How to Measure the Impact of Continuous Improvement Projects

Disclosure: M365.FM host Mirko Peters works at iFAKT GmbH. This page describes the company factually; paid partnerships and links are labelled as advertising.

By Mirko Peters · 2026-11-11

Continuous improvement KPIs that prove impact: set a baseline, pick few indicators, compare before and after data fairly and turn gains into real savings.

To measure the impact of continuous improvement, take a baseline before you change anything, choose one to three KPIs that match the goal, and compare before and after data under similar conditions. Convert gains to money only with finance. This guide is for continuous improvement managers who need to show results to management without overselling them. It covers which continuous improvement KPIs to use, how to compare data fairly, how to calculate savings, and a simple tracking sheet.

Key points

• Measure impact with a baseline taken before the change, one to three KPIs linked to the goal, and a before and after comparison under similar conditions.
• Use both process indicators, such as lead time and changeover time, and outcome indicators, such as cost, quality and delivery.
• Compare the result with the normal range of the baseline, because a change inside normal variation proves nothing.
• Freed capacity is not a saving until it is used or removed, so classify gains as hard, soft or avoided cost and agree the rules with finance.
• Keep a simple tracking sheet for every project, with owner, baseline, target, result and a check at 30, 60 and 90 days.

Which continuous improvement KPIs should you track?

A key performance indicator (KPI) is a measurable value that shows whether you are reaching a business goal. There are many types of KPIs, such as operational KPIs for the shop floor and financial KPIs for the income statement. For improvement work, use two groups: leading indicators that move first and lagging indicators that confirm the result.

• Process KPIs are process performance metrics. They show how the work behaves: lead time, changeover time, first-pass yield, work in process (WIP), number of stops. They move first and respond fast.
• Outcome KPIs show results the business cares about: cost, quality, on-time delivery, safety, customer satisfaction. They move later and depend on more factors.

The Lean Enterprise Institute (LEI) makes the same point in its article on explaining lean's financial benefits to leadership. It advises tracking process indicators, such as whether batches are actually smaller and whether finished goods and work-in-process inventory are falling, and tracking the number of improvements people make. Money follows later, and cost savings rarely show up right away.

Choose few indicators. One to three per improvement initiative is enough, and each needs a definition, a data source and an owner. An effective KPI is specific and measurable, and a leading KPI such as changeover time gives early warning. To drive continuous improvement over the long term, add one KPI for the health of the improvement system itself, such as the number of improvements completed per month, so that you can see whether continuous improvement is alive and a culture of continuous improvement is growing.

How do you set a baseline?

The baseline is the "before" picture. Without it, you cannot prove impact.

1. Define the KPI precisely. Which product, which shift, which unit, which time stamps? Write the definition down.
2. Collect enough data. A few weeks is better than a single day, because you need to see normal variation. Record the lowest and highest values as well as the average.
3. Use the same source later. If the baseline comes from a machine counter and the result from a handwritten sheet, the comparison is not valid.
4. Note the conditions. Product mix, volume, staffing and known problems all affect the number.
5. Store it. Put the baseline on the tracking sheet before the work starts, and set targets at the same time.

Who needs which numbers?

Different stakeholders need different views. Operators want to track progress on the board every day, and they need a number they can influence. Supervisors monitor the process KPIs weekly and use them to identify areas for improvement. Plant management wants to monitor the performance of the plant in a monthly view of key performance indicators in business terms: quality improvement, delivery and cost. Finance wants the money, classified and checked.

Keep one data source for all of them, so the KPI data is the same everywhere. A data-driven review works only when the numbers on the shop floor match the numbers in the report. You can also benchmark between lines or sites, if the definitions are the same, and use simple analytics, such as a trend line, to spot a business process that is drifting. Analyzing KPIs and their trends gives actionable insights, and process improvement becomes a matter of facts instead of opinions. Good performance data also links your improvement efforts to operational efficiency and business goals.

How do you compare before and after data?

A fair comparison asks one question: would the number have changed anyway?

• Compare rates, not totals. Misprints per 1,000 units produced, not total misprints, if volume changed.
• Compare like with like. Same product family, same shift pattern, similar volume.
• Check the normal range. If the baseline weeks vary between 47 and 57 minutes, a result of 55 proves nothing. A result of 41 that holds for several weeks is clearly outside the range.
• Wait for stability. New methods often need a few weeks to settle. Measure after that.
• Look at the long term. Check performance over time at 30, 60 and 90 days to see whether the gain holds. KPI trends tell more than one value.
• Record other changes. A new material, a new operator or a repaired machine can explain a result.

The check step of the PDCA cycle is built for this. The American Society for Quality describes the PDCA cycle as plan, do, check, act, in which you test a change on a small scale and review the results. See the ASQ page on the PDCA cycle.

How do you turn the impact into savings?

Finance needs a clear classification. A guide on hard and soft savings from Airacad describes three levels:

• Hard savings reduce actual spend and can be verified in accounting records. Finance usually gives full credit.
• Soft savings are efficiency or productivity gains that do not immediately show as cash. Finance often gives partial credit or waits for conversion.
• Cost avoidance is spend that does not happen, such as not needing extra space.

The same guide advises agreeing the classification and conversion rules with finance before approval and avoiding double counting across overlapping initiatives. The Kaizen Institute makes the same point for improvement events: gains should be quantified, with validation by the finance department. See its article on the rapid improvement event.

LEI adds an important warning. Removing waste frees up capacity, but unless that capacity is used for something else, it does not appear on the income statement. So state what will happen to the freed time: more output, less overtime or fewer temporary workers.

A simple tracking sheet

Keep one line per improvement. This is the core of tracking and an easy form of process management:

• name, owner and start date
• problem and goal, with the KPI and its definition
• baseline value, range and period
• target value and date
• result at 30, 60 and 90 days
• type of gain: hard, soft or avoided cost
• status of finance review
• standard updated: yes or no

Show the sheet on a board or dashboard in the monthly review. This is a best practice because it keeps goals visible. A good dashboard makes the trend visible and gives actionable information: which initiatives deliver, which are late, which results have faded.

Example: before and after for a changeover

This example uses assumed numbers.

Baseline (8 weeks): the changeover on line 3 takes 52 minutes on average. The weekly average ranged from 47 to 57 minutes. There are 5 changeovers per week.

After the change (4 weeks, stable): the average is 41 minutes.

• Reduction per changeover: 52 − 41 = 11 minutes, or 11 ÷ 52 = about 21%.
• Weekly gain: 5 × 11 = 55 minutes, or about 0.92 hours.
• Yearly gain over 50 weeks: 0.92 × 50 = about 46 hours.
• The result of 41 minutes is below the baseline range of 47 to 57, so it is unlikely to be random.

Money. If the line is a bottleneck with demand, and it makes 120 units per hour with a contribution margin of €6 per unit, 46 hours could add 46 × 120 × 6 = €33,120 per year. If there is no extra demand, the saving is zero in cash terms, and the gain is spare capacity. Report it as a soft gain until the hours are used, and let finance decide how to count it.

Common mistakes

• Starting without a baseline. You cannot prove a change you did not measure.
• Tracking too many KPIs. KPI improvement needs focus. Pick a few, and tie each to the goal.
• Mixing data sources. Before and after must use the same definition and the same source.
• Reporting one good week. Check the range and the long-term trend.
• Counting capacity as cash. Time freed is only a saving once it is used or removed.
• Skipping finance. A number that finance has not seen will be challenged.

How software can help

Measuring the overall performance of an improvement plan needs reliable data, and a management system that uses it for operational excellence. iFAKT's Polaris Analyzer is a module of iFAKT Polaris with KPI dashboards across all modules that compare planned and actual values in production. That gives the team a shared baseline and a view of the result after a change. Before the change starts, iFAKT's Polaris Simulator, a module that tests the production system virtually, for example capacity, staffing and buffer scenarios, can help estimate the effect of a change. Estimates still need a check on the shop floor.

Next step

For your next improvement, write the KPI definition and the baseline on the tracking sheet before the first change, and book the 30, 60 and 90 day checks. That is the simplest way to make continuous improvement KPIs credible.