Most SharePoint tenants are running version history policies that were set during initial deployment and never revisited. A default of 500 major versions was enabled, people started saving documents, and the version history grew quietly for years. Microsoft's Intelligent Versioning feature is now changing that picture by automatically trimming version history based on age rather than count, and for many admins it will be the first time the platform has touched version data they assumed was frozen in place.
What is Microsoft Intelligent Versioning?
Intelligent Versioning (also referred to as Automatic Versioning mode within SharePoint Online settings) is a versioning management approach that replaces the traditional fixed-count limit with a time-based algorithm. Under the traditional model, a document library keeps the last N major versions per file, where N is a number you set (the default is 500). Under Intelligent Versioning, the number of retained versions is not fixed: instead, the system keeps versions at varying density depending on how old they are.
Microsoft introduced this feature to address a structural problem with the count-based approach: a library set to "keep 500 versions" retains the same number of versions whether those versions span two weeks of active editing or five years of occasional updates. The count-based model does not distinguish between a file that was actively revised and one that has 498 autosave artifacts from a single afternoon of collaborative editing.
You can apply Intelligent Versioning at the tenant level (as a default for new libraries), at the site level, or at the individual library level. Libraries on automatic mode use the algorithm; any library switched to manual mode continues with the traditional count-based approach.
How the automatic trimming algorithm works
The exact behaviour of the algorithm is controlled by Microsoft and may be tuned over time, but the operating principle is consistent: recent versions are retained at high density, and older versions are thinned progressively as they age.
Recent versions: kept at full density
For the most recent period of a file's history, Intelligent Versioning retains all or nearly all versions. This ensures that the recovery scenarios most commonly needed by users (restoring to yesterday's draft, undoing a change from last week) are always available. Beyond this window, the algorithm starts thinning.
Older versions: thinned progressively
As versions age beyond the recent window, the algorithm begins thinning: it keeps one version per day rather than every save, then one per week, then one per month for older content. The result is a version history that has high resolution for the recent past and lower resolution as you go further back. A file actively edited over two years might go from having daily snapshots for the last month, to weekly snapshots for months two through six, to monthly snapshots beyond that.
This is structurally similar to how IT teams manage backup retention: full daily backups for the past week, weekly for the past month, monthly for the past year.
What the algorithm always protects
Intelligent Versioning does not trim every old version indiscriminately. The algorithm always keeps certain versions, regardless of age:
- The current (most recent) published version of the file
- Versions that have been explicitly labelled or promoted as major versions where minor versioning is also enabled
- Versions on files subject to Microsoft Purview retention holds or legal holds
That last point is important for compliance-sensitive libraries: Intelligent Versioning respects active holds. However, libraries not currently under a hold are subject to the full trimming algorithm, including older versions that an admin might have assumed were permanent.
Does Intelligent Versioning touch existing versions?
Yes, and this is the part that surprises most admins who hear about the feature and assume it only applies to versions created after it is enabled.
When a library switches to automatic versioning mode (either because the tenant default changed or an admin explicitly switched the mode), the algorithm evaluates the existing version history and begins trimming versions that fall outside the retention profile. This does not happen instantaneously: Microsoft applies the trimming progressively to manage platform load. But over time, a library that switches from manual 500-version mode to automatic mode will see its historical version count reduce.
Intelligent Versioning is not a backup product. It is a storage optimisation policy that happens to preserve more meaningful history than a naive count limit. Admins who assume the automatic mode retains everything they need may find gaps when they look for a specific version from two years ago that was trimmed as part of the monthly-density phase of the algorithm.
The practical implication: if your current version history contains data you need to retain for compliance, audit, or legal reasons, do not rely on Intelligent Versioning to preserve it. Use Microsoft Purview retention policies to protect specific content, or keep those libraries on manual versioning with an explicit count that matches your retention requirement.
How to check your tenant's current versioning mode
You can inspect versioning mode at two levels:
- Tenant default: in the SharePoint admin center, the Settings section includes options for version storage. This controls the default mode applied to newly created document libraries. Changing the tenant default does not retroactively alter existing libraries that were already created with a different setting.
- Library level: in any document library, go to Library settings and then Version settings. The page shows whether the library is on automatic or manual mode, and for manual mode, the configured maximum version count.
For a tenant-wide audit of which libraries are on which mode, the options are PowerShell (using PnP PowerShell's Get-PnPList with versioning properties) or a reporting tool like ShareMaster's Report Master, which exports library settings including version counts and storage consumed by version history into an Excel workbook without requiring script authoring.
What admins should do before Intelligent Versioning changes your storage
There are four practical steps worth taking now, regardless of whether your tenant has already moved to automatic mode or is still on the traditional defaults.
1. Audit your current version history depth. Run a report that shows version counts and estimated version storage per library across the tenant. This gives you a baseline: which sites carry disproportionate version history, which libraries have files with hundreds of versions, and where the storage is concentrated. You cannot make informed decisions about automatic trimming without this baseline.
2. Identify libraries that need manual control. Any library subject to legal hold, regulatory retention, or a documented data governance policy should be reviewed before its versioning mode changes. Flag these libraries explicitly and either apply Purview retention policies or switch them to manual versioning with a count that reflects your actual retention requirement.
3. Consider a pre-emptive trim. If your tenant is carrying years of accumulated version history under the 500-version default, Intelligent Versioning will eventually trim a significant portion of it. Running a deliberate trim now, with a keep policy you define (for example: keep the last 30 versions per file, delete anything older than 12 months beyond that), gives you control over what gets removed rather than leaving it to the algorithm. It also reclaims storage quota immediately rather than waiting for the automatic process to run.
4. Communicate with stakeholders. Version history is often treated as a passive safety net rather than an active governance responsibility. If users or departments rely on the ability to retrieve documents from one or two years ago, they need to know that the default retention model is changing. Set expectations before the trimming happens rather than explaining the gap after the fact.
Explore the Space Master Version Trimmer
How Space Master works alongside Intelligent Versioning
Intelligent Versioning is Microsoft's automated baseline: it handles the obvious waste (hundreds of near-identical autosave versions of the same document from a single session) without requiring admin intervention. That is genuinely useful. But the algorithm operates on general heuristics, not on the specific needs of your organisation.
ShareMaster's Space Master Version Trimmer gives admins explicit, policy-driven control on top of whatever the automatic mode does. The Version Trimmer lets you:
- Audit version counts and ages across every library in the tenant, not just per-site summary figures
- Define a keep policy at the library level: keep the last N versions per file, delete versions older than X days beyond that, always protect the current version
- Run the trim across hundreds of libraries in a scheduled operation, with a preview before execution
- Export before-and-after storage figures to document the reclaimed quota for stakeholders
For libraries where the automatic algorithm is too aggressive or not aggressive enough, the Version Trimmer provides a manual override that does not require PowerShell scripting. For migration preparation, running a targeted trim before a cross-tenant copy reduces the data volume transferred and eliminates the version accumulation debt in the destination from day one. See the storage cost analysis for the numbers on why pre-migration trimming matters.
Frequently Asked Questions
What is Microsoft Intelligent Versioning in SharePoint Online?
Intelligent Versioning (also called Automatic Versioning mode) is a SharePoint Online feature that replaces the traditional fixed-count version limit with a time-based algorithm. Instead of keeping the last N versions regardless of age, it retains versions at higher density for recent content and trims older versions progressively to reduce storage overhead automatically.
Does Intelligent Versioning delete versions I might still need?
Potentially. The algorithm removes versions that are statistically unlikely to be retrieved, but it does not understand your specific compliance or legal hold requirements. Libraries subject to retention policies, audit obligations, or litigation holds should be evaluated before the automatic mode takes effect. Those libraries may need to remain on manual versioning with an explicit count or a Purview retention policy applied to protect the relevant content.
How do I turn off Intelligent Versioning for a specific library?
Open the library in SharePoint Online, go to Library settings, then Version settings. Switch the mode to Manual and set a specific major version count. This overrides the tenant-level default for that library and prevents the algorithm from automatically trimming its version history. Changes apply to that library only; other libraries retain whatever mode they are on.