Microsoft 365 Backup - Simply Explained

Welcome to the M365 FM podcast, where we explore the people, ideas and technologies shaping the future of work. Why being in the cloud is not the same as being backed up. Most people assume Microsoft already backs up their email, and that assumption is where the trouble starts. Your files sit on Microsoft's servers so they feel safe, but a server staying online is a different job from handing you a clean copy of last week. Welcome to the M365 FM podcast, where we explore the people, ideas and technologies shaping the future of work. Today we're talking about Microsoft three hundred sixty five backup, the paid layer Microsoft built for exactly that job. It covers three things costs by the gigabyte and leaves one gap that surprises almost everyone. By the end, you will know what it protects, what it costs, and where it stops. There's a piece of this that trips up even careful teams. When you open outlook or OneDrive, you're looking at a service. Microsoft works hard to keep online and that work is real. Microsoft runs the servers, copies your data across buildings and regions, and swaps in a healthy copy. When hardware fails, none of that hands you a clean copy of your data from an earlier point in time. Microsoft's own position on this hasn't changed, and it's simple enough to say plainly Microsoft keeps the platform running, and you stay responsible for your data and for the recovery decisions your business needs. Those are two different jobs, and only one of them happens automatically. The way to picture it is simple. A live service is like a building that stays open every day with the lights on and the doors unlocked. A backup is a photograph of that building from last Tuesday, kept somewhere safe in case the original changes for the worse. Microsoft is very good at keeping the building open and keeping. The photograph has always been the customer's job. Unless you pay for a tool that does it for you. Available and recoverable are not the same word. A service can be up and running, while a file you deleted three months ago is simply gone, or while an attacker has quietly changed the version you are counting on. Your recycle bin looks like a safety net, but it has a clock attached. And when the clock runs out, the file doesn't come back. So the honest question isn't whether Microsoft keeps your email online because it does. The question is what happens on the day you need yesterday's version back? And that's where a real backup earns its place. What Microsoft three hundred sixty five backup actually is Microsoft three hundred sixty five backup is a paid service Microsoft built and runs itself. It makes copies of your data at points in time, then lets you roll that data back to one of those points when something goes wrong. It isn't a third party tool bolted onto your account, and it isn't a setting you flip once and forget it exists, because Microsoft's own recovery tools, useful as they are, were never built to roll back an entire site or an entire mailbox after a serious incident. Backup fills that one job, and it does it without copying your data outside Microsoft's own boundary. It protects three things, and the list is shorter than you might expect. The first is SharePoint sites, which is where most team documents and shared libraries live. The second is OneDrive accounts, which is each person's own file space. The third is Exchange Mailboxes, which covers email, calendar, contacts, and tasks for mailboxes. It supports both user mailboxes and shared mailboxes, and it doesn't support room mailboxes or group mailboxes. Two words will carry through the rest of this episode, so they need a quick definition. A workload is one of those three areas. So SharePoint is a workload and so is exchange. A restore point is a saved moment in time. You can return to a bit like a bookmark. You can jump back to when a page goes wrong. There's one more shape to this. You can run this straight from the Microsoft three hundred sixty five Admin Center, or you can use a partner application built on the same Microsoft backup platform. Either way, you pay by the gigabyte rather than by the person, and that billing choice sets up the next surprise recycle bins version. History and real backup are three different things. Before you spend anything, it helps to know what you already get for free, because plenty of people pay for backup without understanding why the free options weren't enough. SharePoint and OneDrive give you version history so you can pull up an older copy of a single file. They also give you two recycle bins, and the second one holds deleted content for about ninety three days. There's a files restore feature too, which can roll a library back to a point within the last thirty days. Exchange keeps deleted items for fourteen days by default, and an administrator can stretch that to thirty for an ordinary mistake. That's genuinely enough. Someone deletes the wrong spreadsheet on a Tuesday afternoon and you pull it out of the recycle bin in about five minutes. Nobody needs a paid service for that. The trouble starts when the damage is bigger than one. File picture ransomware that encrypts thousands of documents across dozens of sites in a single night. Now, the recovery job is a whole site or a whole mailbox. And clicking through version history one item at a time doesn't scale. Version history can run out as well because administrators cap how many versions get kept, and every recycle bin has an expiry date. There's a smaller version of the same problem to. A manager. Spends a week editing a policy document, then realizes the version they wanted is six revisions back and the version limit cut it off. Version history is built for the last few edits, not for the kind of memory a business sometimes needs. There's another confusion here. Retention and legal hold are not backup. Retention keeps information around because a policy or a law says you must. And legal hold is built to export data for a case. Neither one is designed to handle a broken site. Back to you. In a working state, even disaster recovery is a different animal because it keeps your data in its current state rather than an older, healthier one. The short way to hold it all together is that the free tools protect you from today's mistake, while backup protects you from a bad day that already spread. That leaves a real gap. And the next question is how far back a proper backup can take you. How far back it can take you, and how fast you get data back. The first number to understand is the recovery window, and it's the setting you choose. You can keep restore points for three months, six months, one year, or two years if you change nothing. Existing policies sit at one year. So a site you protect today may only be recoverable back to about this time next year. The second number is how often those points get made. And this is the one people misread for your SharePoint sites and OneDrive accounts. You get a point roughly every ten minutes across the last two weeks, and then weekly snapshots out to a year for exchange. Those points keep coming about every ten minutes for the entire year. Microsoft is not photographing your entire mailbox every ten minutes, and that distinction matters. What happens is that changes get saved as a new version inside that ten minute window, no matter how many edits happen and deletions are always captured. The ten minute figure describes how much recent change you can recover. Not a fresh copy of everything on a timer. Speed is the next thing people get wrong. A single site smaller than about a terabyte can come back in under twenty minutes. When you use one of the recommended Express Restore points. At larger scale, restores have run at a median of about two terabytes an hour. Exchange restores move at roughly one hundred to five hundred items per mailbox per minute. To make that real, imagine a finance team loses a shared folder of contracts on a Friday morning with a recommended restore point. They can be back in business before lunch, and the restore doesn't get slower just because the folder is large. What slows things down is restoring many sites at once, which is exactly what happens in a big incident. Restore speed depends far more on how many sites you're restoring and which restore point you choose than on how much data you have. Restoring one huge site can be quicker than restoring fifty small ones, which is the opposite of the way it usually gets pictured. One last detail decides how a restore feels. A full site restore rolls everything back to the moment you picked and it overwrites. Anything changed since then? A file version restore moves a single file forward to that moment while keeping the older versions intact. Pick the wrong one and you can undo good work along with the bad. What it costs and how the billing actually works. Now we get to the money. The list price is fifteen cents per gigabyte per month of protected content, and restores themselves are free. That sounds simple, and the simplicity is real, but the way Microsoft counts a gigabyte is where the real cost hides. You don't pay by the user and you don't pay a flat license. You pay through an Azure subscription on a pay as you go basis, which means the bill moves with your data. If you've never touched Azure, that's fine because the setup walks you through linking a subscription, and the backup service is the thing you're buying. The important part is that you're charged for more than the files you can see. The bill includes the first recycle bin, the second recycle bin, and any online archives attached to your mailboxes. Microsoft's own example makes it concrete protect a one gigabyte site that has half a gigabyte sitting in its second recycle bin, and add a one gigabyte mailbox with a one gigabyte online archive, and you're billed for three and a half gigabytes instead of two. There's another catch, and it shows up before the bill does. Once content is protected, you keep paying for the retained copy until the recovery window runs out. Delete half a site and the next bill stays about the same, because the backup is still holding the deleted material for a year. The bill follows your history rather than your current size. Put a number on it for a small team. Protect two hundred gigabytes across your sites and mailboxes, and the list price works out to about thirty dollars a month. The figure moves up or down with the size of what you protect, and it includes the deleted material sitting in your recycle bins. So the total you see in a File Explorer isn't always the total you pay for. One piece of good news is that changing how often restore points are made doesn't meaningfully change the cost, so there's no reason to protect your data less often to save money. Microsoft also publishes a pricing calculator and usage reports, which lets you estimate the bill from real numbers instead of a guess. Once you know the price, the only question left is whether you're paying for the right things, and that's where the coverage list comes in. The three things it protects and the one big thing it doesn't buy. Now the coverage list should feel familiar. Microsoft three hundred sixty five backup protects SharePoint sites, OneDrive accounts, and exchange mailboxes. And that's the entire list. Everything else in Microsoft three hundred sixty five sits outside it, and the biggest name on that list is teams, teams, chats, channel posts, and replies are not backed up. This is the misconception that catches people hardest because teams feels like the center of work, and the assumption that it's protected comes almost for free. It isn't protected, and that gap is easy to miss. The cost of that gap goes beyond annoyance. Teams conversations often hold the reasoning behind a decision, the approval, or the context for a file, and those threads are exactly what backup cannot return. If a channel gets deleted by mistake, you may get the documents back and still lose the story of why they exist. There's a small mercy here, so get this part right. Teams files are a different story from teams conversations. A file shared in a channel usually lives in that team's SharePoint site, and a file shared in a chat usually lives in the sender's OneDrive. If those underlying places are protected, the files can come back. What can't come back is the conversation itself, meaning the messages, the replies, and the thread around the file. You can recover the document and still lose the discussion about it. Restoring individual files and folders is now generally available for SharePoint and OneDrive, which matters a lot for beginners. You can browse a saved point, pick the exact files you need, and put them back in place or into a new folder. There are limits to know, and they're honest ones. Files in document libraries that use information rights management can't be restored this way, and neither can files in locked sites. Exchange is similar but has its own ceiling. You can search a mailbox and restore selected items, but each restore is capped at one thousand items, so a broad search has to be narrowed before you run it. That's less a flaw than a nudge. To be precise, teams is the loudest gap, but it isn't the only one, and the rest of the gaps live somewhere less obvious. The other blind spots, identity settings, and other apps. Backup protects your collaboration data, and it leaves the controls around that data alone. Think about the settings that decide who can sign in, what they can reach and how mail gets routed, conditional access policies, administrator roles, mail forwarding rules, connectors, and tenant wide settings are not part of a data restore that matters in a specific and uncomfortable way. An attacker who gets in often changes the rules before taking anything by adding a forwarding rule, weakening a sign in policy, or creating a hidden administrator account. Restoring your files doesn't undo any of that. Those changes need their own plan. Usually a mix of exports, documentation, and a rebuild checklist. Identity is its own separate service, and this is where a lot of people assume too much. Microsoft Entra Backup handles identity objects like users and groups. It's a different product with its own licensing, and it's still in preview with a short window of about seven days. Microsoft three hundred sixty five backup doesn't recover your logins, your roles, or your access rules, and no amount of data restore will rebuild them. Identity recovery is a solvable problem, and Microsoft is building it out. The catch is that it's a separate piece you have to plan for, rather than something you inherit for free. Other applications sit outside the scope. Two planar forms Power Platform, Loop and Viva aren't automatically protected by the service, even though they may hold work. Your team depends on if a planar board or a power automate flow disappears. Backup won't hand it back. None of this is a reason to panic, and it isn't a knock on the product. It's a map of how your data is protected and knowing which layer covers what is the whole skill. That's a skill you have now. How you turn it on and the mistakes that trip people up. Setting this up takes less time than you might expect, and it follows two steps. First, you link an Azure subscription for pay as you go billing. Then you create backup policies that choose which sites, accounts, and mailboxes get protected. You don't need to be a global administrator to run it. A SharePoint administrator can set it up, and Microsoft added a dedicated Microsoft three hundred sixty five backup administrator role, which lets you hand out exactly the access this job needs. Instead of something far more powerful, you'll need owner or contributor rights on that Azure subscription, so check that before you start. You can protect everything, or you can protect only the things that matter. Protecting selectively keeps the bill down, but it creates a slower risk because new sites and mailboxes keep appearing and can slip outside your policy. If you go selective, someone has to watch the edges. There's a trade off hiding in that choice. Broad protection is easier to forecast and harder to get wrong, while selective protection saves money only if your list of important things stays accurate. New projects, new teams, and new mailboxes all create fresh gaps, and those gaps need an owner? Turn on the admin email alerts and do it on day one. The events that should worry you, like pausing a policy or starting the offboarding of your backups are exactly the events nobody notices in real time. The alert turns a silent mistake into a message in an inbox. A few mistakes show up again and again. People assume the service is on by default, and it isn't. People expect hybrid exchange mailboxes to be protected, and only mailboxes fully hosted in exchange online can be. People assume every SharePoint site is supported and a handful of older or special templates aren't. Some people turn off pay as you go without realizing that turning it off starts the off boarding process for their backups. That offboarding isn't instant, which is the one bit of comfort here. There's a ninety day grace period to recover backups after you switch it off, and shortening that to seven days requires a second administrator to approve. There's also a testing limit that surprises teams. You can run a restore purely to test it about twice a month per protected item, so test on purpose rather than by accident when you actually need it and when you don't. The clearest case for paying is ransomware, where an attack encrypts whole sites and mailboxes in one sweep, and you need a clean rollback to before the damage. A close second is a large accidental deletion where thousands of files vanish at once, and picking through a recycle bin is hopeless. Another is a person leaving when their mailbox or OneDrive still has to be recoverable long after they're gone. A regulated site that needs a longer memory than a recycle bin can offer rounds out the list. Skipping it can make sense in narrower cases if your data is small. If mistakes get caught the same day, and if the free recycle bins and version history already cover your worst realistic case, then a paid layer may be more comfort than necessity. The deciding factor isn't really the price. It's two numbers. People rarely write down how much data you can afford to lose and how long you can afford to be down. Answer those honestly, and the choice tends to make itself. Most organizations end up using a mix, and that isn't a failure. The free tools handle small mistakes, backup handles the big rollback and retention handles the legal reasons each one covers a job the others can't. That honesty is the whole point of this episode. None of it is about fear. It's about knowing which layers of protection you actually have. So you aren't counting on one that was never there. Here's the short version of everything we covered. Your data in the cloud isn't automatically backed up. Microsoft three hundred sixty five backup adds a real recovery layer for SharePoint, OneDrive, and exchange restores from points in time, builds by the gigabyte, and still leaves teams, conversations, identity, and other apps out of the picture. So I'll leave you with one question. If your recycle bin emptied tonight, which file of yours would you not be able to get back if this made the topic feel less intimidating? Subscribe to M365 FM to master the tools and the mindset shaping the future of work.