What Sells on iStock: Real Sales Data From 33,513 Photos

Most "what sells on stock" advice is a list of hunches. This is not. Every number below comes from one real portfolio: NoSystem Images, the account behind PixTagger, an exclusive iStock and Getty Images contributor since October 2007. We exported its full sales history from DeepMeta and matched every download to the photo it came from — 33,513 photos, each with its own keyword list and nearly a decade of sales. The short version: photos of people with a problem sell three to four times the average, and photos of pleasant things sell a fraction of it.

How we measured it

Before the results, the method, because it decides what the numbers can and cannot tell you.

• The photos: every creative photo the portfolio submitted between 2017 and 2023 — 33,513 files. Nothing newer, because a photo needs time to sell and a 2025 upload has not had it. Video is left out; it is a different market and gets its own post.
• The sales: every download of those photos from January 2017 to July 2026, through iStock and Getty's regular sales channels. We left out one partner-distribution channel, because a single deal there can swing a whole subject on its own.
• The subject: the keywords stored on each photo. A photo tagged both Psychotherapy and Mental Health counts under both.
• The measure: downloads per photo, shown as an index where the average photo in the portfolio = 100. A subject at 300 sells three times the average per photo; one at 30 sells less than a third of it. We also show the share of photos that ever sold at all.

The subjects that sell, and the ones that don't

The top of the chart is one idea seen from several angles: a person dealing with something difficult. Therapy sessions, money worries, an insurance meeting, a patient with a nurse. Mental-health photos sell at roughly four times the average, and three in four of them sold at least once. The portfolio as a whole manages a little under one in two.

The bottom of the chart is pleasant, decorative, easy to shoot: flowers, gardening, pottery, tea, forests, beaches. Only about a quarter to a third of those photos ever sold, and the ones that did sold rarely. These are not bad photos. They are photos of things that thousands of other contributors have also photographed, and that a buyer can now generate in a few seconds. We made the same argument from the agencies' own revenue figures in how AI split the stock market in two; this is the same split seen from inside one portfolio.

A worried face outsells a smiling one, four to one

This is the result that surprised us most. Stock is famous for smiling people, and this portfolio is no exception: 14,762 of the photos carry Happiness and 10,700 carry Cheerful. Those photos sell below average. Photos tagged Worried, Emotional Stress, Problems or Anxiety sell two and a half to four times above it.

The reason is supply and demand, not mood. A smiling person is the most common thing on the shelf, so each one competes with millions of others. A convincing photo of someone anxious about a bill, or quiet in a therapy session, is much rarer, and an article about debt, burnout or illness needs exactly that picture. A buyer writing about a problem needs a picture of the problem.

One exception is worth noting: Loneliness sits at 81, below average. Mood alone does not do it. What sells is a recognisable situation a buyer needs to illustrate — a consultation, a diagnosis, a stack of bills.

Two to five people in the frame

Photos with nobody in them sell at about half the average, and single portraits a little under it. The best performers have two to five people — a conversation, a consultation, a small team. That is the shape of most of the winners above: a therapist and a client, an adviser and a couple, a nurse and a patient. Larger groups drop back below average.

Professions and places

The same pattern holds when you look at who is in the photo and where it was taken. Roles that come with a client or a patient sell well; roles that are mostly about the person's own craft sell poorly.

What made almost no difference

Some things contributors worry about barely moved the number. Photos tagged Copy Space sold exactly at the average (99). Looking At Camera came in at 83 and Candid at 115 — a modest difference either way. Christmas photos sold at 95: a holiday is not a shortcut, it is just a subject. The subject and the situation decide far more than the technique.

What this data cannot tell you

We would rather you trust these numbers for the right reasons, so here is where they stop.

• It is one portfolio. Nine in ten of these photos show people, and 93% are model-released: this is a people-and-lifestyle portfolio, and landscapes and nature are a small part of it. A dedicated landscape photographer's numbers could look quite different.
• A few hits can carry a subject. For Financial Advisor, 63% of all downloads came from just ten photos. For Psychotherapy it was 37%, spread across photos uploaded on 39 different days over five years. The mental-health result is a pattern; the financial one leans more on a handful of strong images.
• Keywords are not a perfect label. We grouped photos by the keywords stored on them, and a keyword can be on a photo for a minor reason.
• More downloads is not more money. A subscription download and a single licence count the same here. The index tells you what buyers keep choosing, not what each choice paid.

What we would shoot next

If we were planning a new shoot from these numbers alone, it would be small groups of two to four people in a real situation with a problem in it: a consultation, a worried conversation at a kitchen table, an appointment, a difficult phone call, a team under pressure. It would avoid another set of smiling people with nothing happening, and anything a prompt could produce just as well.

One practical point follows from all of this: a problem photo only sells if a buyer can find it. If a therapy shot is tagged only Women, Sofa and Talking, it is hidden from everyone searching for Mental Health, Psychotherapy or Emotional Stress — the terms that sold best in this portfolio. The concept keywords are where the sales are, so they are worth the extra minute.

Sources & further reading

• Data: the NoSystem Images iStock/Getty Images portfolio, sales and portfolio exports from DeepMeta, October 2026. Analysis by the PixTagger team.
• The NoSystem Images portfolio on iStock
• AI images and stock photography: which half of the market is shrinking
• Getty briefs or your own ideas: where to spend your shooting days
• How many keywords should a stock photo have?