Passing review and being findable are two different outcomes. Approval only means an asset is permitted in the catalogue; appearing in a buyer's results also requires that it has been indexed, that your keywords match what buyers actually type, that the attributes their filters rely on are filled in, and that the licensing status does not exclude it from a commercial search.
Accepted is not the same as indexed
Contributor dashboards report a binary: pending, accepted or rejected. That status describes moderation, not discoverability, and the two run on separate tracks. An asset marked accepted has cleared the content and technical rules. Whether it has been written into the search index, and where it sits once it has, depends on processing that the contributor side of the platform does not expose.
This is why the first question to ask is not what is wrong with the file but how long ago it was approved. Searching for your own brand-new asset by an exact phrase and finding nothing is frequently a timing observation rather than a metadata finding, and acting on it as though it were a metadata problem is how contributors end up rewriting keyword sets that were working.
From approval to findable
The stages an accepted asset passes through before it competes in buyer search, and which of your own metadata decisions matters at each one, with no agency-specific timings implied.
Six reasons an accepted file stays invisible
1. It has not been indexed yet
The simplest explanation and the most common. Rule this out before anything else by searching for an asset you uploaded in the same batch several weeks earlier. If the older file is retrievable by an exact phrase and the new one is not, you are looking at processing rather than metadata, and the correct action is to wait.
2. Your leading keywords are not the ones that matter
Keyword order is not cosmetic on every platform. On Adobe Stock the first ten keywords are the decisive ones for ranking, which means an asset with its most specific terms sitting at positions fifteen to twenty is effectively described by its weakest words. Shutterstock takes a different approach, accepting between 7 and 50 keywords with a title of up to 200 characters, so the same file needs a different metadata shape at each destination. A keyword set exported identically to both is optimised for neither.
3. Buyer filters are excluding it
This is the cause contributors most often miss, because it is invisible from the contributor side. Buyers rarely search a bare phrase; they search and then narrow by orientation, licence type, content type, and whether people appear. Each of those filters reads an attribute on your asset, and an asset with that attribute unset is dropped from the filtered result set no matter how precise its keywords are. A file can be perfectly keyworded, fully indexed and still absent from every search a real buyer runs.
4. It is editorial, and the buyer is searching commercial
Editorial assets are excluded from commercial searches by design. If an identifiable person appeared without a model release, or recognisable private property without a property release, or a visible brand or logo, the asset may well have been accepted as editorial rather than refused. That is a successful submission with a much smaller audience, and nothing about the status in your dashboard announces the difference. Check how each quiet file is classified before you touch its keywords.
5. It is competing with near-identical files, including your own
Agencies actively avoid filling a page of results with variations of one shot. If you uploaded eight frames from the same setup with the same keyword set, you have not bought eight chances at the top of the results; you have made eight files compete for one slot, and seven of them will look invisible. The related rejection-side version of this problem, where the near-duplicates do not get in at all, is covered in images rejected for similar content.
6. The keywords are in the wrong language or the wrong form
A buyer searching in English will not match a keyword set written in another language, and some marketplaces reject non-Western characters outright rather than transliterating them. If your keywording starts in your own language, the translation step is part of the delivery and not an optional polish. How this plays out field by field is in accented characters and non-English keywords.
Why can't your accepted file be found?
A diagnostic that separates an indexing delay from a licensing exclusion, a missing filter attribute, a near-duplicate cluster and a keyword-order problem, using checks you can run from the buyer side.
What to check, in order
The order matters, because each step rules out a whole class of explanation and saves you from editing something that was never broken. Start by confirming an older asset is retrievable, which separates a platform-side delay from an asset-side fault. Then check the licensing classification, since editorial status explains the symptom completely and no keyword work will change it. Then search as a buyer would and apply the filters a buyer would, because a missing attribute looks exactly like a keyword failure and is fixed differently. Only after those three come back clean is it worth opening the keyword set.
Searching as a buyer rather than as the contributor is the single most useful habit here. Your own dashboard shows you the asset you know exists; the public search shows you the asset as the market sees it, filtered, ranked and competing. Many visibility problems are obvious within two minutes on the buyer side and invisible from the contributor side indefinitely.
What contributors use to run this down
There is no tool that reports why a specific asset is not surfacing, because no agency publishes that. What tools can do is answer the narrower questions: what metadata actually shipped, how the asset looks from the buyer side, and how to fix a set of files once you know what is wrong.
The agency's own public search
Best for the diagnosis itself, and consistently underused. It is the only view that shows your asset ranked against competitors and subject to real filters. Pricing model: free. Main limitation: it is manual, it tells you nothing about why, and results can be personalised or regional, so treat a single search as a signal rather than proof.
Contributor dashboard analytics
Best for distinguishing no impressions from impressions without downloads, which is the difference between a visibility problem and a conversion problem. Pricing model: included with your contributor account where it exists. Main limitation: depth varies enormously between agencies, and several report sales without reporting search exposure at all. Getting the most out of what is reported is the subject of which keywords drive your downloads.
ExifTool and Adobe Bridge
Best for confirming what metadata is actually embedded in the file you sent, as opposed to what your keywording tool displayed. More visibility problems than you would expect come from fields that never made it into the delivered file. Pricing model: ExifTool is free and open source; Bridge comes with a Creative Cloud subscription. Main limitation: they read and write your local files and know nothing about what the agency received or how it indexed it.
Xpiks
Best for reworking keyword sets across many files with the library in front of you, including reordering terms so the specific ones lead. Pricing model: a paid desktop licence, no revenue share. Main limitation: you supply the judgement about what each agency wants, since the tool does not hold per-marketplace rules.
StockSubmitter
Best for pushing corrected metadata back out across a long list of agencies once you know what the fix is. Pricing model: paid desktop software, no revenue share. Main limitation: a Windows-centric workflow, and like the others it will deliver a keyword set that is wrong for a given marketplace just as faithfully as a right one.
Rastock AI
Best for the cause in reason two: producing a different metadata shape per destination rather than one set copied everywhere, with the leading keywords ordered for the platform that weights them and the field limits of each marketplace respected. Delivery runs over FTP or SFTP to 10+ agencies with per-file status, so a file that never arrived is distinguishable from one that arrived and is not surfacing. There is no revenue share and no ownership claim, and IPTC and CSV export stay open. Pricing model: a subscription with a 14-day free trial and no card required. Main limitation: it works on metadata and delivery, so it cannot tell you anything about an asset's ranking once the agency has it, and it will not fix a visibility problem caused by editorial status or near-duplicates.
When to use which
Start with the public search every time, because it costs nothing and eliminates the two most common explanations. Go to dashboard analytics when you need to know whether the asset is being seen and ignored rather than not seen at all. Reach for ExifTool or Bridge when you suspect the metadata you think you sent is not the metadata that shipped. Use Xpiks or StockSubmitter when the diagnosis is done and the job is applying and redelivering a fix at volume. Use a policy-aware pipeline when the recurring cause is that one keyword set is being sent to marketplaces that weight and limit keywords differently.
One habit is worth more than any of them: before concluding that an asset has a metadata problem, confirm that it is indexed, that it is commercial rather than editorial, and that its filter attributes are set. Those three checks take a couple of minutes and account for most of the files contributors describe as invisible.
If the asset does appear on the terms buyers use and still earns nothing, the problem has moved from visibility to demand, which is a different diagnosis entirely and is covered in why your stock photos are not selling. If the quiet files turn out to be an older batch with weak metadata across the board, the question becomes which of them are worth the work, and that is the subject of re-optimising an old stock portfolio.
Rastock AI exists for the part of this that repeats: the same file needs a different keyword shape at each marketplace, and nobody wants to maintain that by hand. Metadata is generated against each agency's own field rules and policy, banned and mandatory terms are checked before delivery, every batch can be rolled back, and files go out over FTP or SFTP to 10+ agencies with per-file status. The features page shows how the per-marketplace rules work, and the pricing page lists the plans. The trial is 14 days and needs no card.
Frequently asked questions
How long after approval should an image appear in search?
Agencies do not publish a figure, and it varies by platform and by load, so treat any specific number you read as unreliable. The useful test is comparative rather than absolute: search for an asset from a batch you uploaded several weeks earlier. If the older file is retrievable by exact phrase and the new one is not, you are almost certainly looking at indexing rather than a metadata fault.
Why does my image appear when I search the exact title but not otherwise?
That means it is indexed, so the issue is relevance or filtering rather than visibility. Either your keywords do not match the words buyers type, or a filter attribute such as orientation, content type or people is unset and the asset is being dropped from narrowed result sets. Check the attributes before rewriting the keywords, because the two failures look identical and are fixed differently.
Does keyword order actually matter?
On some platforms, yes. Adobe Stock treats the first ten keywords as decisive for ranking, so an asset whose most specific terms sit near the end of a long list is effectively described by its weakest words. Shutterstock accepts between 7 and 50 keywords with a title of up to 200 characters, so the same file genuinely needs a different metadata shape at each destination.
Could my file have been accepted as editorial without my noticing?
Yes, and it is a common explanation. If an identifiable person appeared without a model release, or recognisable private property without a property release, or a visible brand or logo, an agency may accept the asset as editorial rather than refuse it. Editorial content is excluded from commercial searches by design, so the file is live, findable by editorial buyers, and absent from the searches most people run.
Should I re-upload the file to force it back into the index?
No. A duplicate upload risks a similar-content rejection and, if both copies survive, makes two of your own files compete for the same slot. Edit the existing asset's metadata if the diagnosis points there, and otherwise wait. Re-uploading treats a timing problem as a content problem and usually leaves you worse off than doing nothing.
Why are only some of my images from one shoot visible?
Because agencies avoid filling a results page with variations of the same shot. Several near-identical frames carrying the same keyword set do not give you several chances at the top; they compete for one position and the rest look invisible. Pick the strongest frame, differentiate the keyword sets of the others so they answer genuinely different searches, and consider withdrawing the weakest.