A small brand that asks a freelance photographer to quote twenty product images usually gets one number back, and no way to interrogate it. The quote is real work priced fairly. It is also opaque, which makes it very hard to answer the question underneath: is this the right way to spend two thousand dollars right now, or is it the way it has always been done.
We run Kloti, so we have an obvious interest in the answer and you should read this with that in mind. What follows is deliberately about the other side of the comparison: what a shoot costs in money, what it costs in calendar, how many attempts either route really takes, and the places where generation does not help at all.
How to read the numbers here
Three labels, used consistently.
- Measured means an invoice or a run somebody can point at, with a date.
- List means a figure somebody else published and stands behind: a rate card, a price page, a survey, a press release. We have not reconciled it against anything.
- Estimated means a market range assembled from several sources, or arithmetic done on top of the other two.
The honest disclosure up front: almost nothing in this post is Measured. We have not commissioned a twenty image catalogue shoot, so we have no photographer's invoice of our own to publish. Everything about photography pricing below is market observation, and it is labelled that way rather than dressed up.
What the market charges
Traditional product photography for a small brand working with a freelance photographer runs roughly $75 to $150 per finished image, all in (Estimated, market range). All in matters more than the number. That figure absorbs the day rate, the space, props and styling, basic retouching, and the coordination time that nobody itemises but everybody pays for.
Run it out at the middle of the range and a 20 image catalogue shoot is around $2,000 (Estimated, arithmetic). That is the number most small brands are actually deciding about. Not a per-image price, a four figure commitment made once or twice a year.
Two things make the per-image framing misleading in opposite directions. Ten images does not cost half of twenty, because setup and travel are fixed and get amortised, so small shoots are worse per image than large ones. And the price is per finished image, not per exposure. The photographer shot far more frames than you received. You are buying the selects.
The cost that is not money
A catalogue shoot is two to three weeks of calendar for a small brand (Estimated). Booking against someone else's availability, shipping samples and waiting for them to arrive, the shoot day itself, then selects, then a retouching round, then usually a second retouching round.
For an established product line that is fine. Where it hurts is a launch, a seasonal push, or a marketplace listing that is underperforming and needs new images this week. The shoot cannot compress. That is the constraint people feel first, before they feel the invoice.
There is a second-order effect that matters more than either. When new images cost a booking and three weeks, you stop trying things. You shoot the shot you are sure about. Nobody books a studio day to test whether the product looks better in a kitchen than on a marble slab, so that question just goes unanswered for years.
Every path pays for attempts
Published prices, on both sides of this comparison, quietly assume every output is a keeper. None of them are.
On the generation side, getting one image we would actually publish takes somewhere between two and four generations, depending on how unusual the product is and how specific the brief was. Dark backgrounds, fine print on the label, and unusual silhouettes sit at the top of that range. That multiplier is an estimate from our own usage, not a measured constant, and we have never sat down and counted it properly (Estimated). Treat it as a planning figure and discount it accordingly.
The photography equivalent is not the frame count, it is the reshoot. A shoot that comes back wrong, wrong angle for the marketplace crop, colour that does not match the product, a prop that dates the image, costs another booking and another two weeks. The attempt rate is similar in spirit. The price of a single attempt is what differs, and it differs by orders of magnitude in both money and calendar.
That is the actual argument, and it is not "AI is cheaper than a photographer". It is that the cost of trying an idea collapses, which changes what you are willing to attempt. Twenty scene concepts is not a budget conversation any more. It is an afternoon.
What the rest of the market is doing, and why that is weak evidence for you
Two data points get quoted constantly in this category, and both deserve a caveat attached.
Around 68 percent of e-commerce leaders say they plan to increase AI photography spend in 2026 (List, from a vendor-published survey). The vendor that published it sells into the category. That does not make it false, it makes it a number with an interest in itself, and we would not build a business case on it.
Inditex publicly committed around $400M to AI photography infrastructure by the end of 2026 (List, publicly reported). That one is real money and it does tell you the direction of travel. It tells you almost nothing about whether you should generate your next twenty images, because a group operating at that scale is solving catalogue throughput, not deciding between one photographer and one tool.
Your decision is smaller and more specific than either figure. Treat both as weather, not as advice.
Three honest qualifications
The comparison above is real, and it is routinely overstated. Three things do not get cheaper:
- Your time is the actual cost. Briefing, selecting, correcting and editing is real work, and at any sensible hourly rate it is the largest line in a generated workflow by a wide margin. The savings against a shoot are large, and they are nothing like the ratio the raw prices imply.
- You still need a source photo. Everything here assumes you already have a clean shot of the real product. Someone took that picture. It can be a phone on a windowsill with decent light, and often that is enough, but if the answer is a photographer then a photographer is still in your budget.
- Some shots still want a camera. Fine texture, unusual materials, hands interacting with the product, and anything where a legal or quality claim rests on the image being literal. Generation is good at putting a real product into a scene that never existed. It is not a substitute for evidence.
So when do you still book the shoot
| Your situation | What we would do | Why |
|---|---|---|
| Hero product launch where the material is the selling point | Book it | Texture and finish are where redrawing is weakest and a camera is strongest |
| Forty listings that look inconsistent, needed this month | Generate, and put the text back as layers | The constraint is calendar, and consistency across a catalogue is the thing generation is actually good at |
| No clean packshot of the product yet | Camera first, even a phone | Every generated route starts from a real photo of the real product |
| An image supporting a regulated or comparative claim | Book it | The image needs to be evidence, not a plausible reconstruction |
| Testing twenty scene ideas before committing to anything | Generate | This is the question a shoot was never economic enough to answer |
Most small brands end up doing both, and the split is usually less even than they expect. One short shoot a year for the source material and the shots that need a camera, then generation for the scenes, the seasonal variants, and everything the marketplace crop demands.
Method and caveats
- The photography range, the shoot total, and the calendar estimate are market observation drawn from our own competitor and category research, not from invoices we have paid. They skew toward US and UK freelance rates. Local rates vary widely and a photographer in your city may reasonably quote outside that range in either direction.
- The two to four generations per keeper figure is our own planning estimate. We have not instrumented it, and a properly counted re-roll rate per product type is on our list because we would rather publish a number than a range.
- The survey figure and the Inditex commitment are quoted from public sources and labelled as such. We have not audited either.
- Nothing here counts your own hours, which we think is the largest cost on both sides.
If you have real photography invoices, from either side of this, we would genuinely like to see them. That is the data this comparison is missing. Write to support@kloti.co.
Related: why AI product photos look fake, and how to spot each tell.
Written by the Kloti team. Kloti is an AI visual studio for product and service ads. Every image on this page, cover included, was generated in Kloti.