:39,880 --> 00:26:42,040
tied to a person and a calendar.

682
00:26:42,040 --> 00:26:44,120
If the customer commitment sits close,

683
00:26:44,120 --> 00:26:47,000
that review becomes the dependency controlling delivery.

684
00:26:47,000 --> 00:26:48,440
Overtime can change the answer,

685
00:26:48,440 --> 00:26:50,440
but it isn't free capacity in the abstract.

686
00:26:50,440 --> 00:26:52,200
Someone needs approval to authorise it.

687
00:26:52,200 --> 00:26:55,160
The right operator has to be willing and permitted to work it

688
00:26:55,160 --> 00:26:56,840
and safety rules for T-glimits,

689
00:26:56,840 --> 00:27:00,040
labour agreements and local policy may limit the choice.

690
00:27:00,040 --> 00:27:01,800
A plan that assumes unlimited overtime

691
00:27:01,800 --> 00:27:04,440
is just a plan that has hidden its problem.

692
00:27:04,440 --> 00:27:05,880
Even when overtime is available,

693
00:27:05,880 --> 00:27:08,920
planners need to ask whether it helps the right place.

694
00:27:08,920 --> 00:27:10,440
Bringing in an operator for an operation

695
00:27:10,440 --> 00:27:12,600
with several hours of slack may feel productive,

696
00:27:12,600 --> 00:27:15,160
while a short approval task on the actual path

697
00:27:15,160 --> 00:27:16,520
still waits until the next day.

698
00:27:16,520 --> 00:27:19,160
That is how a plan spends money

699
00:27:19,160 --> 00:27:20,520
without changing the ship date.

700
00:27:20,520 --> 00:27:24,040
Skill rules can also change after production begins.

701
00:27:24,040 --> 00:27:26,840
An experienced operator may be reassigned to a high-risk job,

702
00:27:26,840 --> 00:27:28,840
a trainee may run the machine under supervision

703
00:27:28,840 --> 00:27:30,520
but cannot release the work alone

704
00:27:30,520 --> 00:27:33,160
or a new product revision may require a different certification.

705
00:27:33,160 --> 00:27:36,600
These aren't edge cases in high-mix manufacturing.

706
00:27:36,600 --> 00:27:39,480
They are normal conditions that separate theoretical capacity

707
00:27:39,480 --> 00:27:41,720
from capacity you can safely use.

708
00:27:41,720 --> 00:27:43,560
So when you calculate a live critical path,

709
00:27:43,560 --> 00:27:45,640
don't ask only when the machine is free.

710
00:27:45,640 --> 00:27:48,680
Ask when the machine, material, qualified person

711
00:27:48,680 --> 00:27:51,960
and required approval all become available at the same time,

712
00:27:51,960 --> 00:27:54,920
that overlap determines the earliest legal and practical start.

713
00:27:54,920 --> 00:27:58,440
A machine slot without the right person isn't a slot,

714
00:27:58,440 --> 00:27:59,800
it's a gap in the calendar.

715
00:27:59,800 --> 00:28:01,880
This is also where planning and production

716
00:28:01,880 --> 00:28:03,720
need to work from the same definitions.

717
00:28:03,720 --> 00:28:05,560
Planning may own the schedule logic

718
00:28:05,560 --> 00:28:07,720
but production knows who can run which jobs

719
00:28:07,720 --> 00:28:09,400
under real-shift conditions

720
00:28:09,400 --> 00:28:12,200
and if that knowledge lives only in the supervisor's head,

721
00:28:12,200 --> 00:28:14,760
the schedule will keep discovering it too late.

722
00:28:14,760 --> 00:28:17,560
The dependency model should capture stable qualification rules

723
00:28:17,560 --> 00:28:18,920
and calendar constraints

724
00:28:18,920 --> 00:28:21,400
and local teams still need a way to report the exceptions

725
00:28:21,400 --> 00:28:22,680
that change today's plan.

726
00:28:22,680 --> 00:28:25,800
That gives you a schedule based on actual operating conditions,

727
00:28:25,800 --> 00:28:28,200
not a convenient assumption that any available pair of hands

728
00:28:28,200 --> 00:28:30,280
can run any open machine.

729
00:28:30,280 --> 00:28:31,800
And once the right person is available,

730
00:28:31,800 --> 00:28:34,760
you still have to check whether the equipment is truly ready for production

731
00:28:34,760 --> 00:28:37,080
or whether the schedule only thinks it is.

732
00:28:37,080 --> 00:28:39,320
Machine status is not production readiness.

733
00:28:39,320 --> 00:28:40,360
Here's the problem.

734
00:28:40,360 --> 00:28:44,280
Most manufacturers don't talk about a machine status like running

735
00:28:44,280 --> 00:28:49,000
or available doesn't prove production can actually start the operation you need.

736
00:28:49,000 --> 00:28:51,480
A schedule may show the machine as back in service

737
00:28:51,480 --> 00:28:53,400
but none of those labels guarantees

738
00:28:53,400 --> 00:28:55,400
that the next job is ready to go.

739
00:28:55,400 --> 00:28:57,160
Production readiness is more specific,

740
00:28:57,160 --> 00:28:59,240
the machine needs to be capable of the work

741
00:28:59,240 --> 00:29:01,160
in the right condition with the right tooling,

742
00:29:01,160 --> 00:29:02,440
the approved program,

743
00:29:02,440 --> 00:29:04,520
release material and someone ready to run it.

744
00:29:04,520 --> 00:29:08,280
Miss one of those conditions and the slot on the schedule is only theoretical.

745
00:29:08,280 --> 00:29:09,480
Think of it like this.

746
00:29:09,480 --> 00:29:11,960
You can't start a job just because the spindle spins,

747
00:29:11,960 --> 00:29:14,600
take a machining center that comes back after an alarm.

748
00:29:14,600 --> 00:29:16,040
Maintenance clears the immediate fault,

749
00:29:16,040 --> 00:29:17,720
the controller shows the machine ready

750
00:29:17,720 --> 00:29:19,480
and planning sees capacity again

751
00:29:19,480 --> 00:29:22,440
so they move an urgent job into the next open slot.

752
00:29:22,440 --> 00:29:24,920
Then the team finds the spindle needs a warm-up cycle,

753
00:29:24,920 --> 00:29:27,320
the previous job left a different fixture installed

754
00:29:27,320 --> 00:29:29,880
and the required probe check still needs completion

755
00:29:29,880 --> 00:29:32,360
before the controlled operation can begin.

756
00:29:32,360 --> 00:29:34,760
The equipment runs, the order still waits.

757
00:29:34,760 --> 00:29:38,440
That difference gets lost when people use a single status field for a machine.

758
00:29:38,440 --> 00:29:40,920
Running can mean the spindle turns

759
00:29:40,920 --> 00:29:42,600
or it can mean a job is active

760
00:29:42,600 --> 00:29:45,560
or it can mean the machine is connected to the monitoring system.

761
00:29:45,560 --> 00:29:47,160
None of those meanings tells the planner

762
00:29:47,160 --> 00:29:50,120
whether this specific operation can safely start

763
00:29:50,120 --> 00:29:51,880
and finish within the required window.

764
00:29:51,880 --> 00:29:54,920
A status needs context and that context isn't a green light.

765
00:29:54,920 --> 00:29:55,960
Let's zoom out for a second.

766
00:29:55,960 --> 00:29:57,320
Consider planned maintenance.

767
00:29:57,320 --> 00:29:59,560
The machine may work perfectly at 9 in the morning

768
00:29:59,560 --> 00:30:01,160
but a maintenance window begins at noon

769
00:30:01,160 --> 00:30:03,560
and the operation needs five hours after setup.

770
00:30:03,560 --> 00:30:05,240
Starting it might create a poor handover

771
00:30:05,240 --> 00:30:06,440
and interrupted cycle

772
00:30:06,440 --> 00:30:08,680
or a decision to defer the maintenance task.

773
00:30:08,680 --> 00:30:11,000
Those choices affect more than one work order

774
00:30:11,000 --> 00:30:13,720
and a basic availability flag won't explain them.

775
00:30:13,720 --> 00:30:16,680
In practical terms that means you can't schedule around a status light.

776
00:30:16,680 --> 00:30:18,120
Cleaning, change over and setup

777
00:30:18,120 --> 00:30:19,720
belong in the same conversation.

778
00:30:19,720 --> 00:30:21,720
A coating line may show idle capacity

779
00:30:21,720 --> 00:30:24,360
but it means cleaning before it can process a different finish.

780
00:30:24,360 --> 00:30:25,960
A test cell may sit empty

781
00:30:25,960 --> 00:30:27,800
while its fixture remains in calibration

782
00:30:27,800 --> 00:30:30,600
or its software version doesn't match the product revision.

783
00:30:30,600 --> 00:30:31,800
The resource exists

784
00:30:31,800 --> 00:30:34,760
but it hasn't reached a state where it can process that work.

785
00:30:34,760 --> 00:30:37,560
That distinction changes the earliest possible start

786
00:30:37,560 --> 00:30:39,320
and plan is rarely see it.

787
00:30:39,320 --> 00:30:40,840
Now here's a less obvious issue.

788
00:30:40,840 --> 00:30:42,600
A machine may run within normal limits

789
00:30:42,600 --> 00:30:44,280
while its actual cycle time drifts away

790
00:30:44,280 --> 00:30:46,040
from the standard time used in planning.

791
00:30:46,040 --> 00:30:48,120
Perhaps material requires more passes

792
00:30:48,120 --> 00:30:50,600
or the tool condition changes or a program revision

793
00:30:50,600 --> 00:30:53,720
adds checks that the old time standard didn't include.

794
00:30:53,720 --> 00:30:54,920
The machine doesn't stop

795
00:30:54,920 --> 00:30:57,880
but it consumes more capacity than the plan expected.

796
00:30:57,880 --> 00:31:00,360
Small changes can travel through a loaded schedule.

797
00:31:00,360 --> 00:31:03,800
If a constrained resource runs slightly slower across several jobs

798
00:31:03,800 --> 00:31:07,160
the last job in the queue may lose the slot it needs for the next operation

799
00:31:07,160 --> 00:31:09,000
nothing appears as a major incident.

800
00:31:09,000 --> 00:31:11,400
OEE may still look acceptable for the shift

801
00:31:11,400 --> 00:31:12,760
yet the orders path changes

802
00:31:12,760 --> 00:31:15,960
because the remaining capacity no longer fits the same sequence.

803
00:31:15,960 --> 00:31:18,360
OEE helps you understand equipment performance

804
00:31:18,360 --> 00:31:21,720
it can point to availability loss, speed loss and quality loss

805
00:31:21,720 --> 00:31:24,120
but OEE doesn't calculate a feasible schedule.

806
00:31:24,120 --> 00:31:27,240
A high OEE result doesn't tell you whether the right machine,

807
00:31:27,240 --> 00:31:29,320
tool operator and process release

808
00:31:29,320 --> 00:31:31,320
come together before a customer cutoff.

809
00:31:31,320 --> 00:31:35,080
A lower OEE result doesn't automatically mean an order will ship late.

810
00:31:35,080 --> 00:31:36,680
It's a signal not a planning answer.

811
00:31:36,680 --> 00:31:39,960
This matters when shop floor events flow into planning systems.

812
00:31:39,960 --> 00:31:42,200
A raw machine signal might report an alarm,

813
00:31:42,200 --> 00:31:45,880
a state change, a cycle complete or a long idle period.

814
00:31:45,880 --> 00:31:47,320
Those events can be useful

815
00:31:47,320 --> 00:31:50,280
but they need interpretation before they move a production promise.

816
00:31:50,280 --> 00:31:52,280
An alarm may stop the current operation

817
00:31:52,280 --> 00:31:55,160
or it may clear in seconds and have no impact on output.

818
00:31:55,160 --> 00:31:57,640
A cycle complete signal may mean one part finished

819
00:31:57,640 --> 00:32:01,000
but not that the quantity required for transfer has become available.

820
00:32:01,000 --> 00:32:03,400
An idle state may mean the machine waits for material,

821
00:32:03,400 --> 00:32:05,720
an operator, a tool, a quality release

822
00:32:05,720 --> 00:32:07,400
or simply the next scheduled job.

823
00:32:07,400 --> 00:32:09,400
Raw telemetry can't tell that story alone.

824
00:32:09,400 --> 00:32:11,720
The MES can add execution context

825
00:32:11,720 --> 00:32:14,280
which order ran, what quantity completed,

826
00:32:14,280 --> 00:32:15,640
whether the operation stopped

827
00:32:15,640 --> 00:32:17,960
and whether scrap or a hold changed the result.

828
00:32:17,960 --> 00:32:21,080
Production and maintenance add another layer

829
00:32:21,080 --> 00:32:23,080
whether the asset can take the next job

830
00:32:23,080 --> 00:32:24,760
under which limits and from when.

831
00:32:25,640 --> 00:32:27,320
Only then should planning recalculate.

832
00:32:27,320 --> 00:32:29,480
So what question does a planner actually need answered

833
00:32:29,480 --> 00:32:30,680
after an equipment event?

834
00:32:30,680 --> 00:32:32,280
Not is the machine green?

835
00:32:32,280 --> 00:32:35,640
The useful question is which operations can this resource complete

836
00:32:35,640 --> 00:32:38,120
from what time, at what realistic rate

837
00:32:38,120 --> 00:32:39,960
and under what conditions?

838
00:32:39,960 --> 00:32:41,720
That answer can change the critical path

839
00:32:41,720 --> 00:32:43,640
without a single dramatic breakdown.

840
00:32:43,640 --> 00:32:45,080
A job may lose its place

841
00:32:45,080 --> 00:32:47,080
because setup takes longer than expected.

842
00:32:47,080 --> 00:32:48,360
Another job may move ahead

843
00:32:48,360 --> 00:32:50,280
because it fits the machine's current tooling

844
00:32:50,280 --> 00:32:52,360
and still protects a downstream deadline.

845
00:32:52,360 --> 00:32:55,480
A maintenance constraint may force a split in available time.

846
00:32:55,480 --> 00:32:57,800
Each event changes more than capacity.

847
00:32:57,800 --> 00:32:59,960
It changes the timing of dependent work.

848
00:32:59,960 --> 00:33:01,880
That's why real-time visibility only helps

849
00:33:01,880 --> 00:33:04,120
when the event connects to production meaning.

850
00:33:04,120 --> 00:33:06,840
Otherwise you get a very accurate record of the machine state

851
00:33:06,840 --> 00:33:09,160
and still no clear answer about the customer order.

852
00:33:09,160 --> 00:33:11,080
The dependencies planners miss most often

853
00:33:11,080 --> 00:33:12,360
sit between operations

854
00:33:12,360 --> 00:33:13,800
in the time when a part appears finished

855
00:33:13,800 --> 00:33:15,400
but can't yet move forward.

856
00:33:15,400 --> 00:33:18,280
Transfer batches, queues, and hidden waiting time.

857
00:33:18,280 --> 00:33:20,040
Let me give you a scenario that sounds simple

858
00:33:20,040 --> 00:33:21,560
but causes real headaches.

859
00:33:21,560 --> 00:33:22,840
A part can finish an operation

860
00:33:22,840 --> 00:33:24,200
and still not move forward.

861
00:33:24,200 --> 00:33:26,360
That gap often controls delivery

862
00:33:26,360 --> 00:33:28,280
more than the time spent cutting, assembling,

863
00:33:28,280 --> 00:33:29,720
or testing the part itself.

864
00:33:29,720 --> 00:33:32,440
Consider an order moving from machining into heat treatment.

865
00:33:32,440 --> 00:33:34,680
The machining team completes the first pieces early

866
00:33:34,680 --> 00:33:37,160
and from a local view the job looks ahead of plan.

867
00:33:37,160 --> 00:33:39,880
But heat treatment may require a full transfer batch

868
00:33:39,880 --> 00:33:41,560
not a few finished pieces.

869
00:33:41,560 --> 00:33:43,800
Until the required quantity reaches the staging area

870
00:33:43,800 --> 00:33:45,560
the furnace team can't load the work.

871
00:33:45,560 --> 00:33:46,600
Completion at one operation

872
00:33:46,600 --> 00:33:48,760
doesn't always create a usable handoff.

873
00:33:48,760 --> 00:33:50,120
Transfer batch rules decide

874
00:33:50,120 --> 00:33:52,200
how workflows between operations.

875
00:33:52,200 --> 00:33:54,040
In some processes the whole production lot

876
00:33:54,040 --> 00:33:56,280
must finish before the next process starts.

877
00:33:56,280 --> 00:33:58,840
In others a smaller transfer quantity can move ahead

878
00:33:58,840 --> 00:34:01,320
while the previous work centre continues with the rest.

879
00:34:01,320 --> 00:34:03,480
That overlap can reduce elapsed lead time

880
00:34:03,480 --> 00:34:05,080
but it also adds more tracking work.

881
00:34:05,080 --> 00:34:07,000
You need to know how much quantity has completed

882
00:34:07,000 --> 00:34:08,920
which quantity passed the required checks

883
00:34:08,920 --> 00:34:10,920
and whether the next operation can accept it.

884
00:34:10,920 --> 00:34:13,320
A status of operation complete often hides

885
00:34:13,320 --> 00:34:15,000
more than it explains.

886
00:34:15,000 --> 00:34:17,160
Say a machining order needs 100 parts

887
00:34:17,160 --> 00:34:19,640
and the furnace accepts loads in groups of 50.

888
00:34:19,640 --> 00:34:21,800
If the first 50 parts finish at the right time

889
00:34:21,800 --> 00:34:23,960
the order may catch an early furnace load

890
00:34:23,960 --> 00:34:26,040
while machining continues with the rest.

891
00:34:26,040 --> 00:34:29,480
If only 48 parts finish the furnace load may leave without them.

892
00:34:29,480 --> 00:34:32,120
Two parts then create a weight that affects the whole batch.

893
00:34:32,120 --> 00:34:34,360
That sounds almost silly until you see it happen.

894
00:34:34,360 --> 00:34:35,400
The machine didn't fail.

895
00:34:35,400 --> 00:34:36,600
The workers didn't stop.

896
00:34:36,600 --> 00:34:38,520
The order simply missed a transfer rule.

897
00:34:38,520 --> 00:34:40,360
Yet the path changes because the next process

898
00:34:40,360 --> 00:34:42,600
follows its own capacity and loading pattern

899
00:34:42,600 --> 00:34:44,760
not the timing of one upstream operation.

900
00:34:44,760 --> 00:34:46,520
Cues create the same effect

901
00:34:46,520 --> 00:34:48,120
especially around shared processes

902
00:34:48,120 --> 00:34:50,120
like heat treatment, coating,

903
00:34:50,120 --> 00:34:52,520
washing, painting, or final inspection.

904
00:34:52,520 --> 00:34:54,840
A part may physically arrive at the next department

905
00:34:54,840 --> 00:34:57,160
but the department already has work waiting.

906
00:34:57,160 --> 00:34:59,240
Its planned start time means very little

907
00:34:59,240 --> 00:35:01,160
if the queue contains several other jobs

908
00:35:01,160 --> 00:35:04,040
that need the same furnace, booth, chamber, or test bench.

909
00:35:04,040 --> 00:35:05,000
The part has arrived.

910
00:35:05,000 --> 00:35:06,360
It hasn't entered the process.

911
00:35:06,360 --> 00:35:08,120
Batch furnaces add their own logic

912
00:35:08,120 --> 00:35:10,600
because they don't behave like a single machine processing

913
00:35:10,600 --> 00:35:11,960
one order after another.

914
00:35:11,960 --> 00:35:13,400
The team needs a compatible load.

915
00:35:13,400 --> 00:35:15,560
Parts may need the same cycle, temperature range,

916
00:35:15,560 --> 00:35:17,560
atmosphere, fixture, or customer approval.

917
00:35:17,560 --> 00:35:19,720
A furnace can have open physical space

918
00:35:19,720 --> 00:35:21,480
while the order still can't join the load

919
00:35:21,480 --> 00:35:23,480
because its process requirements don't match.

920
00:35:23,480 --> 00:35:25,800
So capacity isn't just a number of open hours.

921
00:35:25,800 --> 00:35:28,440
It's a set of load opportunities, each with rules.

922
00:35:28,440 --> 00:35:29,560
The planner needs to know

923
00:35:29,560 --> 00:35:31,400
when the next compatible batch can run

924
00:35:31,400 --> 00:35:33,160
whether the order will be ready in time

925
00:35:33,160 --> 00:35:35,800
and whether adding it displaces another order with less slack.

926
00:35:35,800 --> 00:35:38,920
A simple queue position won't always answer that.

927
00:35:38,920 --> 00:35:41,000
Internal transport adds more waiting

928
00:35:41,000 --> 00:35:42,840
that rarely appears in a basic routing.

929
00:35:42,840 --> 00:35:44,760
A completed pallet may wait for a forklift.

930
00:35:44,760 --> 00:35:46,840
It may need a controlled route between areas.

931
00:35:46,840 --> 00:35:49,560
The receiving work centre may lack staging space

932
00:35:49,640 --> 00:35:51,400
In some plans, work cannot move

933
00:35:51,400 --> 00:35:54,520
because the next area already holds its maximum WIP.

934
00:35:54,520 --> 00:35:56,600
That limit can protect flow.

935
00:35:56,600 --> 00:35:59,000
It prevents one department from filling another with parts

936
00:35:59,000 --> 00:36:00,040
that can't process,

937
00:36:00,040 --> 00:36:03,080
but it also means an upstream operation may need to pause

938
00:36:03,080 --> 00:36:05,800
even when its machine and operator remain available.

939
00:36:05,800 --> 00:36:08,120
The order's next step depends on physical space,

940
00:36:08,120 --> 00:36:09,320
transport timing,

941
00:36:09,320 --> 00:36:11,080
and the readiness of the receiving area.

942
00:36:11,080 --> 00:36:12,520
These constraints don't show up well

943
00:36:12,520 --> 00:36:14,920
when every operation gets only a standard runtime.

944
00:36:14,920 --> 00:36:16,200
A planner might see

945
00:36:16,200 --> 00:36:17,480
30 minutes of machining,

946
00:36:17,480 --> 00:36:18,840
2 hours of heat treatment,

947
00:36:18,840 --> 00:36:20,280
and 20 minutes of inspection.

948
00:36:20,280 --> 00:36:22,840
The actual elapsed time can stretch across a day or more

949
00:36:22,840 --> 00:36:24,520
because the order waits for a batch,

950
00:36:24,520 --> 00:36:25,480
a queue position,

951
00:36:25,480 --> 00:36:26,360
a transport move,

952
00:36:26,360 --> 00:36:28,360
and an open staging location.

953
00:36:28,360 --> 00:36:31,240
Touchtime tells you how long someone works on the part.

954
00:36:31,240 --> 00:36:33,400
Flowtime tells you how long the customer waits.

955
00:36:33,400 --> 00:36:34,680
Those are different measures.

956
00:36:34,680 --> 00:36:36,920
Value stream mapping often makes this visible

957
00:36:36,920 --> 00:36:39,160
because it follows the work across departments

958
00:36:39,160 --> 00:36:40,680
and records where it waits.

959
00:36:40,680 --> 00:36:42,440
You don't need a large transformation programme

960
00:36:42,440 --> 00:36:43,640
to learn from that exercise.

961
00:36:43,640 --> 00:36:45,800
Map one late order from release to dispatch,

962
00:36:45,800 --> 00:36:47,720
then ask how much time went into processing

963
00:36:47,720 --> 00:36:50,280
and how much time went into waiting for the next condition.

964
00:36:50,280 --> 00:36:51,640
The answer can be uncomfortable.

965
00:36:51,640 --> 00:36:53,240
It can also change where you focus.

966
00:36:53,240 --> 00:36:55,080
If the order spends most of its time waiting

967
00:36:55,080 --> 00:36:56,600
for a compatible furnace load,

968
00:36:56,600 --> 00:36:59,160
reducing machining cycle time may not protect the due date,

969
00:36:59,160 --> 00:37:00,760
you may need better batch planning,

970
00:37:00,760 --> 00:37:01,640
earlier release,

971
00:37:01,640 --> 00:37:03,080
a different transfer quantity,

972
00:37:03,080 --> 00:37:05,640
or a decision about which order enters the next load.

973
00:37:05,640 --> 00:37:07,640
The critical path includes those waits

974
00:37:07,640 --> 00:37:09,480
because the customer experiences them

975
00:37:09,480 --> 00:37:10,920
just like any other delay.

976
00:37:10,920 --> 00:37:12,440
A part doesn't care whether it waits

977
00:37:12,440 --> 00:37:15,160
because of a machine, a queue, or a pallet location.

978
00:37:15,160 --> 00:37:16,200
It just doesn't move,

979
00:37:16,200 --> 00:37:17,720
and there is one form of waiting

980
00:37:17,720 --> 00:37:19,480
that can appear without warning.

981
00:37:19,480 --> 00:37:22,200
Because it depends on the result of work already completed.

982
00:37:22,200 --> 00:37:23,480
A quality hold?

983
00:37:23,480 --> 00:37:25,880
Quality holds can rewrite the entire network.

984
00:37:25,880 --> 00:37:27,640
A quality hold doesn't just flag a part,

985
00:37:27,640 --> 00:37:29,320
it can stop an order after everyone thought

986
00:37:29,320 --> 00:37:30,760
the hard work was finished.

987
00:37:30,760 --> 00:37:33,080
Imagine a batch leaving surface finishing

988
00:37:33,080 --> 00:37:34,600
and moving to final inspection.

989
00:37:34,600 --> 00:37:35,800
The parts look complete.

990
00:37:35,800 --> 00:37:37,080
Machine time is behind you.

991
00:37:37,080 --> 00:37:38,760
The next work centre has room.

992
00:37:38,760 --> 00:37:40,760
The order seems almost ready for packing.

993
00:37:40,760 --> 00:37:43,240
Then inspection finds a result outside the allowed range.

994
00:37:43,240 --> 00:37:44,600
The order stops right there.

995
00:37:44,600 --> 00:37:47,880
That inspection gate isn't a report about production after the fact.

996
00:37:47,880 --> 00:37:49,800
It is part of the production path itself.

997
00:37:49,800 --> 00:37:51,720
Until quality releases the parts,

998
00:37:51,720 --> 00:37:54,520
the next operation, shipment, or customer acceptance step

999
00:37:54,520 --> 00:37:55,640
may stay blocked,

1000
00:37:55,640 --> 00:37:58,760
even when every physical resource downstream sits idle waiting.

1001
00:37:58,760 --> 00:38:01,400
A hold also creates uncertainty,

1002
00:38:01,400 --> 00:38:03,560
and uncertainty changes how you plan.

1003
00:38:03,560 --> 00:38:05,160
The team might need to measure more parts,

1004
00:38:05,160 --> 00:38:07,240
review the process record, check the material lot,

1005
00:38:07,240 --> 00:38:09,640
or decide whether the result affects the full batch.

1006
00:38:09,640 --> 00:38:11,720
During that time, the schedule can't safely assume

1007
00:38:11,720 --> 00:38:13,240
the work will continue as planned.

1008
00:38:13,240 --> 00:38:15,800
The path now includes a decision and a release condition,

1009
00:38:15,800 --> 00:38:17,320
not just a machine operation.

1010
00:38:17,320 --> 00:38:20,040
Sometimes quality releases the parts after review.

1011
00:38:20,040 --> 00:38:22,280
In that case, the delay may eat into float

1012
00:38:22,280 --> 00:38:24,120
but leave the due date intact.

1013
00:38:24,120 --> 00:38:26,840
Other times, the result triggers a non-conformance process

1014
00:38:26,840 --> 00:38:28,520
and sends the order down a rework route.

1015
00:38:28,520 --> 00:38:30,920
That route can look nothing like the original one.

1016
00:38:30,920 --> 00:38:32,680
A part may return to machining,

1017
00:38:32,680 --> 00:38:34,280
go through a repair operation,

1018
00:38:34,280 --> 00:38:35,560
get another coating pass,

1019
00:38:35,560 --> 00:38:36,920
or require new testing.

1020
00:38:36,920 --> 00:38:38,520
It may need replacement material.

1021
00:38:38,520 --> 00:38:40,680
The original routing described how to build

1022
00:38:40,680 --> 00:38:42,920
a conforming part under normal conditions.

1023
00:38:42,920 --> 00:38:45,400
The rework route describes what the plant must do

1024
00:38:45,400 --> 00:38:47,880
after a specific result changes those conditions.

1025
00:38:47,880 --> 00:38:49,960
This is what I call a conditional dependency.

1026
00:38:49,960 --> 00:38:51,720
It exists only if something goes wrong,

1027
00:38:51,720 --> 00:38:55,080
but when it appears, it can become the longest chain to delivery.

1028
00:38:55,080 --> 00:38:56,360
The rework may need a machine

1029
00:38:56,360 --> 00:38:58,120
that already carries a full schedule.

1030
00:38:58,120 --> 00:39:00,040
It may need the same heat treatment window

1031
00:39:00,040 --> 00:39:01,240
the order missed earlier.

1032
00:39:01,240 --> 00:39:03,080
It may require a specialist who can approve

1033
00:39:03,080 --> 00:39:05,480
the repair method or perform the retest.

1034
00:39:05,480 --> 00:39:07,720
The quality result changes both time and demand.

1035
00:39:07,720 --> 00:39:10,200
That's easy to miss if a system treats quality

1036
00:39:10,200 --> 00:39:13,800
as a simple pass or fail field attached to the end of an operation.

1037
00:39:13,800 --> 00:39:16,680
A failed result does more than change the status of one part.

1038
00:39:16,680 --> 00:39:19,000
It can create new work, consume scarce capacity,

1039
00:39:19,000 --> 00:39:19,960
change material needs,

1040
00:39:19,960 --> 00:39:21,720
and push other orders backward in the queue.

1041
00:39:21,720 --> 00:39:23,720
Think about a constrained inspection bench.

1042
00:39:23,720 --> 00:39:25,640
Planned production work arrives throughout the day,

1043
00:39:25,640 --> 00:39:27,960
each order needing its normal inspection step.

1044
00:39:27,960 --> 00:39:29,160
Then a rework batch arrives

1045
00:39:29,160 --> 00:39:31,960
and needs urgent retesting before a customer deadline.

1046
00:39:31,960 --> 00:39:34,440
If the team inserts that batch ahead of plant work,

1047
00:39:34,440 --> 00:39:36,120
the retest may protect one shipment

1048
00:39:36,120 --> 00:39:38,680
while it reduces the remaining slack on several others.

1049
00:39:38,680 --> 00:39:40,680
Quality work competes for capacity,

1050
00:39:40,680 --> 00:39:41,960
just like production work.

1051
00:39:41,960 --> 00:39:43,640
Same story at a bottleneck machine.

1052
00:39:43,640 --> 00:39:45,560
A repair can feel like an exception

1053
00:39:45,560 --> 00:39:47,400
that needs immediate attention.

1054
00:39:47,400 --> 00:39:48,440
In some cases it does,

1055
00:39:48,440 --> 00:39:51,000
but if the machine can only process one job at a time,

1056
00:39:51,000 --> 00:39:53,000
the rework sequence becomes a planning choice

1057
00:39:53,000 --> 00:39:53,960
with wider effects.

1058
00:39:53,960 --> 00:39:55,880
You can't treat it as invisible work.

1059
00:39:55,880 --> 00:39:57,800
This is where planners need more than an alert

1060
00:39:57,800 --> 00:39:59,320
that reads quality hold.

1061
00:39:59,320 --> 00:40:01,320
They need to know which order quantity is affected,

1062
00:40:01,320 --> 00:40:03,000
which downstream work is blocked,

1063
00:40:03,000 --> 00:40:05,240
what rework or retest steps could follow,

1064
00:40:05,240 --> 00:40:07,160
and which resources those steps will need

1065
00:40:07,160 --> 00:40:08,760
if the hold becomes a failure.

1066
00:40:08,760 --> 00:40:10,200
That doesn't mean every possible failure

1067
00:40:10,200 --> 00:40:11,960
pass belongs in the daily schedule.

1068
00:40:11,960 --> 00:40:14,440
You'd end up planning for every bad day before breakfast.

1069
00:40:14,440 --> 00:40:16,120
It means the model needs enough structure

1070
00:40:16,120 --> 00:40:18,680
to respond when a real quality event occurs.

1071
00:40:18,680 --> 00:40:21,000
The response starts with clear business rules.

1072
00:40:21,000 --> 00:40:23,240
Can the next operation begin under a control deviation

1073
00:40:23,240 --> 00:40:24,680
or must all work stop?

1074
00:40:24,680 --> 00:40:26,440
Does the issue affect one serial number,

1075
00:40:26,440 --> 00:40:28,040
a sample or the full lot?

1076
00:40:28,040 --> 00:40:29,240
Who can release the decision?

1077
00:40:29,240 --> 00:40:32,040
Does rework need an approved instruction before it can start?

1078
00:40:32,040 --> 00:40:34,520
These rules determine whether the hold creates a short pause

1079
00:40:34,520 --> 00:40:36,600
or a new branch in the dependency network.

1080
00:40:37,080 --> 00:40:39,880
A system can calculate timing only after those conditions

1081
00:40:39,880 --> 00:40:41,080
have real meaning.

1082
00:40:41,080 --> 00:40:42,440
There is also a trust issue.

1083
00:40:42,440 --> 00:40:43,800
If production records an operation

1084
00:40:43,800 --> 00:40:46,520
is complete before quality records the hold,

1085
00:40:46,520 --> 00:40:48,360
planning may assume capacity.

1086
00:40:48,360 --> 00:40:51,000
Downstream will receive work that cannot actually move.

1087
00:40:51,000 --> 00:40:54,280
Supervisors then see a schedule that looks feasible,

1088
00:40:54,280 --> 00:40:56,840
but conflicts with the product status they must protect.

1089
00:40:56,840 --> 00:40:59,240
That kind of conflict doesn't come from poor intent.

1090
00:40:59,240 --> 00:41:01,480
It comes from treating quality as separate from flow.

1091
00:41:01,480 --> 00:41:02,680
Quality is part of flow.

1092
00:41:02,680 --> 00:41:04,600
It decides whether the product can proceed.

1093
00:41:04,600 --> 00:41:07,960
Once you model that clearly, a hold becomes more than a red status.

1094
00:41:07,960 --> 00:41:09,720
It becomes an event that can trace through,

1095
00:41:09,720 --> 00:41:12,120
blocked operations, possible rework, shared resources

1096
00:41:12,120 --> 00:41:13,400
and customer commitments.

1097
00:41:13,400 --> 00:41:15,800
The planner can see not only that an order is at risk,

1098
00:41:15,800 --> 00:41:18,440
but why the path changed and where the next decision lives.

1099
00:41:18,440 --> 00:41:20,680
And that decision can reach beyond the current order.

1100
00:41:20,680 --> 00:41:23,800
A change to the product definition can arrive while work already moves

1101
00:41:23,800 --> 00:41:26,440
through the plant bringing a different set of dependencies with it.

1102
00:41:26,440 --> 00:41:28,600
Engineering changes arrive mid-order.

1103
00:41:28,600 --> 00:41:30,360
An engineering change can enter production

1104
00:41:30,360 --> 00:41:31,800
while the order is already moving

1105
00:41:31,800 --> 00:41:33,400
and it changes more than a document number?

1106
00:41:33,400 --> 00:41:36,920
Imagine a component that has completed its first machining step under revision A.

1107
00:41:36,920 --> 00:41:39,160
Before the next operation begins,

1108
00:41:39,160 --> 00:41:41,720
engineering releases revision B because a dimension,

1109
00:41:41,720 --> 00:41:43,880
material callout, process setting,

1110
00:41:43,880 --> 00:41:45,800
or assembly interface needs to change.

1111
00:41:45,800 --> 00:41:48,760
The production order now sits between two valid states.

1112
00:41:48,760 --> 00:41:51,960
Work completed under the earlier revision may still be acceptable

1113
00:41:51,960 --> 00:41:53,160
or it may need review.

1114
00:41:53,160 --> 00:41:55,080
That decision depends on what changed,

1115
00:41:55,080 --> 00:41:56,280
where the path sits,

1116
00:41:56,280 --> 00:41:57,960
what process records exist,

1117
00:41:57,960 --> 00:42:01,240
and whether the earlier work still meets the new product intent.

1118
00:42:01,240 --> 00:42:03,240
A revision change can alter the routing itself,

1119
00:42:03,240 --> 00:42:05,800
maybe the revised part needs an extra inspection step.

1120
00:42:05,800 --> 00:42:08,040
Maybe it now requires a different heat treatment profile,

1121
00:42:08,040 --> 00:42:10,520
a new coating, or a different fixture at assembly.

1122
00:42:10,520 --> 00:42:12,120
The bill of materials may change too,

1123
00:42:12,120 --> 00:42:14,360
which means a component already issued to the order

1124
00:42:14,360 --> 00:42:16,360
no longer fits the release definition.

1125
00:42:16,360 --> 00:42:19,560
That creates new dependencies while the original dependencies still exist.

1126
00:42:19,560 --> 00:42:22,840
Engineering may need to confirm whether completed work can continue.

1127
00:42:22,840 --> 00:42:25,320
Quality may need to approve a deviation.

1128
00:42:25,320 --> 00:42:27,160
Production may need new work instructions

1129
00:42:27,160 --> 00:42:29,000
before the next operation starts.

1130
00:42:29,000 --> 00:42:30,600
Tooling may need adjustment.

1131
00:42:30,600 --> 00:42:32,520
If the change affects a controlled program,

1132
00:42:32,520 --> 00:42:34,840
the machine cannot simply run the next part,

1133
00:42:34,840 --> 00:42:36,600
because someone sent an updated file.

1134
00:42:36,600 --> 00:42:38,200
The order needs a clear release path.

1135
00:42:38,200 --> 00:42:40,120
There are usually a few possible choices

1136
00:42:40,120 --> 00:42:41,560
and none should happen by accident.

1137
00:42:41,560 --> 00:42:44,520
The plant may scrap work that no longer meets the released revision.

1138
00:42:44,520 --> 00:42:47,240
It may rework the part to meet the new requirement.

1139
00:42:47,240 --> 00:42:49,240
It may continue under the earlier revision

1140
00:42:49,240 --> 00:42:50,840
through a controlled deviation.

1141
00:42:50,840 --> 00:42:53,960
If the customer, quality system, and product rules allow it,

1142
00:42:53,960 --> 00:42:55,880
each choice creates a different schedule.

1143
00:42:55,880 --> 00:42:58,760
Scrap creates replacement work and new material demand.

1144
00:42:58,760 --> 00:43:00,680
Re-work creates added operations,

1145
00:43:00,680 --> 00:43:03,720
often on the same constrained resources already under pressure.

1146
00:43:03,720 --> 00:43:06,360
Control-continuation may protect the immediate flow,

1147
00:43:06,360 --> 00:43:08,600
but it needs traceable approval and a clear boundary

1148
00:43:08,600 --> 00:43:10,520
around which units can proceed.

1149
00:43:10,520 --> 00:43:13,560
A planner can't treat those parts as a simple data adjustment.

1150
00:43:13,560 --> 00:43:16,200
The planner needs to know which exact serial numbers,

1151
00:43:16,200 --> 00:43:19,720
lots or quantities reached which operation before the revision changed.

1152
00:43:19,720 --> 00:43:22,200
They need to know what materials those units consumed,

1153
00:43:22,200 --> 00:43:23,720
what instructions govern the work,

1154
00:43:23,720 --> 00:43:25,880
and what evidence proves the state of each unit.

1155
00:43:25,880 --> 00:43:28,680
Without that link, the schedule can claim an order will finish,

1156
00:43:28,680 --> 00:43:32,200
while nobody can confirm whether the finished product matches the release definition.

1157
00:43:32,200 --> 00:43:34,120
That is a product process resource problem.

1158
00:43:34,120 --> 00:43:36,520
The product definition tells you what should exist.

1159
00:43:36,520 --> 00:43:39,000
The process definition tells you how to produce it.

1160
00:43:39,000 --> 00:43:42,280
Resources tell you where and under what conditions the work can happen.

1161
00:43:42,280 --> 00:43:44,360
When engineering changes one part of that chain,

1162
00:43:44,360 --> 00:43:47,800
the system needs to identify the work orders, operations,

1163
00:43:47,800 --> 00:43:49,560
tooling, material lots,

1164
00:43:49,560 --> 00:43:51,320
and instructions touched by the change.

1165
00:43:51,320 --> 00:43:54,280
Otherwise, people reconstruct the impact manually.

1166
00:43:54,280 --> 00:43:55,960
Someone exports the open orders,

1167
00:43:55,960 --> 00:43:58,840
someone else checks paper travelers or MES history.

1168
00:43:58,840 --> 00:44:03,320
A supervisor walks to the work centre and asks which pallets already went through the first operation.

1169
00:44:03,320 --> 00:44:05,720
Engineering checks whether the toolpath changed,

1170
00:44:05,720 --> 00:44:08,200
quality searches for the last approved inspection plan,

1171
00:44:08,200 --> 00:44:10,920
that work can be necessary, especially in a controlled environment.

1172
00:44:10,920 --> 00:44:13,800
But if every change begins with a scavenger hunt across systems,

1173
00:44:13,800 --> 00:44:16,920
the plant loses time before it can even choose a safe response.

1174
00:44:16,920 --> 00:44:20,120
The dependency model should carry revision identity through the order.

1175
00:44:20,120 --> 00:44:23,080
That means an operation isn't only machining complete.

1176
00:44:23,080 --> 00:44:25,480
It is machining complete for this product revision,

1177
00:44:25,480 --> 00:44:28,360
using this process revision with this approved instruction

1178
00:44:28,360 --> 00:44:30,600
and where needed this tooling or program version.

1179
00:44:30,600 --> 00:44:33,640
The level of detail depends on the product and risk.

1180
00:44:33,640 --> 00:44:36,520
You don't need to model every screw as a separate drama,

1181
00:44:36,520 --> 00:44:39,480
but you do need enough traceability to answer our practical question.

1182
00:44:39,480 --> 00:44:43,720
Which work can continue and which work must stop until someone approves the next step?

1183
00:44:43,720 --> 00:44:46,280
Approval itself becomes a schedule dependency.

1184
00:44:46,280 --> 00:44:48,840
A revised drawing might exist in the engineering system,

1185
00:44:48,840 --> 00:44:51,560
while the released manufacturing instruction still waits for review.

1186
00:44:51,560 --> 00:44:53,240
The technical change may be correct,

1187
00:44:53,240 --> 00:44:56,520
but production cannot act on an unreleased or unclear instruction.

1188
00:44:56,520 --> 00:44:59,480
In a live critical path, document release and approval time

1189
00:44:59,480 --> 00:45:02,040
need the same treatment as any other blocking condition.

1190
00:45:02,040 --> 00:45:03,400
They consume elapsed time.

1191
00:45:03,400 --> 00:45:07,160
This can feel frustrating when an order is close to its due date.

1192
00:45:07,160 --> 00:45:08,920
People want to keep material moving.

1193
00:45:08,920 --> 00:45:11,720
Yet pushing work forward without clear revision control

1194
00:45:11,720 --> 00:45:13,960
can create a far larger problem later,

1195
00:45:13,960 --> 00:45:16,760
especially when traceability, customer requirements,

1196
00:45:16,760 --> 00:45:18,600
or safety related product rules apply.

1197
00:45:18,600 --> 00:45:21,720
Speed isn't the only goal, correct flow matters.

1198
00:45:21,720 --> 00:45:24,120
Once the revised order reaches its final operations,

1199
00:45:24,120 --> 00:45:27,720
the remaining path no longer ends just because the product leaves the last machine.

1200
00:45:27,720 --> 00:45:31,480
It ends when the product can meet the customer promise under the right revision

1201
00:45:31,480 --> 00:45:33,880
with the right records and release state.

1202
00:45:33,880 --> 00:45:36,760
Delivery dates depend on more than the last operation.

1203
00:45:36,760 --> 00:45:38,840
Here's the real catch most people miss.

1204
00:45:38,840 --> 00:45:42,120
A product that's finished production isn't necessarily ready to ship.

1205
00:45:42,120 --> 00:45:45,320
Picture an order that clears final assembly on Thursday afternoon.

1206
00:45:45,320 --> 00:45:46,760
Production marks it complete,

1207
00:45:46,760 --> 00:45:49,000
and it looks done from a manufacturing standpoint,

1208
00:45:49,000 --> 00:45:51,480
but packing still needs the right protective material

1209
00:45:51,480 --> 00:45:54,760
to labels and customer specific documents before dispatch can release it.

1210
00:45:54,760 --> 00:45:58,760
Finished goods and shipable goods are two different states entirely.

1211
00:45:58,760 --> 00:46:01,560
For a simple local delivery, that gap might be small,

1212
00:46:01,560 --> 00:46:03,560
but if you're dealing with a regulated product

1213
00:46:03,560 --> 00:46:06,920
in export order or a customer with strict documentation rules,

1214
00:46:06,920 --> 00:46:09,080
that gap controls the delivery date,

1215
00:46:09,080 --> 00:46:11,480
even when every production operation finished early.

1216
00:46:11,480 --> 00:46:14,840
Think about a unit that needs a final test certificate,

1217
00:46:14,840 --> 00:46:17,640
a material certificate, and a signed release record.

1218
00:46:17,640 --> 00:46:20,760
The physical item sits in the finished goods area ready to leave,

1219
00:46:20,760 --> 00:46:24,120
yet one missing document can hold the shipment because the customer,

1220
00:46:24,120 --> 00:46:28,440
carrier, or border authority expects those records to travel with the product.

1221
00:46:28,440 --> 00:46:31,800
So the product exists, but the business commitment still isn't met.

1222
00:46:31,800 --> 00:46:34,360
Packing has its own capacity and timing constraints.

1223
00:46:34,360 --> 00:46:36,680
A plant might have limited space for large items,

1224
00:46:36,680 --> 00:46:38,120
a specialist packing team,

1225
00:46:38,120 --> 00:46:40,440
or packaging that must match customer specs.

1226
00:46:40,440 --> 00:46:42,600
If that team only works a certain shift,

1227
00:46:42,600 --> 00:46:45,640
a late final inspection pushes packing into the next day,

1228
00:46:45,640 --> 00:46:47,800
even if the inspection delay itself was just minutes.

1229
00:46:48,520 --> 00:46:50,680
That elapsed time needs a place in the schedule.

1230
00:46:50,680 --> 00:46:52,520
Carrier cutoffs add another hard edge.

1231
00:46:52,520 --> 00:46:55,080
This patch may need freight ready by a fixed collection time,

1232
00:46:55,080 --> 00:46:58,280
missed that window, and the shipment waits for the next carrier run.

1233
00:46:58,280 --> 00:46:59,560
For some destinations,

1234
00:46:59,560 --> 00:47:01,560
dispatch only happens on selected days,

1235
00:47:01,560 --> 00:47:03,960
and export checks or booking rules need completion

1236
00:47:03,960 --> 00:47:06,280
before the product can even enter the dispatch queue.

1237
00:47:06,280 --> 00:47:09,480
A plan that ends at machine completion stops too early,

1238
00:47:09,480 --> 00:47:11,240
the critical path should run through the event

1239
00:47:11,240 --> 00:47:13,240
that fulfills the actual promise.

1240
00:47:13,240 --> 00:47:15,880
Goods packed and released, loaded onto a named carrier,

1241
00:47:15,880 --> 00:47:19,560
or a customer acceptance step, especially if the contract uses acceptance

1242
00:47:19,560 --> 00:47:21,400
to trigger delivery or payment.

1243
00:47:21,400 --> 00:47:23,320
You need to define the finish line clearly.

1244
00:47:23,320 --> 00:47:24,520
Why does this matter?

1245
00:47:24,520 --> 00:47:27,400
Because different teams use the word complete in different ways.

1246
00:47:27,400 --> 00:47:30,280
Production means the final operation recorded good quantity.

1247
00:47:30,280 --> 00:47:32,200
Quality means the item passed inspection.

1248
00:47:32,200 --> 00:47:34,120
Logistics means it sits in a dispatch lane.

1249
00:47:34,120 --> 00:47:36,760
Customer service means the carrier confirmed collection.

1250
00:47:36,760 --> 00:47:38,280
All of those states are useful,

1251
00:47:38,280 --> 00:47:39,800
but they aren't interchangeable.

1252
00:47:39,800 --> 00:47:42,760
Consider a customer order that allows partial shipment.

1253
00:47:42,760 --> 00:47:44,360
If half the quantity is ready,

1254
00:47:44,360 --> 00:47:46,600
packing and dispatch can protect part of the promise

1255
00:47:46,600 --> 00:47:47,960
while the rest follows later.

1256
00:47:47,960 --> 00:47:50,440
But that depends on commercial terms, freight cost,

1257
00:47:50,440 --> 00:47:51,640
customer preference,

1258
00:47:51,640 --> 00:47:53,640
and whether a partial delivery creates a problem

1259
00:47:53,640 --> 00:47:54,920
at the customer's assembly line.

1260
00:47:54,920 --> 00:47:58,120
Partial shipment isn't just a logistics choice,

1261
00:47:58,120 --> 00:48:00,040
it changes the dependency network.

1262
00:48:00,040 --> 00:48:02,920
The first available quantity follows one path to dispatch

1263
00:48:02,920 --> 00:48:05,160
while the remaining quantity stays on a separate path

1264
00:48:05,160 --> 00:48:07,160
through production and release.

1265
00:48:07,160 --> 00:48:10,200
Planning needs to know whether the customer accepts that split,

1266
00:48:10,200 --> 00:48:12,520
because an internally sensible recovery plan

1267
00:48:12,520 --> 00:48:14,520
can still fail the actual commitment.

1268
00:48:14,520 --> 00:48:18,520
Final inspection often becomes the pressure point near a due date.

1269
00:48:18,520 --> 00:48:19,880
The physical work finishes,

1270
00:48:19,880 --> 00:48:22,120
but the order can't pack, load, or ship

1271
00:48:22,120 --> 00:48:23,960
until someone releases the result.

1272
00:48:23,960 --> 00:48:26,120
If the inspection resource has a queue,

1273
00:48:26,120 --> 00:48:28,360
that queue can control customer delivery

1274
00:48:28,360 --> 00:48:29,800
more than the production operation

1275
00:48:29,800 --> 00:48:32,040
that consumed most of the labor and machine time.

1276
00:48:32,040 --> 00:48:34,760
A short gait can control a long route.

1277
00:48:34,760 --> 00:48:36,360
There are practical details too.

1278
00:48:36,360 --> 00:48:38,360
Labels may need a specific customer format,

1279
00:48:38,360 --> 00:48:40,600
export classification may need verification.

1280
00:48:40,600 --> 00:48:43,480
Dangerous goods rules change packing methods and carrier choices.

1281
00:48:43,480 --> 00:48:46,120
A carrier booking can expire if they're ready-date shifts.

1282
00:48:46,120 --> 00:48:48,040
None of this belongs in the machine routing,

1283
00:48:48,040 --> 00:48:50,840
but all of it affects whether the order meets the promise.

1284
00:48:50,840 --> 00:48:52,680
Those dependencies need a place in the plan

1285
00:48:52,680 --> 00:48:55,320
that doesn't mean a production planner has to personally manage

1286
00:48:55,320 --> 00:48:57,240
every freight document or packing task.

1287
00:48:57,240 --> 00:48:59,720
It means the order's delivery model needs to connect production

1288
00:48:59,720 --> 00:49:02,040
completion to the conditions that turn a completed product

1289
00:49:02,040 --> 00:49:03,560
into a fulfilled commitment.

1290
00:49:03,560 --> 00:49:05,880
Otherwise, the plan celebrates a finished order

1291
00:49:05,880 --> 00:49:08,360
while customer service explains why it didn't ship.

1292
00:49:08,360 --> 00:49:10,200
So when someone asks for the critical pass,

1293
00:49:10,200 --> 00:49:12,200
don't stop at the last physical operation,

1294
00:49:12,200 --> 00:49:14,680
follow the order until the commitment becomes real.

1295
00:49:14,680 --> 00:49:16,120
That's where the due date lives,

1296
00:49:16,120 --> 00:49:18,360
and it's also why another late order dashboard

1297
00:49:18,360 --> 00:49:20,280
rarely gives planners the answer they need.

1298
00:49:20,280 --> 00:49:23,480
Why dashboards don't solve a dependency problem?

1299
00:49:23,480 --> 00:49:26,200
I've seen the scenario play out more times than I can count.

1300
00:49:26,200 --> 00:49:28,920
A late order dashboard shows you which orders need attention,

1301
00:49:28,920 --> 00:49:31,640
downtime, queue size, oe, missed plan,

1302
00:49:31,640 --> 00:49:34,200
and the customer due date sitting uncomfortably close.

1303
00:49:34,200 --> 00:49:36,600
That view helps, but it doesn't tell the planner

1304
00:49:36,600 --> 00:49:38,040
what can still happen next.

1305
00:49:38,040 --> 00:49:40,600
Because a dashboard describes the state it receives

1306
00:49:40,600 --> 00:49:43,960
while a dependency problem needs the system to reason across time,

1307
00:49:43,960 --> 00:49:45,800
rules, and limited capacity.

1308
00:49:45,800 --> 00:49:47,800
Take a familiar morning meeting,

1309
00:49:47,800 --> 00:49:49,720
production sees a machining delay,

1310
00:49:49,720 --> 00:49:51,720
quality sees work waiting for release,

1311
00:49:51,720 --> 00:49:55,320
customer service sees an order that will miss its requested ship date,

1312
00:49:55,320 --> 00:49:57,480
and planning sees a gap in the schedule.

1313
00:49:57,480 --> 00:49:59,160
Each team looks at a correct report,

1314
00:49:59,160 --> 00:50:01,080
yet they draw different conclusions.

1315
00:50:01,080 --> 00:50:02,760
Production asks for more machine time,

1316
00:50:02,760 --> 00:50:04,440
quality asks for a faster review,

1317
00:50:04,440 --> 00:50:07,480
customer service asks whether a partial shipment can leave,

1318
00:50:07,480 --> 00:50:10,360
and planning sees that none of those actions protects the date

1319
00:50:10,360 --> 00:50:12,760
because the real constraint sits later in the flow

1320
00:50:12,760 --> 00:50:14,440
at a shared inspection resource.

1321
00:50:14,440 --> 00:50:17,240
The reports don't disagree because anyone built them badly,

1322
00:50:17,240 --> 00:50:20,280
they disagree because each report answers a local question.

1323
00:50:20,280 --> 00:50:22,200
Power BI can bring those views together.

1324
00:50:22,200 --> 00:50:23,800
It helps people explore trends,

1325
00:50:23,800 --> 00:50:25,480
compare plan against actual,

1326
00:50:25,480 --> 00:50:27,080
and get a common view of the facts.

1327
00:50:27,080 --> 00:50:28,280
If the data model is sound,

1328
00:50:28,280 --> 00:50:30,520
it can even show where delay starts to build,

1329
00:50:30,520 --> 00:50:34,040
but a report doesn't automatically calculate a feasible response.

1330
00:50:34,040 --> 00:50:35,560
Knowing that a machine stopped for an hour

1331
00:50:35,560 --> 00:50:38,440
doesn't tell you which orders lost the next available slot.

1332
00:50:38,440 --> 00:50:40,840
Knowing that an operation finished late doesn't tell you whether

1333
00:50:40,840 --> 00:50:42,760
the downstream resource can still take the work,

1334
00:50:42,760 --> 00:50:44,600
whether the required operator is on shift,

1335
00:50:44,600 --> 00:50:46,920
or whether the order can still catch dispatch.

1336
00:50:46,920 --> 00:50:48,440
Those links need explicit logic.

1337
00:50:48,440 --> 00:50:51,080
This is the difference between a status view and a dependency model.

1338
00:50:51,080 --> 00:50:54,760
A status view tells you an order is late, at risk, or complete.

1339
00:50:54,760 --> 00:50:57,480
A dependency model traces why that state exists,

1340
00:50:57,480 --> 00:50:58,760
what other work it affects,

1341
00:50:58,760 --> 00:51:01,560
and which conditions must change before the date can recover.

1342
00:51:01,560 --> 00:51:03,800
That takes more than joining tables by order number.

1343
00:51:03,800 --> 00:51:07,320
The system needs to understand that an operation requires a particular material state,

1344
00:51:07,320 --> 00:51:09,080
a qualified resource, a time window,

1345
00:51:09,080 --> 00:51:10,360
and maybe a quality release.

1346
00:51:10,360 --> 00:51:13,400
It needs to know that finishing one operation unlocks another,

1347
00:51:13,400 --> 00:51:15,000
only when the right quantity,

1348
00:51:15,000 --> 00:51:17,400
batch rule, or approval condition has been met.

1349
00:51:17,400 --> 00:51:18,600
Without those relationships,

1350
00:51:18,600 --> 00:51:21,320
the dashboard becomes a well-organized list of symptoms.

1351
00:51:21,320 --> 00:51:23,720
This is partly why Excel survives in manufacturing,

1352
00:51:23,720 --> 00:51:25,560
even after a Power BI rollout.

1353
00:51:25,560 --> 00:51:28,040
The spreadsheet isn't always better at storing data.

1354
00:51:28,040 --> 00:51:31,560
It survives because an experienced planner uses it to connect context

1355
00:51:31,560 --> 00:51:34,120
that the enterprise systems don't connect for them.

1356
00:51:34,120 --> 00:51:37,720
They add a note about a tool change a sequence based on a call with the supervisor,

1357
00:51:37,720 --> 00:51:40,280
mark a material lot as not really usable.

1358
00:51:40,280 --> 00:51:42,840
Or remember that a customer accepts a partial shipment,

1359
00:51:42,840 --> 00:51:45,560
but only if the first delivery includes a certain sub-assembly.

1360
00:51:45,560 --> 00:51:49,080
That spreadsheet becomes a working model of the factory.

1361
00:51:49,080 --> 00:51:50,120
That knowledge matters,

1362
00:51:50,120 --> 00:51:52,440
but it often lives in one person's file,

1363
00:51:52,440 --> 00:51:55,080
with manual updates, unclear source dates,

1364
00:51:55,080 --> 00:51:58,040
and assumptions nobody else can inspect when the planner is away.

1365
00:51:58,040 --> 00:52:01,400
A polished report beside that spreadsheet doesn't solve the issue,

1366
00:52:01,400 --> 00:52:03,880
if the report lacks the relationships inside it,

1367
00:52:03,880 --> 00:52:05,880
you haven't removed the manual planning work.

1368
00:52:05,880 --> 00:52:07,560
You've just placed it next to nicer charts.

1369
00:52:07,560 --> 00:52:10,200
A dependency problem also needs time logic,

1370
00:52:10,200 --> 00:52:11,560
not just current state.

1371
00:52:11,560 --> 00:52:12,840
If the furnace queue changes,

1372
00:52:12,840 --> 00:52:15,880
you need to know which orders can reach the next compatible load,

1373
00:52:15,880 --> 00:52:17,000
which ones miss it,

1374
00:52:17,000 --> 00:52:20,120
and where the resulting delay travels through the rest of the schedule.

1375
00:52:20,120 --> 00:52:23,800
That means calculating possible start and finish times against real constraints.

1376
00:52:23,800 --> 00:52:26,200
Reports can display the result of that calculation.

1377
00:52:26,200 --> 00:52:29,480
They can explain a change path in plain language and direct attention

1378
00:52:29,480 --> 00:52:31,560
toward the orders with no slack left.

1379
00:52:31,560 --> 00:52:33,560
But the calculation needs to happen somewhere,

1380
00:52:33,560 --> 00:52:36,520
using defined process rules and trusted execution events.

1381
00:52:36,520 --> 00:52:39,240
A visual layer can't invent missing relationships.

1382
00:52:39,240 --> 00:52:41,800
This matters for IT and OT convergence too.

1383
00:52:41,800 --> 00:52:43,800
You can connect the shop floor to Azure,

1384
00:52:43,800 --> 00:52:45,560
Land Data and Microsoft Fabric,

1385
00:52:45,560 --> 00:52:47,320
and build a very good Power BI report.

1386
00:52:47,320 --> 00:52:50,200
Yet if an equipment event has no link to the affected work order,

1387
00:52:50,200 --> 00:52:53,240
resource condition, remaining duration, and downstream commitments,

1388
00:52:53,240 --> 00:52:56,040
the architecture still can't answer the planner's question.

1389
00:52:56,040 --> 00:52:58,760
It can report the event, it can't explain the impact.

1390
00:52:58,760 --> 00:53:00,520
So before you choose a dashboard layout,

1391
00:53:00,520 --> 00:53:03,400
agree on the information the planning process needs to trust.

1392
00:53:03,400 --> 00:53:06,280
Define the objects in the flow, the relationships between them,

1393
00:53:06,280 --> 00:53:08,520
and the events that change those relationships.

1394
00:53:08,520 --> 00:53:11,160
Only then does reporting become useful for decisions,

1395
00:53:11,160 --> 00:53:14,200
rather than another place to watch orders turn red.

1396
00:53:14,200 --> 00:53:17,000
The data model behind a live critical path.

1397
00:53:17,000 --> 00:53:19,080
Here's the problem most people don't think about.

1398
00:53:19,080 --> 00:53:21,240
You can connect all your data sources,

1399
00:53:21,240 --> 00:53:24,440
but that alone won't calculate a live critical path.

1400
00:53:24,440 --> 00:53:27,000
You need a data model that gives that data meaning,

1401
00:53:27,000 --> 00:53:31,000
because the schedule can only reason about dependencies the model can actually represent.

1402
00:53:31,000 --> 00:53:33,880
So you start with the order, not just the sales order number,

1403
00:53:33,880 --> 00:53:36,760
but the production order, the quantity required,

1404
00:53:36,760 --> 00:53:39,320
the due date, the promised shipment condition,

1405
00:53:39,320 --> 00:53:42,840
and the product revision that applies to that specific work.

1406
00:53:42,840 --> 00:53:44,120
Then you move into operations.

1407
00:53:44,120 --> 00:53:45,960
Each operation needs a plan sequence,

1408
00:53:45,960 --> 00:53:48,360
a remaining duration, a current execution state,

1409
00:53:48,360 --> 00:53:50,120
and rules for what lets it start.

1410
00:53:50,120 --> 00:53:52,040
That might include a predecessor operation,

1411
00:53:52,040 --> 00:53:55,080
a material lot, an approved tool, a qualified resource,

1412
00:53:55,080 --> 00:53:56,280
or a quality release.

1413
00:53:56,280 --> 00:53:58,360
Some operations need all those conditions.

1414
00:53:58,360 --> 00:54:02,440
Others can start when only part of the previous order quantity becomes available.

1415
00:54:02,440 --> 00:54:04,840
The detail should follow the decision you need to make.

1416
00:54:04,840 --> 00:54:06,920
If your problem sits around a bottleneck machine,

1417
00:54:06,920 --> 00:54:09,480
you need to model that machine as a resource with capacity,

1418
00:54:09,480 --> 00:54:11,640
calendars, capabilities, current setup state,

1419
00:54:11,640 --> 00:54:13,080
and planned maintenance periods.

1420
00:54:13,080 --> 00:54:15,400
If an operation needs a fixture or a special tool,

1421
00:54:15,400 --> 00:54:17,560
the model needs to know that the tool exists,

1422
00:54:17,560 --> 00:54:19,800
where it is, and whether it can support the job,

1423
00:54:19,800 --> 00:54:22,440
otherwise you create capacity that nobody can use.

1424
00:54:22,440 --> 00:54:24,040
People belong in the model too.

1425
00:54:24,040 --> 00:54:25,800
But not as a vague labour total.

1426
00:54:25,800 --> 00:54:28,280
The schedule needs the specific skills and authorisations

1427
00:54:28,280 --> 00:54:29,880
that matter for the operation.

1428
00:54:29,880 --> 00:54:32,840
A name person doesn't need to appear in every planning record,

1429
00:54:32,840 --> 00:54:35,960
but a required qualification, approval role, or shift rule,

1430
00:54:35,960 --> 00:54:38,760
often does because it changes the earliest start time.

1431
00:54:38,760 --> 00:54:40,360
Quality states need the same treatment.

1432
00:54:40,360 --> 00:54:43,640
A part can exist physically, while its status blocks the next step.

1433
00:54:43,640 --> 00:54:46,440
Material can sit in stock while the lot remains unreleased.

1434
00:54:46,440 --> 00:54:49,320
An operation can report completed quantity while inspection

1435
00:54:49,320 --> 00:54:51,240
still holds that quantity from transfer.

1436
00:54:51,240 --> 00:54:53,160
These aren't comments attached to an order.

1437
00:54:53,160 --> 00:54:55,800
They are conditions that change what the order can do next.

1438
00:54:55,800 --> 00:54:58,440
Think of the model as a set of business objects

1439
00:54:58,440 --> 00:55:00,120
with explicit links between them.

1440
00:55:00,120 --> 00:55:01,880
An operation follows another operation,

1441
00:55:01,880 --> 00:55:03,400
a work order consumes material,

1442
00:55:03,400 --> 00:55:05,080
a machine qualifies for an operation,

1443
00:55:05,080 --> 00:55:06,840
a tool supports a machine setup,

1444
00:55:06,840 --> 00:55:08,600
a quality hold blocks a transfer,

1445
00:55:08,600 --> 00:55:11,080
a calendar limits when a resource can work.

1446
00:55:11,080 --> 00:55:12,840
A product revision governs the routing

1447
00:55:12,840 --> 00:55:14,680
and the instruction the operator needs.

1448
00:55:14,680 --> 00:55:16,520
Those links need clear verbs.

1449
00:55:16,520 --> 00:55:18,680
Requires tells the system that an operation

1450
00:55:18,680 --> 00:55:20,440
cannot start without something else.

1451
00:55:20,440 --> 00:55:24,200
Producers tells it what quantity becomes available after work completes.

1452
00:55:24,200 --> 00:55:27,080
Consumes connects material to a production step.

1453
00:55:27,080 --> 00:55:30,680
Qualifies states that a resource can run a defined operation.

1454
00:55:30,680 --> 00:55:32,840
Blocks captures a hold or release condition.

1455
00:55:32,840 --> 00:55:35,240
Follows describes the process dependency.

1456
00:55:35,240 --> 00:55:37,160
Without those links you have records.

1457
00:55:37,160 --> 00:55:38,920
With them you can calculate a path.

1458
00:55:38,920 --> 00:55:40,680
Time also needs more than one field.

1459
00:55:40,680 --> 00:55:42,440
Plan start and plan finish still matter

1460
00:55:42,440 --> 00:55:43,960
because they show the original intent.

1461
00:55:43,960 --> 00:55:46,840
Actual start, actual completion, reported quantity,

1462
00:55:46,840 --> 00:55:48,680
and actual downtime tell you what happened.

1463
00:55:48,680 --> 00:55:51,880
Then you need an estimate for remaining time based on the current state.

1464
00:55:51,880 --> 00:55:54,360
Not a standard duration copied from the routing.

1465
00:55:54,360 --> 00:55:56,520
That estimate can change during the shift.

1466
00:55:56,520 --> 00:55:59,240
A partly completed operation may need less time than planned.

1467
00:55:59,240 --> 00:56:02,040
It may need more because scrap reduced the good quantity,

1468
00:56:02,040 --> 00:56:04,440
a tool change took longer or a setup failed.

1469
00:56:04,440 --> 00:56:07,000
The model should preserve the difference between plan,

1470
00:56:07,000 --> 00:56:09,480
actual event, and current forecast.

1471
00:56:09,480 --> 00:56:11,560
If those states merge into one date,

1472
00:56:11,560 --> 00:56:13,400
nobody can explain why the schedule moved.

1473
00:56:13,400 --> 00:56:16,360
Identity becomes a practical issue very quickly.

1474
00:56:16,360 --> 00:56:19,800
Your ERP, MES, maintenance system, quality system,

1475
00:56:19,800 --> 00:56:23,400
and shop floor data source may all refer to the same item,

1476
00:56:23,400 --> 00:56:25,960
machine, or order in different ways.

1477
00:56:25,960 --> 00:56:28,120
One system uses a work order number,

1478
00:56:28,120 --> 00:56:29,960
another uses an operation ticket.

1479
00:56:29,960 --> 00:56:33,240
A machine event may only carry a program name or a resource ID.

1480
00:56:33,240 --> 00:56:35,880
The model needs a reliable way to connect those references.

1481
00:56:35,880 --> 00:56:38,360
That doesn't always mean replacing every local identifier.

1482
00:56:38,360 --> 00:56:40,680
It means defining how they map to a shared identity

1483
00:56:40,680 --> 00:56:41,960
that planning can trust.

1484
00:56:41,960 --> 00:56:45,160
If the same machining center appears under different names in two systems,

1485
00:56:45,160 --> 00:56:46,600
the schedule may count at twice,

1486
00:56:46,600 --> 00:56:49,320
or miss the fact that a disruption affects the cute work.

1487
00:56:49,320 --> 00:56:50,760
Master DataWork isn't glamorous,

1488
00:56:50,760 --> 00:56:52,440
and neither is correcting a resource code

1489
00:56:52,440 --> 00:56:53,720
that changed five years ago.

1490
00:56:53,720 --> 00:56:55,240
But live planning depends on it.

1491
00:56:55,240 --> 00:56:56,920
Versioning matters for the same reason.

1492
00:56:56,920 --> 00:56:59,320
A routing can change, a product revision can change,

1493
00:56:59,320 --> 00:57:00,440
a calendar can change,

1494
00:57:00,440 --> 00:57:01,800
a schedule assumption can change

1495
00:57:01,800 --> 00:57:03,880
after a planner approves an alternate machine

1496
00:57:03,880 --> 00:57:05,320
or a partial transfer batch.

1497
00:57:05,320 --> 00:57:07,880
You need to know which version governed the decision.

1498
00:57:07,880 --> 00:57:10,040
Otherwise, a planner looks back at a missed date

1499
00:57:10,040 --> 00:57:11,960
and sees only that the schedule changed.

1500
00:57:11,960 --> 00:57:14,440
They can't tell whether the change came from a late event,

1501
00:57:14,440 --> 00:57:15,800
a revised process rule,

1502
00:57:15,800 --> 00:57:17,480
or an altered planning assumption.

1503
00:57:17,480 --> 00:57:19,000
That makes learning nearly impossible

1504
00:57:19,000 --> 00:57:21,240
because the history has lost its meaning.

1505
00:57:21,240 --> 00:57:23,960
So the goal isn't a giant model of every fact in the factory.

1506
00:57:23,960 --> 00:57:26,840
Build the minimum model that can explain why work cannot start,

1507
00:57:26,840 --> 00:57:27,880
what it depends on,

1508
00:57:27,880 --> 00:57:29,960
and what changes when one condition moves.

1509
00:57:29,960 --> 00:57:31,320
Once that model exists,

1510
00:57:31,320 --> 00:57:33,640
you can follow events from operations and resources

1511
00:57:33,640 --> 00:57:35,960
into a planning decision without asking the planner

1512
00:57:35,960 --> 00:57:38,360
to rebuild the dependency chain by hand.

1513
00:57:38,360 --> 00:57:40,520
Following the data from OT to the planner,

1514
00:57:40,520 --> 00:57:42,840
the model only works if current events reach it

1515
00:57:42,840 --> 00:57:45,400
with enough meaning to affect a real planning decision,

1516
00:57:45,400 --> 00:57:46,600
start on the shop floor.

1517
00:57:46,600 --> 00:57:49,640
A programmable logic controller, a PLC,

1518
00:57:49,640 --> 00:57:51,720
may know that a machine changed state,

1519
00:57:51,720 --> 00:57:54,600
finished a cycle, raised an alarm, or stopped for a period.

1520
00:57:54,600 --> 00:57:56,600
IoT connections can collect those signals

1521
00:57:56,600 --> 00:57:58,920
and move them into an industrial data service.

1522
00:57:58,920 --> 00:58:01,000
That data matters, but raw signals alone

1523
00:58:01,000 --> 00:58:02,360
don't schedule production.

1524
00:58:02,360 --> 00:58:05,080
A state change from running to idle could mean a completed part,

1525
00:58:05,080 --> 00:58:06,120
a missing operator,

1526
00:58:06,120 --> 00:58:07,000
a tool issue,

1527
00:58:07,000 --> 00:58:09,720
a material shortage, or an intentional pause.

1528
00:58:09,720 --> 00:58:12,520
If every state change triggers a schedule update,

1529
00:58:12,520 --> 00:58:15,560
the planning system will spend the day reacting to noise.

1530
00:58:15,560 --> 00:58:17,480
Nobody needs a critical path calculation

1531
00:58:17,480 --> 00:58:19,880
that panics every time someone opens a machine door,

1532
00:58:19,880 --> 00:58:22,040
the event needs an operational meaning first.

1533
00:58:22,040 --> 00:58:24,280
This is where the MES sits close to the work.

1534
00:58:24,280 --> 00:58:26,280
The manufacturing execution system

1535
00:58:26,280 --> 00:58:28,520
records the context around execution,

1536
00:58:28,520 --> 00:58:29,960
which work order ran,

1537
00:58:29,960 --> 00:58:31,480
which operation started,

1538
00:58:31,480 --> 00:58:32,760
what quantity completed,

1539
00:58:32,760 --> 00:58:33,880
what quantity failed,

1540
00:58:33,880 --> 00:58:35,960
and whether the order entered a hold state.

1541
00:58:35,960 --> 00:58:39,160
That turns equipment activity into a production event.

1542
00:58:39,160 --> 00:58:41,320
For example, an MES record may confirm

1543
00:58:41,320 --> 00:58:43,800
that operation 30 started on a specific resource

1544
00:58:43,800 --> 00:58:47,160
at a defined time against a known work order and revision.

1545
00:58:47,160 --> 00:58:50,360
Later, it may report that part of the required quantity completed,

1546
00:58:50,360 --> 00:58:53,160
while some units moved to scrap or await inspection

1547
00:58:53,160 --> 00:58:54,680
a planner can work with that.

1548
00:58:54,680 --> 00:58:57,480
A machine cycle counter cannot answer the same question by itself.

1549
00:58:57,480 --> 00:58:59,720
The ERP contributes a different part of the story.

1550
00:58:59,720 --> 00:59:00,920
It carries demand,

1551
00:59:00,920 --> 00:59:02,120
customer promise dates,

1552
00:59:02,120 --> 00:59:03,160
order priority,

1553
00:59:03,160 --> 00:59:04,360
material supply,

1554
00:59:04,360 --> 00:59:05,480
purchasing commitments,

1555
00:59:05,480 --> 00:59:06,520
and commercial rules

1556
00:59:06,520 --> 00:59:08,680
that shape what delivery actually means.

1557
00:59:08,680 --> 00:59:10,360
It may also hold the planned routing

1558
00:59:10,360 --> 00:59:11,960
and the original production dates.

1559
00:59:11,960 --> 00:59:13,800
Those records don't replace execution data.

1560
00:59:13,800 --> 00:59:15,720
They provide the wider business condition

1561
00:59:15,720 --> 00:59:17,320
around the execution data.

1562
00:59:17,320 --> 00:59:19,320
A late operation means one thing for an order

1563
00:59:19,320 --> 00:59:20,440
with a flexible internal date

1564
00:59:20,440 --> 00:59:21,720
and something else for an order tied

1565
00:59:21,720 --> 00:59:23,480
to a committed customer shipment.

1566
00:59:23,480 --> 00:59:25,240
Planning needs both views at once.

1567
00:59:25,240 --> 00:59:27,160
So think about the flow as a chain of decisions,

1568
00:59:27,160 --> 00:59:28,840
not a chain of databases.

1569
00:59:28,840 --> 00:59:31,160
Equipment and IoT data report physical events,

1570
00:59:31,160 --> 00:59:33,800
the MES connects those events to production execution.

1571
00:59:33,800 --> 00:59:35,400
The ERP connects production work

1572
00:59:35,400 --> 00:59:37,720
to demand, supply, and customer commitments.

1573
00:59:37,720 --> 00:59:40,280
A shared data layer can then bring those facts together,

1574
00:59:40,280 --> 00:59:41,880
without pretending that every system

1575
00:59:41,880 --> 00:59:43,880
should become the master of every fact.

1576
00:59:43,880 --> 00:59:45,640
As your can support the integration

1577
00:59:45,640 --> 00:59:47,880
and event handling part of this architecture,

1578
00:59:47,880 --> 00:59:50,360
Microsoft Fabric can provide shared access

1579
00:59:50,360 --> 00:59:52,520
to governed operational and business data

1580
00:59:52,520 --> 00:59:55,800
for analysis, planning support, and reporting.

1581
00:59:55,800 --> 00:59:57,720
The exact services depend on your plant,

1582
00:59:57,720 --> 01:00:00,520
security model, existing MES, and data volumes.

1583
01:00:00,520 --> 01:00:02,920
The point isn't to move every signal into one place,

1584
01:00:02,920 --> 01:00:04,440
the point is to move the events

1585
01:00:04,440 --> 01:00:06,040
that can change a decision

1586
01:00:06,040 --> 01:00:08,120
with their source, time, identity,

1587
01:00:08,120 --> 01:00:09,480
and business meaning intact.

1588
01:00:09,480 --> 01:00:11,000
A machine alarm may matter only

1589
01:00:11,000 --> 01:00:13,080
when it affects an active or cute operation.

1590
01:00:13,080 --> 01:00:14,680
A completed quantity may matter only

1591
01:00:14,680 --> 01:00:17,240
when it meets the transfer rule for the next process,

1592
01:00:17,240 --> 01:00:19,560
filtering belongs close to the source where possible.

1593
01:00:19,560 --> 01:00:21,640
You don't want planners trying to interpret

1594
01:00:21,640 --> 01:00:22,840
thousands of controller tags

1595
01:00:22,840 --> 01:00:24,200
and you don't want a cloud platform

1596
01:00:24,200 --> 01:00:25,400
reaching into a control loop

1597
01:00:25,400 --> 01:00:27,080
to tell a machine how to run.

1598
01:00:27,080 --> 01:00:28,920
OT and IT have different jobs.

1599
01:00:28,920 --> 01:00:31,640
OT protects safe, stable production.

1600
01:00:31,640 --> 01:00:34,280
IT helps connect the dots between IT and OT

1601
01:00:34,280 --> 01:00:35,960
so planning can use trusted facts

1602
01:00:35,960 --> 01:00:37,400
without disrupting the process.

1603
01:00:37,400 --> 01:00:39,080
That boundary needs respect.

1604
01:00:39,080 --> 01:00:40,840
A useful event flow might begin

1605
01:00:40,840 --> 01:00:43,400
when the machine operator reports an execution change.

1606
01:00:43,400 --> 01:00:45,480
The MES validates the order context

1607
01:00:45,480 --> 01:00:47,000
and records the result.

1608
01:00:47,000 --> 01:00:49,400
An integration layer publishes the relevant event.

1609
01:00:49,400 --> 01:00:51,800
The planning model then checks whether that event changes

1610
01:00:51,800 --> 01:00:54,360
remaining duration, resource availability,

1611
01:00:54,360 --> 01:00:57,320
quantity ready for transfer, or a release condition.

1612
01:00:57,320 --> 01:01:00,360
Only after that does the critical path need recalculation.

1613
01:01:00,360 --> 01:01:03,000
That sequence also makes data just easier to manage.

1614
01:01:03,000 --> 01:01:04,760
When a planner sees a changed date,

1615
01:01:04,760 --> 01:01:07,080
they should be able to ask where the input came from.

1616
01:01:07,080 --> 01:01:08,520
Was it a maintenance estimate?

1617
01:01:08,520 --> 01:01:10,280
An MES completion record,

1618
01:01:10,280 --> 01:01:12,040
a supplier update from ERP,

1619
01:01:12,040 --> 01:01:13,560
or a manual planning assumption?

1620
01:01:13,560 --> 01:01:15,640
Different sources carry different authority.

1621
01:01:15,640 --> 01:01:17,800
A sensor can report that a spindle stopped.

1622
01:01:17,800 --> 01:01:19,480
Maintenance owns the repair estimate.

1623
01:01:19,480 --> 01:01:21,320
The MES owns the execution record.

1624
01:01:21,320 --> 01:01:22,760
Quality owns the release state.

1625
01:01:22,760 --> 01:01:25,080
Planning owns the approved response in the schedule.

1626
01:01:25,080 --> 01:01:27,720
If all those facts arrive without ownership and timestamps,

1627
01:01:27,720 --> 01:01:29,800
the planner gets a confident looking answer

1628
01:01:29,800 --> 01:01:31,000
with no way to challenge it.

1629
01:01:31,000 --> 01:01:32,200
That isn't decision support.

1630
01:01:32,200 --> 01:01:33,720
It's just faster confusion.

1631
01:01:33,720 --> 01:01:35,560
Real-time visibility becomes useful

1632
01:01:35,560 --> 01:01:37,720
when it gives the planner events they can trust.

1633
01:01:37,720 --> 01:01:39,400
At a level they can act on.

1634
01:01:39,400 --> 01:01:41,080
Not every signal deserves a replay,

1635
01:01:41,080 --> 01:01:42,440
but a confirmed machine loss,

1636
01:01:42,440 --> 01:01:44,280
a missed quantity, a material release,

1637
01:01:44,280 --> 01:01:46,760
or an inspection hold can change the earliest possible finish

1638
01:01:46,760 --> 01:01:47,560
for real work.

1639
01:01:47,560 --> 01:01:50,280
Those are the events worth carrying through the architecture.

1640
01:01:50,280 --> 01:01:52,680
Even then, shared data doesn't automatically understand

1641
01:01:52,680 --> 01:01:54,520
why one event affects one order,

1642
01:01:54,520 --> 01:01:56,040
several orders, or none at all.

1643
01:01:56,040 --> 01:01:57,240
Raw facts need a structure

1644
01:01:57,240 --> 01:01:59,480
that can follow relationships across the factory.

1645
01:01:59,480 --> 01:02:01,640
Digital twin and knowledge graph.

1646
01:02:01,640 --> 01:02:02,760
Different jobs.

1647
01:02:02,760 --> 01:02:05,240
Once those events land in a shared data layer,

1648
01:02:05,240 --> 01:02:07,480
people start throwing around two terms.

1649
01:02:07,480 --> 01:02:09,160
Digital twin and knowledge graph.

1650
01:02:09,160 --> 01:02:11,320
You see them used like they're the same thing, they're not.

1651
01:02:11,320 --> 01:02:15,480
A digital twin gives you a structured representation

1652
01:02:15,480 --> 01:02:17,160
of something physical or operational

1653
01:02:17,160 --> 01:02:19,400
like a machine, a production line,

1654
01:02:19,400 --> 01:02:22,120
a furnace, a tool, or a step in the process.

1655
01:02:22,120 --> 01:02:23,720
It captures what the thing is,

1656
01:02:23,720 --> 01:02:24,920
what state it's in,

1657
01:02:24,920 --> 01:02:26,920
and often what data belongs to it.

1658
01:02:26,920 --> 01:02:28,360
Take a machining center.

1659
01:02:28,360 --> 01:02:31,080
Its digital twin might include the resource identity,

1660
01:02:31,080 --> 01:02:33,080
current state, approve capabilities,

1661
01:02:33,080 --> 01:02:34,680
maintenance condition, active program,

1662
01:02:34,680 --> 01:02:37,800
install tool set, and the work order currently assigned.

1663
01:02:37,800 --> 01:02:39,880
So now the asset has a usable digital identity

1664
01:02:39,880 --> 01:02:43,080
that helps, but it still doesn't answer the wider planning question.

1665
01:02:43,080 --> 01:02:45,080
A knowledge graph jumps into the relationships.

1666
01:02:45,080 --> 01:02:47,800
It records how orders, products, operations, resources,

1667
01:02:47,800 --> 01:02:49,640
people, tools, materials, quality rules,

1668
01:02:49,640 --> 01:02:51,560
and customer commitments all connect.

1669
01:02:51,560 --> 01:02:54,280
So instead of just asking by what's the state of this machine,

1670
01:02:54,280 --> 01:02:57,160
it can ask which work orders depend on this machine,

1671
01:02:57,160 --> 01:02:58,360
which operations follow,

1672
01:02:58,360 --> 01:03:01,560
and which delivery commitments sit behind those operations.

1673
01:03:01,560 --> 01:03:03,000
That's a different job entirely.

1674
01:03:03,000 --> 01:03:04,600
Take the same machine outage from earlier,

1675
01:03:04,600 --> 01:03:07,400
a digital twin can describe the machine's current state.

1676
01:03:07,400 --> 01:03:10,600
Unavailable, under repair, no confirmed return time.

1677
01:03:10,600 --> 01:03:12,840
A knowledge graph traces outward from that state,

1678
01:03:12,840 --> 01:03:15,160
connecting the machine to eligible operations,

1679
01:03:15,160 --> 01:03:16,200
then to work orders,

1680
01:03:16,200 --> 01:03:18,440
then to material and quality conditions,

1681
01:03:18,440 --> 01:03:20,360
and finally to shipment commitments.

1682
01:03:20,360 --> 01:03:21,880
One tells you what the asset is doing.

1683
01:03:21,880 --> 01:03:24,040
The other explains who and what depends on it.

1684
01:03:24,040 --> 01:03:25,240
You can use both together.

1685
01:03:25,240 --> 01:03:28,040
The digital twin gives current state a structured home,

1686
01:03:28,040 --> 01:03:29,640
while the knowledge graph puts that state

1687
01:03:29,640 --> 01:03:31,320
into a wider network of meaning.

1688
01:03:31,320 --> 01:03:33,880
When the machine changes from available to unavailable,

1689
01:03:33,880 --> 01:03:35,480
the graph can follow the relationships

1690
01:03:35,480 --> 01:03:37,560
that turn a local event into a planning problem.

1691
01:03:37,560 --> 01:03:39,400
But neither one replaces your MES.

1692
01:03:39,400 --> 01:03:42,680
Your MES still records and controls production execution.

1693
01:03:42,680 --> 01:03:44,680
It knows which operation started,

1694
01:03:44,680 --> 01:03:46,680
what quantity passed, what quantity failed,

1695
01:03:46,680 --> 01:03:48,200
and where the work sits.

1696
01:03:48,200 --> 01:03:51,240
Your ERP still carries demand, supply, commercial rules,

1697
01:03:51,240 --> 01:03:52,680
and planned order data.

1698
01:03:52,680 --> 01:03:54,360
A scheduling engine still needs to calculate

1699
01:03:54,360 --> 01:03:56,360
feasible sequences against real constraints.

1700
01:03:56,360 --> 01:03:58,520
A twin or graph doesn't magically schedule the plant

1701
01:03:58,520 --> 01:04:00,360
just because it has a clever name.

1702
01:04:00,360 --> 01:04:01,880
What it does provide is the structure

1703
01:04:01,880 --> 01:04:04,680
that lets those systems exchange and interpret information

1704
01:04:04,680 --> 01:04:07,640
without reducing the factory to disconnected status fields.

1705
01:04:07,640 --> 01:04:10,280
That's where it becomes useful for a live critical path.

1706
01:04:10,280 --> 01:04:13,480
At the center of this sits the product process resource relationship.

1707
01:04:13,480 --> 01:04:15,320
The product tells you what you're building.

1708
01:04:15,320 --> 01:04:18,440
The process tells you which operations and conditions apply.

1709
01:04:18,440 --> 01:04:20,360
The resource tells you where the work can happen,

1710
01:04:20,360 --> 01:04:23,560
with what capability, tooling, skills, and release rules.

1711
01:04:23,560 --> 01:04:25,080
When those links stay separate,

1712
01:04:25,080 --> 01:04:27,560
an engineering revision can appear in one system,

1713
01:04:27,560 --> 01:04:30,360
while planning still follows an older route in another.

1714
01:04:30,360 --> 01:04:32,360
A resource can show free capacity

1715
01:04:32,360 --> 01:04:34,920
while a quality rule blocks the work it needs to run.

1716
01:04:34,920 --> 01:04:37,160
A machine event can arrive with no clear path

1717
01:04:37,160 --> 01:04:38,920
to the customer order it affects.

1718
01:04:38,920 --> 01:04:40,600
The relationship model joins those facts

1719
01:04:40,600 --> 01:04:42,360
without pretending they are identical.

1720
01:04:42,360 --> 01:04:43,640
Here's a concrete example.

1721
01:04:43,640 --> 01:04:46,520
A product revision may require operation 40,

1722
01:04:46,520 --> 01:04:49,560
which needs a specific fixture and a qualified test cell.

1723
01:04:49,560 --> 01:04:52,280
That test cell depends on calibration status

1724
01:04:52,280 --> 01:04:53,880
and an operator approval role.

1725
01:04:53,880 --> 01:04:56,440
And the completed unit then requires a quality release

1726
01:04:56,440 --> 01:04:57,800
before it enters packing.

1727
01:04:57,800 --> 01:05:00,200
Those aren't just data fields on one large record.

1728
01:05:00,200 --> 01:05:01,400
They're linked conditions

1729
01:05:01,400 --> 01:05:04,360
and the links explain why the work can or can't move.

1730
01:05:04,360 --> 01:05:06,760
I would start much smaller than the usual vision

1731
01:05:06,760 --> 01:05:08,760
of a full digital copy of the plant.

1732
01:05:08,760 --> 01:05:10,840
That vision sounds impressive until someone asks

1733
01:05:10,840 --> 01:05:12,600
who will keep every relationship current,

1734
01:05:12,600 --> 01:05:13,720
including the awkward ones

1735
01:05:13,720 --> 01:05:15,640
that only appear during real production.

1736
01:05:15,640 --> 01:05:17,560
Start with one constrained value stream.

1737
01:05:17,560 --> 01:05:19,560
Choose a flow where late delivery keeps returning

1738
01:05:19,560 --> 01:05:21,400
to the same kind of dependency problem.

1739
01:05:21,400 --> 01:05:23,320
Maybe a bottleneck machining cell,

1740
01:05:23,320 --> 01:05:24,280
a batch furnace,

1741
01:05:24,280 --> 01:05:27,160
a control test process, or a final release gate.

1742
01:05:27,160 --> 01:05:29,480
Model the orders, operations, resources,

1743
01:05:29,480 --> 01:05:32,280
and release conditions that change decisions in that flow,

1744
01:05:32,280 --> 01:05:34,840
then test whether the model can answer a practical question

1745
01:05:34,840 --> 01:05:36,280
a planner asks every day.

1746
01:05:36,280 --> 01:05:38,760
If it can't explain which dependency blocks an order,

1747
01:05:38,760 --> 01:05:40,600
it isn't ready for a larger scope.

1748
01:05:40,600 --> 01:05:42,920
But if it can trace a machine loss or quality hold

1749
01:05:42,920 --> 01:05:44,200
through the affected work

1750
01:05:44,200 --> 01:05:45,400
and show the conditions around it,

1751
01:05:45,400 --> 01:05:47,240
you have something people can actually use.

1752
01:05:47,240 --> 01:05:48,840
That brings us to the next step.

1753
01:05:48,840 --> 01:05:50,280
Once the relationships exist,

1754
01:05:50,280 --> 01:05:52,200
the system can recalculate the path

1755
01:05:52,200 --> 01:05:54,200
when a real event changes the plan.

1756
01:05:54,200 --> 01:05:56,680
Re-calculating the path when an event changes.

1757
01:05:56,680 --> 01:05:58,040
Once the relationships exist,

1758
01:05:58,040 --> 01:06:00,520
recalculation starts with an event that has planning meaning.

1759
01:06:00,520 --> 01:06:02,600
Maybe a machine stops with a repair estimate,

1760
01:06:02,600 --> 01:06:04,200
a supplier moves a delivery date,

1761
01:06:04,200 --> 01:06:05,800
inspection places a lot on hold,

1762
01:06:05,800 --> 01:06:08,920
or an operation reports less good quantity than the plan expected.

1763
01:06:08,920 --> 01:06:10,600
The event changes an assumption

1764
01:06:10,600 --> 01:06:12,520
and the schedule needs to find every dependency

1765
01:06:12,520 --> 01:06:13,800
that used that assumption.

1766
01:06:13,800 --> 01:06:16,680
Take a machining center that goes down halfway through a shift.

1767
01:06:16,680 --> 01:06:19,880
The first question isn't whether the machine turns red on the status board.

1768
01:06:19,880 --> 01:06:22,440
Planning needs to identify what's running on it now,

1769
01:06:22,440 --> 01:06:25,800
what's waiting for it and which orders depend on work leaving that machine.

1770
01:06:25,800 --> 01:06:27,240
Some orders might absorb the delay,

1771
01:06:27,240 --> 01:06:28,760
but others have no slack left.

1772
01:06:28,760 --> 01:06:30,600
The model traces from the disrupted resource

1773
01:06:30,600 --> 01:06:32,120
to eligible operations,

1774
01:06:32,120 --> 01:06:35,400
then from those operations to work orders and their remaining route.

1775
01:06:35,400 --> 01:06:38,440
It also checks shared links that don't sit in a simple order sequence

1776
01:06:38,440 --> 01:06:40,360
like a fixture, a qualified operator,

1777
01:06:40,360 --> 01:06:43,240
a material lot or a batch window later in the flow.

1778
01:06:43,240 --> 01:06:47,560
That's how one equipment event turns into a set of specific planning impacts.

1779
01:06:47,560 --> 01:06:49,560
Then the system needs new time estimates

1780
01:06:49,560 --> 01:06:52,440
starting with the repair estimate changing the resource calendar.

1781
01:06:52,440 --> 01:06:55,000
The current order may have part of its quantity complete

1782
01:06:55,000 --> 01:06:57,480
and a queued order may move to an alternate machine,

1783
01:06:57,480 --> 01:07:00,040
but only if that alternate has the right capability,

1784
01:07:00,040 --> 01:07:01,640
tooling and time available.

1785
01:07:01,640 --> 01:07:03,800
Nothing moves just because there's an open slot.

1786
01:07:03,800 --> 01:07:05,080
For each affected operation,

1787
01:07:05,080 --> 01:07:08,360
the calculation asks when all start conditions next overlap.

1788
01:07:08,360 --> 01:07:10,280
Whether the predecessor quantity is ready,

1789
01:07:10,280 --> 01:07:11,560
the material is released,

1790
01:07:11,560 --> 01:07:13,880
the assigned resource can perform the operation.

1791
01:07:13,880 --> 01:07:16,280
Its calendar allows enough uninterrupted time

1792
01:07:16,280 --> 01:07:19,000
and whether holds, maintenance windows or shift limits

1793
01:07:19,000 --> 01:07:20,280
change the answer.

1794
01:07:20,280 --> 01:07:23,480
That produces a revised earliest start and earliest finish.

1795
01:07:23,480 --> 01:07:26,600
The system then works forward through the remaining dependencies.

1796
01:07:26,600 --> 01:07:29,720
If operation 20 finishes later, operation 30 may start later.

1797
01:07:29,720 --> 01:07:32,040
If operation 30 misses a compatible furnace load,

1798
01:07:32,040 --> 01:07:34,440
the delay grows far beyond the original outage.

1799
01:07:34,440 --> 01:07:36,280
If final release misses a dispatch cutoff,

1800
01:07:36,280 --> 01:07:38,040
the customer promise moves again.

1801
01:07:38,040 --> 01:07:40,280
A delay doesn't travel at a fixed speed.

1802
01:07:40,280 --> 01:07:41,960
It meets conditions as it moves.

1803
01:07:41,960 --> 01:07:45,560
At the same time, the calculation needs to look at slack or float.

1804
01:07:45,560 --> 01:07:48,520
The time and operation can move without moving the final commitment.

1805
01:07:48,520 --> 01:07:50,280
When an event consumes that slack,

1806
01:07:50,280 --> 01:07:52,840
the operation moves onto the current critical path.

1807
01:07:52,840 --> 01:07:54,040
That word current matters.

1808
01:07:54,040 --> 01:07:56,280
An operation that controls delivery this morning

1809
01:07:56,280 --> 01:07:58,760
may no longer control it after a resource change,

1810
01:07:58,760 --> 01:08:01,320
a material release or a revised sequence.

1811
01:08:01,320 --> 01:08:03,080
Another operation may be one that looked harmless

1812
01:08:03,080 --> 01:08:05,000
in the original plan can become the path

1813
01:08:05,000 --> 01:08:07,480
that now decides the earliest possible ship date.

1814
01:08:07,480 --> 01:08:10,040
The critical path isn't a permanent tag on a routing step.

1815
01:08:10,040 --> 01:08:11,800
It's the chain that controls completion

1816
01:08:11,800 --> 01:08:13,240
under the conditions you have now.

1817
01:08:13,240 --> 01:08:15,640
This is also why planners need an explanation,

1818
01:08:15,640 --> 01:08:16,840
not just a new date.

1819
01:08:16,840 --> 01:08:19,560
If a system tells them an order moved from Thursday to Friday,

1820
01:08:19,560 --> 01:08:21,240
they need to know what changed in the chain.

1821
01:08:21,240 --> 01:08:23,080
Was it the machining outage directly?

1822
01:08:23,080 --> 01:08:24,200
A lost alternate slot?

1823
01:08:24,200 --> 01:08:25,400
A missed furnace batch?

1824
01:08:25,400 --> 01:08:28,200
A quality release that became the final constraint?

1825
01:08:28,200 --> 01:08:30,840
Without that explanation, people don't know where to act.

1826
01:08:30,840 --> 01:08:32,520
A useful result might say,

1827
01:08:32,520 --> 01:08:33,960
this order now finishes later

1828
01:08:33,960 --> 01:08:36,120
because the plan machine became unavailable.

1829
01:08:36,120 --> 01:08:39,240
The approved alternate resource can't begin until the next shift,

1830
01:08:39,240 --> 01:08:40,840
and the resulting completion time

1831
01:08:40,840 --> 01:08:42,520
misses the next heat treatment load.

1832
01:08:42,520 --> 01:08:46,280
The final delivery risk comes from the missed batch window,

1833
01:08:46,280 --> 01:08:48,360
not just from the original machine event.

1834
01:08:48,360 --> 01:08:49,960
That gives the plan a something to test.

1835
01:08:49,960 --> 01:08:52,680
They can ask maintenance for a more reliable return estimate,

1836
01:08:52,680 --> 01:08:55,080
check whether an alternate fixture can move or review

1837
01:08:55,080 --> 01:08:57,080
whether another furnace load is possible.

1838
01:08:57,080 --> 01:08:59,320
Each action changes the stated condition

1839
01:08:59,320 --> 01:09:01,080
and the model can calculate the result

1840
01:09:01,080 --> 01:09:03,640
rather than relying on a hopeful manual adjustment.

1841
01:09:03,640 --> 01:09:06,040
Not every event needs an immediate full replay.

1842
01:09:06,040 --> 01:09:08,200
A short interruption inside available slack

1843
01:09:08,200 --> 01:09:10,040
may change nothing that matters to delivery.

1844
01:09:10,040 --> 01:09:12,440
The model needs to separate local noise from events

1845
01:09:12,440 --> 01:09:13,880
that alter feasible dates

1846
01:09:13,880 --> 01:09:16,200
or consume the remaining buffer on an order.

1847
01:09:16,200 --> 01:09:18,440
Otherwise, planners receive alerts all day

1848
01:09:18,440 --> 01:09:19,640
and learn to ignore them.

1849
01:09:19,640 --> 01:09:21,640
There's also a discipline around uncertainty.

1850
01:09:21,640 --> 01:09:23,400
A repair estimate may be tentative,

1851
01:09:23,400 --> 01:09:25,640
a supplier promise may lack confirmation.

1852
01:09:25,640 --> 01:09:27,880
In those cases, the calculation should retain

1853
01:09:27,880 --> 01:09:29,960
the source and confidence of the assumption

1854
01:09:29,960 --> 01:09:33,320
rather than presenting a guest finished date as fact.

1855
01:09:33,320 --> 01:09:34,840
Planning can then see the exposure

1856
01:09:34,840 --> 01:09:37,400
and decide whether to wait, investigate or act.

1857
01:09:37,400 --> 01:09:39,160
A recalculated critical path tells you

1858
01:09:39,160 --> 01:09:40,920
what the current conditions allow.

1859
01:09:40,920 --> 01:09:43,320
It doesn't decide which response the plan should choose

1860
01:09:43,320 --> 01:09:46,760
when several feasible responses create different consequences.

1861
01:09:46,760 --> 01:09:49,800
Prediction, scheduling and optimization are different.

1862
01:09:49,800 --> 01:09:52,520
When a recalculated path tells you the current schedule

1863
01:09:52,520 --> 01:09:54,120
no longer fits the facts.

1864
01:09:54,120 --> 01:09:56,760
People often bundle three different jobs together.

1865
01:09:56,760 --> 01:09:59,640
Prediction, scheduling and optimization.

1866
01:09:59,640 --> 01:10:02,120
They're connected, but each one answers a different question.

1867
01:10:02,120 --> 01:10:03,880
Prediction asks what's likely to happen

1868
01:10:03,880 --> 01:10:05,640
if conditions continue as they are.

1869
01:10:05,640 --> 01:10:08,200
For example, given the current repair estimate,

1870
01:10:08,200 --> 01:10:11,320
remaining work, queue position, and recent execution times,

1871
01:10:11,320 --> 01:10:13,240
when will this order probably finish?

1872
01:10:13,240 --> 01:10:14,840
It can also estimate late risk

1873
01:10:14,840 --> 01:10:16,840
before the order actually misses its date,

1874
01:10:16,840 --> 01:10:19,240
and that helps the planner focus attention.

1875
01:10:19,240 --> 01:10:21,240
A prediction can use historical patterns,

1876
01:10:21,240 --> 01:10:23,960
current events and statistical or machine learning models

1877
01:10:23,960 --> 01:10:25,560
to detect patterns like a resource

1878
01:10:25,560 --> 01:10:27,240
that runs slower after a setup

1879
01:10:27,240 --> 01:10:29,960
or orders that enter a certain queue late in the day,

1880
01:10:29,960 --> 01:10:32,040
often waiting until the next shift.

1881
01:10:32,040 --> 01:10:34,360
Still, prediction doesn't prove a plan can work.

1882
01:10:34,360 --> 01:10:36,200
If a model predicts an order will finish Friday,

1883
01:10:36,200 --> 01:10:38,600
it hasn't checked every rule needed to produce that result.

1884
01:10:38,600 --> 01:10:40,840
It may not know whether the machine has the approved fixture,

1885
01:10:40,840 --> 01:10:42,600
whether a qualified operator is available

1886
01:10:42,600 --> 01:10:44,440
or whether the job can join the furnace load

1887
01:10:44,440 --> 01:10:45,480
that Friday afternoon.

1888
01:10:45,480 --> 01:10:46,840
It describes a likely outcome,

1889
01:10:46,840 --> 01:10:49,080
but it doesn't build a feasible sequence.

1890
01:10:49,080 --> 01:10:52,440
Scheduling takes the next step by assigning work to resources and time.

1891
01:10:52,440 --> 01:10:55,000
It respects the route, the order of operations,

1892
01:10:55,000 --> 01:10:58,840
resource capacity, shifts, process rules, material status,

1893
01:10:58,840 --> 01:11:01,560
and all the other jobs competing for the same equipment.

1894
01:11:01,560 --> 01:11:03,720
In practical terms, scheduling asks,

1895
01:11:03,720 --> 01:11:06,440
can we actually run this work here at this time in this order?

1896
01:11:06,440 --> 01:11:09,240
That question is less glamorous than an AI prediction,

1897
01:11:09,240 --> 01:11:11,480
but it's where most real production pressure sits.

1898
01:11:11,480 --> 01:11:13,080
A planner can't tell the supervisor

1899
01:11:13,080 --> 01:11:15,880
to start an operation merely because a risk score looks bad.

1900
01:11:15,880 --> 01:11:18,360
They need a schedule that accounts for what the resource can run

1901
01:11:18,360 --> 01:11:20,200
and what work must happen first.

1902
01:11:20,200 --> 01:11:22,680
A finite capacity schedule treats a machine,

1903
01:11:22,680 --> 01:11:25,720
furnace, test cell, or skilled operator as limited.

1904
01:11:25,720 --> 01:11:27,240
When you put one order into a slot,

1905
01:11:27,240 --> 01:11:29,800
another order must use a different slot or move later,

1906
01:11:29,800 --> 01:11:32,520
so the system can't quietly schedule 10 hours of work

1907
01:11:32,520 --> 01:11:35,320
into a four hour window and call it a plan.

1908
01:11:35,320 --> 01:11:37,720
EIP plans express demand, due dates,

1909
01:11:37,720 --> 01:11:40,040
and broad production intent very well,

1910
01:11:40,040 --> 01:11:42,520
yet a live shop floor sequence needs more rules,

1911
01:11:42,520 --> 01:11:44,520
more current facts, and much tighter timing.

1912
01:11:44,520 --> 01:11:46,920
That's why manufacturers often use an advanced planning

1913
01:11:46,920 --> 01:11:49,400
and scheduling system, usually called APS,

1914
01:11:49,400 --> 01:11:51,320
alongside EIP and MES.

1915
01:11:51,320 --> 01:11:52,920
Then optimization adds another layer.

1916
01:11:52,920 --> 01:11:55,000
Optimization starts with feasible schedules

1917
01:11:55,000 --> 01:11:57,800
and compares choices against a stated goal.

1918
01:11:57,800 --> 01:11:59,640
Maybe you want to reduce late deliveries,

1919
01:11:59,640 --> 01:12:01,080
protect a named customer order,

1920
01:12:01,080 --> 01:12:03,080
or aim for fewer setups, less overtime,

1921
01:12:03,080 --> 01:12:04,680
or lower work in progress.

1922
01:12:04,680 --> 01:12:06,840
You can't maximize all of those goals at once.

1923
01:12:06,840 --> 01:12:09,240
Move similar jobs together and you reduce setup time,

1924
01:12:09,240 --> 01:12:11,000
but an urgent order might wait too long.

1925
01:12:11,000 --> 01:12:13,640
Expedite the urgent order and you protect its delivery date,

1926
01:12:13,640 --> 01:12:15,320
while several other orders lose capacity.

1927
01:12:15,320 --> 01:12:17,080
The optimizer needs a clear objective,

1928
01:12:17,080 --> 01:12:19,400
priorities, and rules that it must never break.

1929
01:12:19,400 --> 01:12:22,040
That is a decision problem, not just a math problem.

1930
01:12:22,040 --> 01:12:23,960
A solver or specialist scheduling engine

1931
01:12:23,960 --> 01:12:27,080
can evaluate many combinations faster and more consistently

1932
01:12:27,080 --> 01:12:29,800
than someone manually shifting bars in a spreadsheet.

1933
01:12:29,800 --> 01:12:32,360
It can test alternate resources, sequence rules,

1934
01:12:32,360 --> 01:12:34,360
batch limits, and stated priorities,

1935
01:12:34,360 --> 01:12:37,240
then return schedules that obey the constraints you gave it,

1936
01:12:37,240 --> 01:12:39,720
but the phrase you gave it deserves attention.

1937
01:12:39,720 --> 01:12:41,240
If the model misses a tooling rule,

1938
01:12:41,240 --> 01:12:43,400
an inspection gate or a skill requirement,

1939
01:12:43,400 --> 01:12:45,560
the solver can produce a beautiful answer

1940
01:12:45,560 --> 01:12:47,240
that nobody can run.

1941
01:12:47,240 --> 01:12:48,600
The scheduling math may be correct,

1942
01:12:48,600 --> 01:12:50,600
but the factory model can still be wrong.

1943
01:12:50,600 --> 01:12:52,600
That's why the dependency model comes first.

1944
01:12:52,600 --> 01:12:55,720
Generative AI has a useful role around this process.

1945
01:12:55,720 --> 01:12:58,440
Copilot or Azure AI can help a planner ask questions

1946
01:12:58,440 --> 01:12:59,560
in normal language,

1947
01:12:59,560 --> 01:13:01,560
summarize the change behind a move date,

1948
01:13:01,560 --> 01:13:03,720
explain which conditions created the delay,

1949
01:13:03,720 --> 01:13:06,840
or guide someone to what the data and rules behind a recommendation.

1950
01:13:06,840 --> 01:13:08,280
It can help people investigate.

1951
01:13:08,280 --> 01:13:10,360
I would be careful about asking a language model

1952
01:13:10,360 --> 01:13:11,720
to invent the schedule itself

1953
01:13:11,720 --> 01:13:14,600
because generative AI produces plausible language,

1954
01:13:14,600 --> 01:13:16,760
but not deterministic constraint logic,

1955
01:13:16,760 --> 01:13:18,040
proof of feasibility,

1956
01:13:18,040 --> 01:13:21,160
or a defensible trade-off between competing customer commitments.

1957
01:13:21,160 --> 01:13:24,280
Fluent text isn't a production schedule.

1958
01:13:24,280 --> 01:13:26,120
A stronger design keeps the roles clear.

1959
01:13:26,120 --> 01:13:27,720
The planning and optimization engine

1960
01:13:27,720 --> 01:13:31,000
calculates feasible choices from approved rules and current data,

1961
01:13:31,000 --> 01:13:33,320
while the AI layer explains those results,

1962
01:13:33,320 --> 01:13:35,320
helps users explore scenarios

1963
01:13:35,320 --> 01:13:37,560
and brings the right context into the conversation.

1964
01:13:37,560 --> 01:13:40,680
That approach also lets people challenge the answer.

1965
01:13:40,680 --> 01:13:44,360
A planner can ask why order 184 moved behind another order,

1966
01:13:44,360 --> 01:13:46,120
and the system should point to capacity,

1967
01:13:46,120 --> 01:13:49,640
process rules, priority policy, or a change dependency,

1968
01:13:49,640 --> 01:13:51,400
not respond with a confident paragraph

1969
01:13:51,400 --> 01:13:53,640
that nobody can trace back to a source.

1970
01:13:53,640 --> 01:13:55,240
So prediction tells you what may happen,

1971
01:13:55,240 --> 01:13:56,920
scheduling tells you what can happen,

1972
01:13:56,920 --> 01:13:59,000
and optimization helps choose among the schedules

1973
01:13:59,000 --> 01:14:01,480
that can happen based on goals somebody has explicitly said.

1974
01:14:02,120 --> 01:14:05,320
Once you see the difference, the next question becomes unavoidable.

1975
01:14:05,320 --> 01:14:07,720
Which goal should win when every feasible replan

1976
01:14:07,720 --> 01:14:09,480
causes pain somewhere else?

1977
01:14:09,480 --> 01:14:11,560
The trade-offs behind every replan,

1978
01:14:11,560 --> 01:14:14,360
every replan picks a winner and creates pressure somewhere else.

1979
01:14:14,360 --> 01:14:16,040
That sounds obvious, but it gets lost

1980
01:14:16,040 --> 01:14:18,040
when a late order appears in a morning meeting

1981
01:14:18,040 --> 01:14:19,880
and someone asks for it to be moved to the front.

1982
01:14:19,880 --> 01:14:22,840
Moving it forward may protect one customer date,

1983
01:14:22,840 --> 01:14:25,000
yet the capacity has to come from another order,

1984
01:14:25,000 --> 01:14:27,160
another shift, or another part of the process.

1985
01:14:27,160 --> 01:14:28,520
A schedule can't remove that cost.

1986
01:14:28,520 --> 01:14:29,720
It can only show it.

1987
01:14:29,720 --> 01:14:32,760
Consider a machining cell with work already grouped by setup family,

1988
01:14:32,760 --> 01:14:34,760
keeping a similar set of tools and programs together

1989
01:14:34,760 --> 01:14:37,320
to reduce changeovers and give the cell a stable flow.

1990
01:14:37,320 --> 01:14:39,080
Then a customer order becomes urgent

1991
01:14:39,080 --> 01:14:40,760
and needs a different setup family.

1992
01:14:40,760 --> 01:14:42,200
You can insert it at the front,

1993
01:14:42,200 --> 01:14:44,040
but the machine now needs a changeover,

1994
01:14:44,040 --> 01:14:45,560
the grouped work moves back,

1995
01:14:45,560 --> 01:14:47,560
and the orders behind it may lose the time

1996
01:14:47,560 --> 01:14:49,560
they needed to meet their own commitments.

1997
01:14:49,560 --> 01:14:52,040
Protecting delivery date performance

1998
01:14:52,040 --> 01:14:54,360
can conflict with setup efficiency.

1999
01:14:54,360 --> 01:14:55,320
Neither goal is wrong,

2000
01:14:55,320 --> 01:14:57,240
setup efficiency protects capacity,

2001
01:14:57,240 --> 01:14:58,360
reduces disruption,

2002
01:14:58,360 --> 01:15:00,760
and often helps the plant keep a predictable rhythm,

2003
01:15:00,760 --> 01:15:02,520
while customer delivery matters too,

2004
01:15:02,520 --> 01:15:04,840
especially where a late shipment stops a customer line

2005
01:15:04,840 --> 01:15:06,760
or affects a contract commitment.

2006
01:15:06,760 --> 01:15:08,600
The planner needs a rule for choosing,

2007
01:15:08,600 --> 01:15:10,920
not just permission to drag an order higher in a queue.

2008
01:15:10,920 --> 01:15:13,400
Sometimes the right response is overtime,

2009
01:15:13,400 --> 01:15:15,000
which creates extra capacity,

2010
01:15:15,000 --> 01:15:16,840
but also needs people, approvals,

2011
01:15:16,840 --> 01:15:18,360
budget, and safe work limits.

2012
01:15:18,360 --> 01:15:20,680
A plan that assumes overtime every time a date moves

2013
01:15:20,680 --> 01:15:22,840
will eventually turn overtime into the normal schedule,

2014
01:15:22,840 --> 01:15:24,120
which is usually a sign that

2015
01:15:24,120 --> 01:15:26,680
the capacity model needs a more honest conversation.

2016
01:15:26,680 --> 01:15:28,360
Extra hours aren't free capacity,

2017
01:15:28,360 --> 01:15:30,440
subcontracting can change the path as well.

2018
01:15:30,440 --> 01:15:31,560
It may protect a due date

2019
01:15:31,560 --> 01:15:33,720
when an internal resource cannot recover,

2020
01:15:33,720 --> 01:15:35,880
but the order now depends on supplier lead time,

2021
01:15:35,880 --> 01:15:37,560
transport, external quality checks,

2022
01:15:37,560 --> 01:15:39,480
and clear control of the product definition.

2023
01:15:39,480 --> 01:15:41,080
You haven't removed dependencies,

2024
01:15:41,080 --> 01:15:43,000
you've moved some of them outside the plant,

2025
01:15:43,000 --> 01:15:44,200
that may still be the best choice,

2026
01:15:44,200 --> 01:15:46,120
but it needs to be visible.

2027
01:15:46,120 --> 01:15:48,360
Alternate routing creates a similar trade-off,

2028
01:15:48,360 --> 01:15:50,120
a different machine may run the operation,

2029
01:15:50,120 --> 01:15:51,400
but at a slower rate,

2030
01:15:51,400 --> 01:15:54,040
with more setup time or with a limited qualification.

2031
01:15:54,040 --> 01:15:56,600
The alternate route may keep the order moving,

2032
01:15:56,600 --> 01:15:58,120
while it consumes a resource

2033
01:15:58,120 --> 01:15:59,960
that another product needs even more.

2034
01:15:59,960 --> 01:16:02,040
A free machine isn't always spare capacity.

2035
01:16:02,040 --> 01:16:03,880
Batch splitting can help in some flows.

2036
01:16:03,880 --> 01:16:06,520
If part of an order can move to the next process early,

2037
01:16:06,520 --> 01:16:08,120
you may protect a partial shipment

2038
01:16:08,120 --> 01:16:10,200
or start a downstream operation sooner.

2039
01:16:10,200 --> 01:16:11,800
Yet, splitting may increase handling,

2040
01:16:11,800 --> 01:16:13,400
inspection effort, traceability work,

2041
01:16:13,400 --> 01:16:14,760
and the risk that small quantities

2042
01:16:14,760 --> 01:16:16,360
wait longer than the main batch.

2043
01:16:16,360 --> 01:16:18,840
The right choice depends on the actual order rules,

2044
01:16:18,840 --> 01:16:20,920
not a generic preference for speed.

2045
01:16:20,920 --> 01:16:22,760
Some constraints should never become trade-offs.

2046
01:16:22,760 --> 01:16:24,600
Safety rules don't move because a due date

2047
01:16:24,600 --> 01:16:25,720
looks uncomfortable,

2048
01:16:25,720 --> 01:16:27,640
and quality requirements don't disappear

2049
01:16:27,640 --> 01:16:30,120
because a planner sees a red delivery status.

2050
01:16:30,120 --> 01:16:32,280
A resource without the right capability

2051
01:16:32,280 --> 01:16:33,880
doesn't become qualified,

2052
01:16:33,880 --> 01:16:35,880
because the schedule needs an answer.

2053
01:16:35,880 --> 01:16:38,360
Those boundaries need to sit inside the planning logic.

2054
01:16:38,360 --> 01:16:39,960
Compliance rules belong there too.

2055
01:16:39,960 --> 01:16:41,640
If a product needs approved documentation,

2056
01:16:41,640 --> 01:16:43,800
a controlled process, or a named release step,

2057
01:16:43,800 --> 01:16:45,640
the replan must respect that condition.

2058
01:16:45,640 --> 01:16:47,400
A schedule that ignores it may look better

2059
01:16:47,400 --> 01:16:50,120
for an hour and create a much worse issue later.

2060
01:16:50,120 --> 01:16:52,920
Not every feasible looking option is an allowed option.

2061
01:16:53,720 --> 01:16:55,240
This brings us to decision authority.

2062
01:16:55,240 --> 01:16:57,960
A planner can often re-sequence work with in agreed rules,

2063
01:16:57,960 --> 01:17:00,920
but over time, may need a production manager's approval,

2064
01:17:00,920 --> 01:17:04,280
a subcontract decision may need purchasing and quality involvement,

2065
01:17:04,280 --> 01:17:05,880
and a customer promise change

2066
01:17:05,880 --> 01:17:08,440
may belong with customer service or commercial leadership.

2067
01:17:08,440 --> 01:17:11,480
The system should know when to stop and ask

2068
01:17:11,480 --> 01:17:13,240
that doesn't mean slowing every decision down

2069
01:17:13,240 --> 01:17:14,440
with a workflow maze,

2070
01:17:14,440 --> 01:17:17,240
but setting approval thresholds that match the consequence.

2071
01:17:17,240 --> 01:17:19,720
A small local sequence change may happen quickly,

2072
01:17:19,720 --> 01:17:21,720
while a change that risks another customer

2073
01:17:21,720 --> 01:17:23,320
uses unapproved over time,

2074
01:17:23,320 --> 01:17:25,800
or moves controlled work onto an alternate route

2075
01:17:25,800 --> 01:17:28,520
needs clear ownership and a traceable decision.

2076
01:17:28,520 --> 01:17:31,320
People on the floor need to know which schedule is real.

2077
01:17:31,320 --> 01:17:33,240
If a planning tool produces a new sequence,

2078
01:17:33,240 --> 01:17:34,920
but the supervisor hasn't accepted it,

2079
01:17:34,920 --> 01:17:37,640
or cannot execute it with the available crew and tools,

2080
01:17:37,640 --> 01:17:39,480
the schedule remains a suggestion.

2081
01:17:39,480 --> 01:17:41,560
Re-planning only works when the approved decision

2082
01:17:41,560 --> 01:17:43,480
reaches the people who must carry it out

2083
01:17:43,480 --> 01:17:45,480
and when their feedback can correct assumptions

2084
01:17:45,480 --> 01:17:47,080
that no system can see yet.

2085
01:17:47,080 --> 01:17:48,680
That is where planning becomes operational.

2086
01:17:48,680 --> 01:17:51,320
So the question isn't, can we recover this order?

2087
01:17:51,320 --> 01:17:52,600
The better question is,

2088
01:17:52,600 --> 01:17:55,320
which approved action protects the most important commitment

2089
01:17:55,320 --> 01:17:57,560
and what does that action cost the rest of the plan?

2090
01:17:57,560 --> 01:18:00,920
A live dependency model gives you a way to test that question

2091
01:18:00,920 --> 01:18:02,840
before the plant commits to a change.

2092
01:18:02,840 --> 01:18:05,720
scenario, a bottleneck machine stops mid-shift.

2093
01:18:05,720 --> 01:18:07,800
Here's a real pressure test for the dependency models,

2094
01:18:07,800 --> 01:18:09,640
so you've got a high priority work order

2095
01:18:09,640 --> 01:18:11,480
running on the only machining center qualified

2096
01:18:11,480 --> 01:18:13,880
for that part family and you're halfway through the shift.

2097
01:18:13,880 --> 01:18:16,600
The operator stops the machine after an alarm

2098
01:18:16,600 --> 01:18:18,440
and maintenance starts checking the fault.

2099
01:18:18,440 --> 01:18:20,520
Nobody yet knows whether that takes 20 minutes

2100
01:18:20,520 --> 01:18:21,800
or most of the day.

2101
01:18:21,800 --> 01:18:23,800
Pushing every order on that machine to tomorrow

2102
01:18:23,800 --> 01:18:24,920
isn't the answer.

2103
01:18:24,920 --> 01:18:27,560
Planning needs a clear view of what's already completed,

2104
01:18:27,560 --> 01:18:29,080
what remains on the current job,

2105
01:18:29,080 --> 01:18:30,440
which jobs are in the queue,

2106
01:18:30,440 --> 01:18:33,560
and what maintenance can credibly estimate about the return time.

2107
01:18:33,560 --> 01:18:35,400
Those facts define the starting point.

2108
01:18:35,400 --> 01:18:38,120
Suppose the current order still needs another hour of machining

2109
01:18:38,120 --> 01:18:40,040
before it moves to its next process.

2110
01:18:40,040 --> 01:18:41,880
Behind it, sit two other work orders.

2111
01:18:41,880 --> 01:18:45,000
One has a later due date but feeds the customer assembly shipment.

2112
01:18:45,000 --> 01:18:46,520
The other has an earlier due date,

2113
01:18:46,520 --> 01:18:49,160
but its next operation has capacity later in the week.

2114
01:18:49,160 --> 01:18:52,040
A simple priority flag won't settle that sequence.

2115
01:18:52,040 --> 01:18:54,920
The high priority order may have an alternate machining center,

2116
01:18:54,920 --> 01:18:57,960
but that resource can only run the part with a specific fixture.

2117
01:18:57,960 --> 01:19:00,200
The fixture currently sits on another machine.

2118
01:19:00,200 --> 01:19:01,480
Even if someone moves it,

2119
01:19:01,480 --> 01:19:03,160
the alternate machine needs a setup,

2120
01:19:03,160 --> 01:19:06,520
and only one operator on the next shift holds the required qualification.

2121
01:19:06,520 --> 01:19:08,040
So the alternate route exists.

2122
01:19:08,040 --> 01:19:09,400
It just doesn't exist right now.

2123
01:19:09,400 --> 01:19:11,240
Here's what changes the critical path.

2124
01:19:11,240 --> 01:19:13,640
The stopped machine creates the first disruption,

2125
01:19:13,640 --> 01:19:16,280
but the delivery impact may come from the fixture move,

2126
01:19:16,280 --> 01:19:17,400
the operator calendar,

2127
01:19:17,400 --> 01:19:19,720
or the lost place in a downstream process queue.

2128
01:19:19,720 --> 01:19:21,400
The machine failure starts the chain.

2129
01:19:21,400 --> 01:19:23,240
It doesn't always control the final date.

2130
01:19:23,240 --> 01:19:26,760
Let's say maintenance estimates a return late in the next shift.

2131
01:19:26,760 --> 01:19:28,920
Planning now has a few possible responses.

2132
01:19:28,920 --> 01:19:31,640
They can hold the current work order and wait for the repair.

2133
01:19:31,640 --> 01:19:34,760
They can transfer the remaining quantity to the alternate machine.

2134
01:19:34,760 --> 01:19:36,120
Or they can split the order,

2135
01:19:36,120 --> 01:19:38,280
run the completed quantity through the next process

2136
01:19:38,280 --> 01:19:40,040
and decide later what to do with the rest.

2137
01:19:40,040 --> 01:19:42,520
Each option creates a different set of dependencies.

2138
01:19:42,520 --> 01:19:44,600
Waiting may keep the preferred sequence intact,

2139
01:19:44,600 --> 01:19:47,960
but it risks the customer date if the repair estimate moves again.

2140
01:19:47,960 --> 01:19:50,440
Moving to the alternate machine may protect the first order

2141
01:19:50,440 --> 01:19:52,600
while displacing work already planned there.

2142
01:19:52,600 --> 01:19:55,160
Splitting quantity may start downstream work earlier,

2143
01:19:55,160 --> 01:19:57,880
but only if the next process accepts partial transfer

2144
01:19:57,880 --> 01:19:59,480
and the traceability rules allow it.

2145
01:19:59,480 --> 01:20:02,520
There isn't one correct answer hiding inside the alarm code.

2146
01:20:02,520 --> 01:20:04,040
A live scheduling model can calculate

2147
01:20:04,040 --> 01:20:06,360
the earliest feasible completion for each option.

2148
01:20:06,360 --> 01:20:08,040
It can check the remaining machining time,

2149
01:20:08,040 --> 01:20:10,360
setup duration, tool and fixture availability,

2150
01:20:10,360 --> 01:20:12,840
operator shifts, downstream capacity,

2151
01:20:12,840 --> 01:20:15,240
and the shipment commitments linked to each order.

2152
01:20:15,240 --> 01:20:17,400
Then it can show the planner the consequences

2153
01:20:17,400 --> 01:20:19,000
in plain operational terms.

2154
01:20:19,000 --> 01:20:22,760
For example, waiting for repair might place the high priority order at risk,

2155
01:20:22,760 --> 01:20:24,840
but preserve the dates of two other orders.

2156
01:20:24,840 --> 01:20:26,360
Moving the work to the alternate machine

2157
01:20:26,360 --> 01:20:27,800
might protect that priority order

2158
01:20:27,800 --> 01:20:30,920
while pushing another customer order past its dispatch cutoff.

2159
01:20:30,920 --> 01:20:33,640
Splitting the batch may keep a partial quantity moving,

2160
01:20:33,640 --> 01:20:36,120
but create extra handling and inspection work.

2161
01:20:36,120 --> 01:20:37,720
Those are decisions people can discuss.

2162
01:20:37,720 --> 01:20:39,960
Notice what the planner still needs from the shop floor.

2163
01:20:39,960 --> 01:20:41,800
A repair estimate from maintenance,

2164
01:20:41,800 --> 01:20:43,640
confirmation that the alternate machine

2165
01:20:43,640 --> 01:20:45,240
can produce conforming parts.

2166
01:20:45,240 --> 01:20:46,680
The real status of the fixture,