When a marketplace rewrites a rule, nothing happens to your existing files on the day of the change. The risk arrives later, when a resubmission, an edit or a fresh moderation pass puts an old asset in front of the new policy. Re-checking a portfolio is mostly a scoping exercise: find what changed, find which files touch it, fix those in batches.
Why rule changes break portfolios quietly
Agencies almost never re-review an entire catalogue after a policy update, because the cost of doing so would be enormous. What they do instead is apply the new rule at every point of contact: new submissions, resubmissions of previously rejected files, metadata edits, and whatever subset of the library gets pulled into a periodic compliance sweep. That means a portfolio can sit in a non-compliant state for months without a single notification, and then fail in a cluster the moment you start editing keywords across it.
The changes that cause this fall into three groups, and they need different responses. Metadata rule changes alter what a compliant keyword set or title looks like, which is a bulk-edit problem. Content policy changes decide whether an asset is allowed at all, which is a triage problem with no edit that fixes it. Technical requirement changes move a format or resolution floor, which usually means re-exporting from masters rather than touching metadata.
Step one: establish exactly what changed
Most portfolio damage done in response to a rule change comes from acting on a summary rather than the rule. Second-hand accounts of agency policy are frequently wrong in the details that matter, and the details are what determine how many of your files are affected. Before you edit anything, pin down four things from the agency's own contributor documentation.
- Whether the rule is retroactive. Some changes apply only to content submitted after a date; others apply to the whole catalogue. This single question decides whether you have a project or nothing to do.
- Which content types it covers. A rule written for photographs often says nothing about vectors or footage, and a rule about generative AI may treat AI-assisted editing differently from fully generated work.
- What the enforcement looks like. Being delisted from search is a different problem from being removed from the catalogue, and a flag on a declaration field is different again.
- Whether the change is a limit rather than a prohibition. Adobe, for example, has told contributors that submissions pause temporarily once a weekly submission limit or the cap on content pending moderation is reached. A throughput limit changes how you schedule a re-upload campaign; it does not make existing files non-compliant.
Both patterns are visible in live documentation. Adobe's contributor announcement on submission limits describes a throughput pause rather than a content ban, while Motion Array's published upload rules state outright that AI-generated images are not accepted as stock photos. A contributor who read a general summary about agencies accepting AI stills, and applied it to every marketplace they deliver to, would have a compliance problem at the second one and no idea why.
The re-check sequence, step by step
Five stages for auditing a portfolio against a rule you did not make: read the primary source, classify the change, scope the affected files, fix in batches, then record what you changed and why.
Step two: scope the exposure from your own records
The instinct is to start clicking through the contributor dashboard. That is the slowest possible route, and on a large portfolio it is also the least reliable, because agency dashboards rarely let you filter on the things a rule change actually turns on. The faster path is to answer the question locally, from your own delivery records and the metadata embedded in your masters, and only then go to the agency to confirm a sample.
Four filters account for most rule changes. Submission date separates content delivered before and after a cut-off. Content type separates photo, vector, footage and template, since rules are usually written per type. Generation method separates camera work from AI-generated and AI-assisted files, which is the division most policy changes of the last few years have turned on. Release status separates assets with a model or property release on file from those that can only run as editorial.
If you cannot answer the generation-method question from your own records, that is the finding, and it is worth more than any individual fix. A portfolio that does not track how each asset was made cannot be audited against a declaration rule at all. The field-level requirements for that declaration differ by marketplace, which we broke down in metadata for generative AI images and agency rules.
Does this rule change affect your files?
A short diagnostic that routes a rule change to the right response: no action, a metadata bulk edit, re-export from masters, editorial re-filing, or removal with no edit that saves the asset.
Step three: fix in batches, and verify the first one
The failure mode here is applying a portfolio-wide edit before anything has been confirmed. If your reading of the rule is slightly wrong, a batch of fifty files tells you that cheaply and a batch of five thousand tells you expensively. Take the smallest affected bucket, fix a sample, push it, and wait for the review outcome before touching the rest. Keep the previous values for every field you overwrite, because a rollback path is the difference between a bad afternoon and a lost portfolio.
What contributors use for the re-check itself
No single tool does all of this well, and the right choice depends on whether your bottleneck is editing, delivering or knowing the rules. These are the options contributors actually reach for, with what each is good at and where each stops.
Adobe Bridge and Lightroom Classic
Best for editing embedded metadata across a large local library with a visual interface, and for applying a keyword change to a selection in one pass. Pricing model: bundled into a Creative Cloud subscription you most likely already hold. Main limitation: they know nothing about agency policy. They will happily write a keyword set that a marketplace rejects, and they have no concept of which agency a file is destined for.
ExifTool
Best for auditing rather than editing. A single command can report which of ten thousand files carry a given field, which is exactly the scoping question step two asks, and it can export that into a spreadsheet you can reason about. Pricing model: free and open source. Main limitation: it is a command-line tool with no safety net, no agency knowledge, and no undo beyond the backups you remembered to make.
Xpiks
Best for contributors who want a desktop application built specifically for microstock keywording and upload, with the whole library in front of them. Pricing model: a paid desktop licence rather than a royalty share. Main limitation: it is driven by you. It gives you better tools for applying your judgement about a rule, but it does not hold the rule.
StockSubmitter
Best for breadth of delivery. It is the tool most often named when contributors talk about pushing the same batch to a long list of agencies, and for a re-upload campaign after a technical rule change that breadth is the point. Pricing model: paid desktop software, no revenue share. Main limitation: a Windows-centric desktop workflow, and the same gap as the others on policy awareness.
Wirestock
Best for contributors who would rather not own the compliance problem. It is a managed pipeline: you submit to Wirestock, and it handles distribution and much of the agency-facing detail. Pricing model: it takes a share of the royalties your work earns. Main limitation: that share is permanent for the files that go through it, and you give up direct control over how each asset is presented to each marketplace, which is the lever a rule change makes you want.
Rastock AI
Best for the case this article describes: generating metadata against each marketplace's own policy, flagging banned and mandatory terms before anything is delivered, and keeping a per-file record you can roll a batch back from. Delivery is over FTP or SFTP to 10+ agencies with per-file status, and there is no revenue share, no ownership claim over your files, and IPTC and CSV export stay open. Pricing model: a subscription, with a 14-day free trial and no card required. Main limitation: it is a metadata and delivery layer, not an editor. If your problem is that your masters fall below a new resolution floor, nothing here helps and you are back in your raw processor.
When to use which
Use ExifTool when the question is how many files are affected, because scoping is a reporting job and nothing does reporting across a large library more cheaply. Use Bridge or Lightroom when the fix is a visual judgement on a few hundred files and you want to see what you are changing. Use Xpiks or StockSubmitter when the fix is mechanical and the bottleneck is getting corrected files back out to many agencies. Use a managed service like Wirestock if you would rather trade royalty share for not thinking about this again. Use a policy-aware pipeline when the rule changes often enough that the knowledge, not the editing, is the hard part.
In practice most working contributors end up with two of these rather than one: something that reports on the library, and something that delivers to the agencies. The mistake is assuming the delivery tool also understands the rules, which is how a compliant batch at one marketplace becomes a rejected batch at the next.
Keep a record, or you will do this twice
The part everyone skips is logging what was changed and why. Six months later, when a file comes back rejected, the only way to tell whether you are looking at a new problem or the one you thought you fixed is a record of the edit, the previous value and the rule that prompted it. The mechanics of doing that safely across a large library are in our guide to bulk editing keywords on a large archive, and if the re-check turns into a broader cleanup, re-optimising an old portfolio covers what is worth the effort and what is not.
Rastock AI was built around the assumption that marketplace rules move and your portfolio does not. Metadata is written against each agency's own policy, banned and mandatory terms are checked before delivery rather than after a rejection, every batch is reversible, and files go out over FTP or SFTP to 10+ agencies with per-file status. The features page shows how the policy checks and rollback work, and the pricing page lists the plans. The trial runs 14 days and does not ask for a card.
Frequently asked questions
Will an agency re-review my whole portfolio after a rule change?
Almost never, because the cost would be prohibitive. What happens instead is that the new rule gets applied wherever a file is touched: new submissions, resubmissions, metadata edits and periodic compliance sweeps. That is why a portfolio can sit non-compliant for months with no notification and then fail in a cluster as soon as you start editing across it.
How do I know whether a rule change is retroactive?
Read the agency's own contributor documentation rather than a summary, and look for language about content already in the catalogue. If it does not say, treat it as retroactive but do not bulk-edit yet. Ask contributor support in writing and keep the reply, because a written answer is what justifies either editing or leaving thousands of files alone.
Should I delete files that no longer comply?
Not as a first move. Check whether a permitted route still exists, such as re-filing the asset as editorial or supplying a release that keeps it commercial. Also check the other marketplaces you deliver to, since policies differ and content refused at one agency is often still accepted at another. Withdrawal is the last option, not the default.
What is the fastest way to find which files are affected?
Query your own records rather than the agency dashboard. Filter by submission date, content type, generation method and release status, since those four dimensions cover most policy changes. A command-line metadata reader such as ExifTool can report across tens of thousands of files in one pass, which no contributor dashboard will let you do.
Can I bulk-edit the whole portfolio at once?
You can, and it is the most common way a rule change turns into real damage. If your reading of the rule is slightly off, a fifty-file batch tells you cheaply while a five-thousand-file batch does not. Fix a sample, confirm it cleared review, then run the rest in batches, keeping the previous field values so a rollback stays possible.
Do different agencies interpret the same rule the same way?
No, and assuming they do is the main source of cross-agency rejections. Rules are written per content type and per marketplace, so a declaration that satisfies one agency can be incomplete at another, and content accepted at one may be refused outright elsewhere. Scope and fix per agency rather than applying one interpretation across every destination.