Three figures set Microsoft 365 Backup restore points: 10-minute steps for two weeks, weekly steps to 52 weeks, and a 52-week hold after removal. Read those three, and you will know exactly how far back your deal file can be pulled before the quarter closes.
Key takeaways
- For OneDrive and SharePoint, the trailing two weeks gives 10-minute steps, so anything broken in that window can be pulled back at fine granularity.
- Beyond two weeks, points thin to a weekly rhythm from 2 to 52 weeks in the past, so the finest landing spot is a whole week back.
- Exchange Online runs on its own figure: a 10-minute recovery point objective, roughly 10 minutes' worth of data lost in a destruction event.
- When a site is removed from the backup policy, its backup is held for 52 weeks from the time each restore point was created for that site.
- A DR copy maintains current state, not historical versions, so it cannot rescue you from a bad overwrite that is already live — reach for the restore point instead.
A quarter is closing, and a forecast folder has gone missing. You need it back, now, not after a ticket. That is the moment your Microsoft 365 Backup plan either saves you or leaves you guessing. Most people glance at the service, see "backup," and assume safety. Then they discover that not every restore point behaves the way they pictured.
There is a second thing to understand before the numbers. A disaster-recovery copy inside Microsoft 365 is different from a backup restore point. A DR copy holds the live state of your content, not a store of older versions. So a wrong save that is already live can still be live in your DR copy. The restore point is what lets you walk backward.
What a restore point actually is
A restore point is a marked moment you can return your content to. In Microsoft 365 Backup, restore points are collected over time so you can choose a healthy state from before the trouble started. Recovery time objective (RTO) covers how fast the data comes back. Recovery point objective (RPO) covers how much you might lose. Microsoft's own summary of the service says it lets you "restore data to a previous healthy state quickly, with fast RTO (recovery time objectives) and short RPO (recovery point objectives)".
Notice that the copy you normally think of — the current, live state — is not a restore point. A DR copy inside Microsoft 365 "maintains the current state of content, not any historical versions from prior points in time" according to the same FAQ. If someone saves over a deal memo this morning, the current state holds the overwrite. To get the memo before that save, you go to a restore point.
For a revenue leader, the practical habit is simple. Ask two questions before you pick a moment. First: how fresh is the newest point I can choose? Second: how coarse are the points as I move back? Those two answers tell you whether your folder will come back clean or with a gap.
This guide reads those two questions against the published figures, so you can judge your own exposure before the quarter closes. For the CRM side of the same problem, see deleted sales record recovery.
Why the quarter makes this urgent
Your deal files and forecast folders change fast at the end of a quarter. Reps drop in final notes. Forecasts get reworked. Someone exports a slide deck and pastes over an earlier draft. Small edits, frequent saves, many hands.
Three pressures collide in this window. First, edits pile up, so the live state is rarely the state you want. Second, the clock is short; nobody has time to rebuild a forecast from email. Third, more people are touching the same folder, which raises the odds of a bad overwrite. A restore point covers the bad-overwrite risk, provided you pick the right one. For the record freeze itself, see forecast snapshot cutoff checks for quarter close.
One more reason to care: if you remove a site from the backup policy, its backup is still held for 52 weeks from when a given restore point was created for that site, according to the Microsoft 365 Backup FAQ. That hold gives you room to change your mind.
So the three figures below are not trivia. They define the window you can reach into, the step size at each depth, and how long a site keeps its history after you drop it.
The three restore-point figures
Figure one: 10 minutes for the trailing two weeks
For OneDrive and SharePoint, the RPO for the trailing two weeks is 10 minutes. That is your worst-case loss inside that window, as stated in the Microsoft 365 Backup FAQ. In practice it means if it is Monday at 8:00 AM, you can step back to any 10-minute period up to two weeks in the past.
This is the figure that matters most at quarter close, because it covers the days when your folder is moving fastest. If a forecast deck is mangled this morning, the restore point from before the damage sits within that same fortnight, at 10-minute granularity.
Two clarifications help. First, this figure is about OneDrive and SharePoint. Second, it is a recovery point objective, not a claim about how the data is stored. A 10-minute RPO does not mean a picture of your site is snapped every 10 minutes. Microsoft's own note on the number starts with: "Let's start with what it doesn't mean: We're not taking snapshots every 10 minutes."
For you: if your deal folder lives in SharePoint or OneDrive, treat the trailing two weeks as your safe zone. Anything you break inside that window can be pulled back at 10-minute steps. Anything older than that sits behind a different ladder.
Figure two: the weekly ladder out to 52 weeks
Past the two-week mark, the step size changes. Beyond two weeks, you can choose any one-week period from 2 to 52 weeks in the past. That is the rule in the Microsoft 365 Backup FAQ.
Read that as a ladder. The trailing two weeks are fine-grained at 10 minutes. Beyond two weeks, the points thin to a weekly rhythm. So a folder damaged three weeks ago can still be recovered, but your best landing spot is a whole week back. A bad paste from week six can be undone, at weekly resolution.
Here is the side-by-side view.
| Content type | Recovery point objective | How far back |
|---|---|---|
| OneDrive and SharePoint | 10 minutes | Trailing two weeks |
| OneDrive and SharePoint | One-week period | From 2 to 52 weeks in the past |
Exchange Online runs on its own figure: the RPO is 10 minutes, meaning the most amount of data that can be lost due to a data destruction event is roughly 10 minutes' worth of data. That line comes from the Microsoft 365 Backup FAQ.
A short planning note for your side of the desk: the ladder above is what the published RPO figures describe.
Figure three: 52 weeks of hold after removal
The third figure is about what happens when your scope shrinks. When a site is removed from the backup policy, the backup of the SharePoint site is held for 52 weeks from the time a given restore point was created for that site, according to the Microsoft 365 Backup FAQ.
That line protects two common scenarios. Scenario one: you trim a site out of policy during a cleanup, then realize a month later that a deal folder on it still needs protection. The history from the point when it was covered remains available for up to a year from each such point. Scenario two: you remove a site, close a quarter, and later face an audit question about a forecast version from before the removal.
The hold is measured from each restore point's creation time, not from the removal date. So a point created early in the coverage period anchors its own 52-week clock. Plan with that in mind if you are sequencing removals.
Read the ladder this way: the trailing fortnight gives you 10-minute steps, the stretch beyond two weeks, out to week 52, gives you weekly steps, and removal from policy does not erase the history — it preserves it on a separate 52-week clock for each point.
How the three figures work together
Picture a revenue leader with three distinct worries, and see which figure answers each.
Worry one: a forecast folder is corrupted this week. The trailing two weeks at 10-minute granularity cover it. Pick a point from before the corruption, restore, and move on.
Worry two: a deal file was quietly overwritten three weeks ago, and nobody noticed until now. The weekly ladder covers it. Your recovery lands on a full week back, which is the finest step available at that depth.
Worry three: a site was dropped from the backup policy last quarter, and a stakeholder asks for the forecast version that existed while it was covered. The 52-week hold covers it, because each restore point keeps its own year-long clock.
Now consider a mixed case. A rep's mailbox is hit by a destructive event, and the related SharePoint folder shows signs of tampering. Exchange Online follows the 10-minute recovery point objective. The folder follows whichever ladder applies based on when the tampering happened. If the tampering is inside the trailing two weeks, the folder step is 10 minutes. If it sits past that, the folder step is a week. Matching the right ladder to the right content type is the whole skill here.
One gap worth naming: Microsoft describes recovery time (RTO) as fast. Confirm your expected recovery time with your Microsoft contact before you promise a timing to your leadership.
A second practical point: a DR copy maintains the current state, so it will not rescue you from a bad overwrite that is already live. Reach for the restore point. The DR copy helps when you need the live state itself, not the past.
Illustrative example: a quarter-close restore
Assume the trailing two weeks rule applies to your folder.
Tuesday, 8:00 AM: a rep pastes the wrong numbers into the shared forecast deck. The live state now carries the error. A restore point from 7:50 AM sits inside the trailing fortnight, at 10-minute granularity. Pick a moment from before the paste. The pre-paste version returns. The rep reopens the file, and you confirm the clean figures.
Thursday: someone asks whether you could get a state from late last quarter. That state sits beyond the two-week mark, so the ladder turns weekly. Your best landing spot is a full week back, at the finest step available at that depth. If the answer matters for a forecast review, plan the review around the weekly step instead of expecting a fine-grained pull.
What to check before the quarter closes
Treat this as a short pre-close review of your own environment. For a companion checklist, see backup and recovery checks.
- Policy coverage: confirm which SharePoint sites and OneDrive accounts sit under your backup policy today.
- Window match: check that your most likely mishaps fall inside the trailing two weeks, where 10-minute steps apply. If a likely mishap sits older than that, expect weekly steps.
- Content type: note which of your deal files live in SharePoint or OneDrive, and which key messages live in Exchange Online. Each follows its own recovery point figure.
- Removal clock: if you plan to remove any site, track each restore point's creation time, because the 52-week hold runs from that moment for each point.
- Target time: write down the moment you intend to restore to before you touch the restore control. A stated target keeps you from clicking into a nearby point and pulling back more than you meant.
- DR copy role: remember the DR copy holds current state. It is not your time machine when the current state is already wrong.
- Recovery time: confirm your expected recovery time with your Microsoft contact before you promise a timing to your leadership.
Run that list while you still have quiet. It is easier on a Tuesday than at quarter close. Start today by checking one deal folder against the list above.
Related guides: Pipedrive deleted products: check the 30-day restore window · Backup frequency: 4 survey shares for forecasts.
Frequently asked questions
Does a 10-minute recovery point objective mean the data is copied every 10 minutes?
No. The figure describes how much data you could lose, not how the data is stored. Your planning rule stays the same: treat 10 minutes as your worst-case gap for Exchange Online, and for OneDrive and SharePoint inside the trailing two weeks.
What happens to restore points if I remove a site from the policy?
Each point keeps its own clock. The backup of the site is held for 52 weeks from the time a given restore point was created for that site. So a point created early anchors its own year-long hold, independent of when you removed the site.
Can a DR copy bring back an overwritten deal file?
Not as a past version. The DR copy maintains current state, not historical versions. An overwrite that is live in the current state will be live in the DR copy too. For a prior version, choose a restore point.
Where does my mailbox sit on the ladder?
Exchange Online runs at a 10-minute recovery point objective, meaning the most data lost in a destruction event is roughly 10 minutes' worth.
How far back can I go for a SharePoint folder?
Inside the trailing two weeks, the steps are 10 minutes. From 2 to 52 weeks in the past, they are weekly. Beyond 52 weeks, ask your Microsoft contact what is available at that depth.
Which figure should I weigh first before the quarter closes?
Weigh the trailing-two-week window first, because that is where your folder moves fastest and where the 10-minute steps give you the finest control. Confirm your likely mishaps land inside it. If they do not, plan around the weekly step instead.