One Photo, Two Agencies: The Keyword Lists Agree on 28%

If you sell the same photograph on Getty and on Adobe Stock, the tempting move is to write one keyword list and paste it into both. We can now say what that costs, because we measured it: across 48 images that were tagged for both marketplaces in our own system, the two finished lists share 28% of their keywords. The picture is identical. Roughly seven keywords in ten are not.

What we measured

PixTagger asks which marketplace a batch is for before it tags, and tags differently depending on the answer. That means some files have been through the pipeline twice — once aimed at Getty, once at Adobe — and both results are stored. Those pairs are a controlled experiment we did not set out to run: same photograph, same photographer, same shoot notes, same models, one variable changed.

We found the pairs by matching files on their content hash rather than their name, so a re-upload under a different filename still counts as the same picture. 48 images qualified. For each one we compared the finished keyword list on the Getty side against the finished list on the Adobe side.

The lists are almost exactly the same length. That is the part that fools people. Two lists of about forty terms, describing one photograph, look interchangeable at a glance — and they are not.

Why one picture produces two lists

The divergence is not the tagger being inconsistent. It follows from a structural difference between the two marketplaces that most contributors know about in the abstract and underestimate in practice.

• Getty and iStock keyword against a controlled vocabulary. A fixed list — the copy we ship carries 24,631 terms — and a term outside it is not a keyword at all. You are not describing the picture in your own words; you are choosing from a menu.
• Adobe Stock takes free text. No menu. Adobe caps the list at 49 keywords, its own guidance points at roughly 15–35 as the useful range, and it states plainly that the first ten carry the most weight in search placement. What matters is matching the words a buyer types.

So the two systems reward opposite instincts. Getty rewards finding the official term for a thing. Adobe rewards finding the phrase a customer would search for. Where those two happen to be the same string, the keyword survives on both lists. Where they differ — which is most of the time — you get two entries that mean the same thing and match each other nowhere.

The same idea, twice, in two spellings

Here are the most frequent terms that appeared on only one side across those 48 images. We checked every left-hand term against the Getty vocabulary and every right-hand term as well: in each row the Getty form is in the controlled vocabulary and the Adobe form is not.

Look at the first and last rows. Body Building and Bodybuilding differ by a single space, and Women and Woman by a single letter. A human reading both lists would call them the same list. A search index matching strings would not.

The detail that surprised us most

The bare word "Fitness" is not in the Getty controlled vocabulary. Five compounds built on it are — Fitness Ball, Fitness Instructor, Fitness Tracker, Fitness Trail, Fitness Boot Camp — but the plain noun, the most obvious keyword in the world for this shoot, is not a term you can use. On the Adobe side it appeared on 21 of the 48 lists, exactly as you would expect it to.

The same trap in reverse: Gym is in the vocabulary, so it is one of the terms that survives on both lists — but Gym workout is not. Compounding a valid term with an ordinary English word almost always takes you outside the vocabulary. This is the same family of problem as the terms Getty refuses outright, which we covered separately in keywords Getty refuses even from its own vocabulary.

One more, because it is expensive to get wrong

The Getty vocabulary contains Women, Men, Females and Males — and no singular forms at all. It is tempting to read the plural as a headcount and leave it off a solo portrait. It is not a headcount; in this vocabulary those are the gender terms, and the count is carried by separate terms. Drop the plural from a picture of one woman and you have not made the list more accurate, you have removed the only term that says who is in the frame.

What the two lists do agree on

The 28% that survives is not random. It is the terms where the official vocabulary form happens to also be the phrase a buyer would type:

People · Gym · Healthy Lifestyle · Adult · Strength Training · Athlete · Sport · Concentration · Weight Training · Determination

Plain, single-concept nouns. The moment a concept needs two words, or a grammatical number, or a house style, the two systems part company.

And here is the number with the sharpest practical edge. Adobe weights the first ten keywords above all the others. We took the first ten keywords of each Getty list and asked how many appear anywhere in the matching Adobe list — not in its top ten, anywhere in all 40 terms.

3.9 of 10.

If you paste your Getty list into Adobe, six of the ten slots that decide your Adobe search placement are spent on terms the Adobe list would not have contained at all. The list is not slightly wrong. The weighted part of it is mostly wrong.

Age: the trap that looks like it is already solved

Getty's vocabulary describes age with numeric ranges — 20-29 Years, 30-34 Years, 55-59 Years. These are correct, they are what Getty wants, and they are what a well-built Getty list contains.

No Adobe Stock customer has ever typed "30-34 Years" into a search box. They type young woman, or mature adult. Adobe takes free text, so the range will be accepted — it simply will not be found. A keyword that is accepted and never matched is a slot spent for nothing, and on a 49-slot list with ten weighted places, slots are the scarce resource.

This one is quietly fixable, which is why we fixed it in the exporter rather than leaving it to the reader: on the Adobe and Shutterstock CSV exports, numeric ranges are translated to the descriptive bucket the marketplace uses — Young Adult, Mid Adult, Mature Adult, Senior Adult — while the Getty and qHero exports keep the ranges untouched. Worth knowing whether or not you use our tool, because if you keyword by hand it is a translation you have to make yourself, on every file.

What to do if you sell on both

1. Stop treating one list as portable. The overlap is 28% on our sample and the weighted top ten overlaps less than that. Reusing a list is not a shortcut with a small penalty; it is two different jobs done once.
2. Decide the marketplace before you keyword, not after. The target changes which term is correct, so it cannot be applied at the end. A Getty list trimmed for Adobe is still a Getty list.
3. On the Getty side, bind every term to the vocabulary before you export, not after a rejection — and remember that being in the vocabulary does not by itself mean a term will be accepted.
4. On the Adobe side, spend the first ten on searchable phrases and bring the words from your title into them. Our guide to Adobe keyword order covers why those ten places behave differently from the other 39.
5. Write the title twice too. Only 11 of our 48 pairs ended up with the same title, and that is the right outcome, not a bug: Adobe titles feed the URL and the top-ten keyword bump, while Getty titles sit alongside a controlled vocabulary that is already carrying the descriptive load.

Sources & further reading

• Adobe Stock — Title and keyword tips (first ten keywords carry the most weight)
• Getty Images — Preparing your content (contributor login required)
• How the Getty controlled vocabulary works
• Keywords Getty refuses — even from its own vocabulary
• Adobe Stock keyword order: why the first keywords decide
• How many keywords should stock photos have?
• Stock agency metadata formats: CSV columns & keyword limits