What product photography actually costs in 2026, and the alternatives

For a typical ecommerce brand, expect $35 to $80 per finished image through a studio marketplace, $500 to $1,500 for a half day booking a freelance product photographer, and $1,500 to $3,500 for a full day with a specialist who owns proper lighting and does their own retouching. Those are US figures, August 2026, and they have been fairly stable for two years.

The spread inside those ranges comes from three things: how much styling the shot needs, who supplies the props and surfaces, and how many retouched finals are included. A photographer quoting $900 for a half day who includes 15 retouched images is a better deal than one quoting $600 with 5, and almost nobody prices that out before booking.

Studio marketplaces like Soona made pricing legible by charging per final image, around $40 each, with props and models as add ons. You ship the product to them, pick shots from a live feed, and get files back in days. For straightforward catalog work this is the honest benchmark to compare everything else against.

Mostly lighting and judgement. Anyone can put a bottle on a table. The photographer earns their rate by controlling reflections, keeping color accurate to the physical product, and knowing which angle makes a jar of moisturizer look expensive instead of medical. Retouching is the other half: dust removal, color correction, and background cleanup routinely take as long as the shoot itself.

This is also why the cheap end of the market disappoints people. A $150 shoot on a marketplace gig site usually means natural window light, minimal retouching, and color that drifts from the real product. If the images are going on a product page where returns hurt you, color accuracy is not a nice to have.

Some work should stay with a human holding a camera, and it is worth being specific rather than polite about where that line sits. Hire a photographer for hero campaign images that will run for a year, for jewelry and anything else that is mostly reflective surfaces, for food that depends on steam, pour, or motion, and for fabric where a buyer will zoom into the weave before paying $200 for a coat. Those shots sell on physical truth, and faking physical truth is exactly where generated imagery is weakest.

The same goes for regulated categories. If you sell supplements or skincare with claims on the label, the label in your ads needs to be the label, pixel for pixel. A photograph is the simplest way to guarantee that.

If your whole business is ten hero products that change once a year, book the full day shoot. The economics below only get interesting when you need many images, often.

The DIY route costs about $300 once: a foldable lightbox around $60, a tripod around $40, and any recent phone. For plain white background shots this genuinely works, and marketplaces like Amazon were built around exactly that shot. Where DIY dies is styled scenes. Getting a credible kitchen counter, believable shadows, and props that do not look like a garage sale takes the skill you were trying not to pay for, and most people quietly stop after the first afternoon.

AI product photography sits between the two. Tools run $20 to $100 a month and work from photos you already have: you give them a clean shot of your product, and they rebuild the scene around it. Styled backgrounds, seasonal variations, lifestyle staging, done in minutes instead of a shoot day. For ad creative specifically, where you want twenty variations to test rather than one perfect image, the cost per usable image lands under a dollar, which no camera based option can touch.

Good enough that the interesting question has moved from whether the image looks real to whether your product stays exactly your product. The failure mode that matters is not fake looking backgrounds anymore, it is a generator quietly redrawing your label, shifting your bottle's proportions, or inventing a slightly wrong shade of your packaging. If you evaluate a tool, evaluate that: give it your product photo and zoom in on the label in the output.

This is the specific problem we built Dovey around. It studies ads already proven in your space, then generates creatives where the product is placed exactly as photographed, with the scene, lighting, and text built around it rather than over it. The $1 trial exists so you can run that label test yourself on your own product before paying real money.

Where AI still loses cleanly: macro texture, reflective surfaces, and any shot whose entire job is physical accuracy. For those, book the photographer. For the other 80 percent of images a brand needs in a year, generated is now the default and the shoot is the exception.

Price the year, not the shoot. A brand refreshing 40 catalog images twice a year at marketplace rates spends about $3,000 to $6,500 annually before a single ad variation. The same brand covers catalog refreshes and ad testing with a $300 DIY setup for white background shots plus an AI tool at $600 to $1,200 a year, and books one real shoot for the handful of hero images that deserve it. That split is what most small brands we see end up at, and it is usually the right one.

If you want to test the AI half of that split, Dovey is what we built for it. Give it your product photo and it studies ads already working in your category, then generates creatives with your product placed exactly as shot, checked against your original so the label never comes out approximated. Text stays editable, so changing a headline costs nothing. It is $1 for three days, which is enough to run your own product through it and decide from the actual output rather than from this article.