State of the photography business 2026: sourced only

Most state of the industry articles run on a survey that either does not exist or was paid for by whoever wants you to buy something. This one uses figures you can click through to, and it is shorter for it. Where a number has never been measured, that is said plainly, because knowing which numbers do not exist turns out to be the more useful half.

How this was put together

Three rules, and they are the reason this piece is thinner than most of its competitors.

First, every figure here links to the organization that produced it, with the year attached. Second, where the only available figure is a vendor's own marketing, it is named as marketing rather than presented as a finding. Third, where nothing has been measured, the article says so instead of estimating, rounding, or borrowing a number from an article that borrowed it from somewhere else.

Applied honestly, those rules leave four solid numbers and one clear change in the tools. They also leave a list of things about this trade that nobody has ever counted, which is section six, and which is probably the most useful part of this page. If you are looking for the survey where four hundred photographers report their burnout levels and their AI adoption rates, it does not exist. It gets quoted constantly anyway.

Who actually runs a photography business

The US Bureau of Labor Statistics counted 151,200 photographers in 2024, with median pay of $42,520 a year, or $20.44 an hour. About 66 percent are self-employed. Projected growth is 2 percent from 2024 to 2034, which is 2,800 additional jobs and slower than the average across all occupations.

Read the self-employment figure first, because it changes how you read everything else. Two thirds of this trade has no employer, no payroll department and nobody to cover a sick Saturday. When an article says the industry is consolidating or the industry is investing in AI, it is really making a claim about a hundred thousand sole traders who each decide alone, on a Tuesday, whether to pay ten dollars a month for something.

It also explains why the median is so slippery. That single figure covers the weekend shooter grossing eight thousand dollars and the studio owner clearing six figures, and it is a wage statistic in a population that mostly sets its own prices. We pulled it apart properly in how much photographers make. The structural consequence for how you actually run the week is in solo photographer versus studio.

Two percent growth over a decade is the other number worth sitting with. This is not a collapsing trade and it is not a booming one. It is flat, which means share is taken from other photographers rather than handed out by a growing market.

What the work pays, where a figure exists

For weddings there is one genuinely large dataset. The Knot's Real Weddings Study, based on 10,474 US couples married during 2025, puts average spend on a professional wedding photographer at $3,000. The quartiles are more useful than the average: $1,500 at the lower end, $2,800 in the middle, $4,700 at the upper end.

Note what that measures. It is what couples reported spending, not what photographers invoiced, and certainly not what anybody kept. Set it beside the BLS median and an obvious question appears: if the average wedding pays $3,000 and the median photographer earns $42,520, then either most photographers are not shooting anything like a full wedding calendar, or a large share of what they charge does not survive costs and tax. Both are true for somebody. No public dataset separates them.

Five rows is the whole reliable inventory for this trade. The fuller breakdown of the wedding figures sits in wedding photography statistics, and what to do with them when you set your own prices starts there too.

The time split everyone quotes is from 2020

You have seen the figures: 4 percent of a photographer's working time spent shooting, 55 percent editing, 18 percent on admin, 11 percent picking photos, 7 percent on communication. They come from a single survey, run by Your Perfect Wedding Photographer among more than 300 UK wedding photographers and reported by PetaPixel in February 2020. The same respondents averaged 28 weddings a year at a starting full-day rate around £1,590.

It is a good survey and the shape of it still rings true, which is why we cite it ourselves in where a photographer's time actually goes. It is also six years old, UK only, and was run before any of today's selection tools were in common use. Nobody has repeated it.

So when you read that photographers spend 11 percent of their time culling, including when you read it on this site, understand what you are being handed: a 2020 UK reading carried forward because it is the only one there is. It is the last known measurement, not the current one. If the tools have done what their vendors say, that share should have fallen. Nobody has checked.

The practical response is to measure your own, which takes one shoot and a timer. The method is in how to audit your own culling time.

The one clear change in 2026: selection came bundled

If you want a single verifiable shift in this business over the past year, it is not demand and it is not pricing. It is that Adobe started shipping photo selection inside a subscription most photographers already hold.

Assisted Culling is now in Lightroom Classic and cloud Lightroom, labeled early access. In Classic you enable it under Catalog Settings, then Metadata. It scores Subject Focus, Eye Focus and Eyes Open, and it rejects misfires, exposure problems, documents and receipts. Version 15.4 in June 2026 extended it to run during import. Adobe states it is optimized for portraits and headshots.

Prices checked August 2026 from each vendor's own pricing page.

Two things follow. The line that Lightroom cannot detect a blink or a missed focus is now simply false, and any article still running on it was written before June. And a standalone selection subscription is now priced against a feature bundled into a plan you are probably already buying, which is a harder sale than it was a year ago. That does not make the dedicated tools redundant: they are faster on high volume work, better with long bursts, and none of them carry an early access label. Our test of Adobe's version is in the Lightroom Assisted Culling review.

Six numbers that do not exist

These get quoted in this trade constantly. None of them has a source you can open.

How photographers' time splits in 2026. One UK survey from 2020, never repeated. Everything more recent traces back to it.

How many working photographers use AI selection. Vendors publish user counts in their own marketing material. There is no independent census, no trade body collecting it, and no way to check any of the figures being cited.

How accurate any selection tool is. FilterPixel publishes accuracy comparisons against rival products. Those come from its own testing and appear on its own marketing pages, and there is no neutral benchmark to check them against. When you see a percentage attached to a competitor's accuracy, you are reading a competitor's estimate of them. The only accuracy test that means anything is the one you run on a shoot you have already picked by hand.

Average gallery turnaround. Asserted everywhere, measured nowhere. Trace any four to six week industry standard back through its citations and it ends in a blog post citing another blog post.

Burnout prevalence among photographers. Occupational burnout is a well studied field with real instruments behind it. Photographers are not one of the populations that has been studied with them. Anybody quoting a percentage of burned out photographers made it up or repeated somebody who did.

What share of revenue survives as profit. The rule of thumb that a lean solo business nets 50 to 65 percent of gross is exactly that, a rule of thumb passed between photographers. There is no public dataset behind it for this trade.

None of this means the underlying ideas are wrong. Editing probably does eat the majority of the week. Selection tools probably do save time. It means the decimal points attached to those ideas are decoration. Ask for the link every time, and notice how often there is not one.

Burnout has no statistic, and shows up in the same places anyway

Since there is no prevalence figure, treat what follows as a description of a pattern rather than as data.

It rarely announces itself as exhaustion. It shows up first in operations: a backup skipped because it was late, an invoice that sits unsent for three weeks, retouching that drifts between galleries, a gallery past the date you promised, and a specific reluctance to open the catalog at all. Those are process signals, which is useful, because process signals can be counted and feelings cannot.

So set a threshold you can actually check. Three shoots sitting unpicked, or any gallery past its promised date, is a reasonable line. Crossing it triggers something specific and decided in advance: a week with no new editing commitments, a subcontracted edit, or a mini session weekend that quietly does not get scheduled. The value is that the decision was made when you were rested. The longer version of this argument is in the photographer burnout guide.

Five numbers to keep on yourself this season

The industry figures do not exist at the resolution you need, and they never will, because nobody is funding the survey. Your own numbers are cheap to collect and they answer the questions the industry ones only gesture at.

1. Total hours per job, from first inquiry to delivered gallery, including the ten minute tasks. Five jobs gives you a median worth pricing against.
2. Your keep rate, delivered frames divided by frames captured. It tells you whether your problem is at the camera or at the desk.
3. Days from shoot to gallery, measured on every job, not remembered. This is the promise you are selling, so it should be the number you know best.
4. Tool cost per booking, your annual software total divided by jobs booked last year. It reframes every subscription decision in about ten seconds.
5. Weeks with nothing outstanding. Count them across the year. If the answer is under four, no pricing change is going to fix how the season feels.

On our own product, the same rules apply. Kepla's picking app for iPhone, iPad and Mac is still being built, so we have no time-saved figure to publish, and we will not publish one until the app exists and somebody has actually measured it. What is live today is the booking page, which is free while we build the rest. If a company will not tell you which of its numbers are measured and which are hoped for, that is the number worth noticing.

FAQ