Studio-quality product photos — without a studio

What a product photo actually has to do

A product photograph has one job that matters more than any other: it has to make a stranger confident enough to spend money on something they cannot touch. Everything else — the styling, the background, the crop — exists to serve that. This is why the technically perfect shot sometimes converts worse than a slightly plainer one. The buyer is not judging your photography. They are trying to work out what will arrive in the box.

That reframes what "good" means. A photograph that flatters the product but misrepresents its colour will earn a return and a bad review. One that shows the product honestly, at a size the buyer can judge, on a background that does not compete with it, will outperform a more beautiful image that leaves a question unanswered. The most common failure in e-commerce photography is not ugliness. It is ambiguity.

• Scale: can the buyer tell how big it is without reading the dimensions?
• Colour: does the photographed colour match the physical product under normal light?
• Finish: is it matte or gloss, soft or rigid, and can you tell from the image?
• Completeness: does the photograph show what is actually included?

Why white backgrounds still matter — and where they stop

The plain white packshot is not a stylistic choice; it is a compliance requirement. Amazon states the main image must "have a pure white background (RGB color values: 255, 255, 255)", and Google Shopping applies comparable rules. Getting this right is the price of entry, and it is unglamorous work: even lighting, clean edges, no shadow pooling at the base, product filling most of the frame without touching the edges.

But a catalogue made entirely of white packshots tells the buyer nothing about the product in use. This is where the second and third images earn their place. A held shot answers scale instantly — a bottle in a hand needs no dimensions table. A styled surface shot places the product in the life the buyer imagines. A macro crop proves the finish is what the description claims. Sellers who treat image two onward as an afterthought are leaving the most persuasive frames empty.

The practical sequence most sellers converge on is the same: white packshot first for compliance and for the grid, then scale, then texture, then context. You can argue about the order of the last three. You cannot skip the first.

Sources: Amazon Seller Central — Product image guide — main image must be on pure white, RGB 255, 255, 255 · Google Merchant Center — Image requirements

What changes by category

Product photography advice is usually given as though all products behave the same way. They do not, and the differences are mostly about how surfaces handle light.

Clothing is a geometry problem. Fabric drapes, folds and stretches, and the failure mode is a pattern that changes scale across a fold or a seam that dissolves. What you are protecting is the weave and the true colour, because those are what a returns request is written about.

Electronics is a reflection problem. Brushed metal, glass and gloss plastic each throw light differently, and the tell of a bad render is highlights that do not agree with each other — as though the object were lit by three suns. Screens are their own trap: clip them to white and the product looks switched off.

Beauty is a typography problem more than anything else. Buyers read labels. An image that softens the text on a serum bottle into approximate letterforms has failed, however attractive the lighting. Glass translucency matters for the same reason — it is how a buyer judges how much product is in the bottle.

Food and beverage is a context problem. The packaging is usually simple; what sells is the surface it sits on and the temperature the image implies. Condensation on a cold can does more work than another degree of sharpness.

Where AI genuinely helps, and where it does not

The honest case for generating product photography is consistency and cost at volume, not artistry. A preset applied across twenty-five products produces a grid that looks deliberate, because every frame shares the same lighting, crop and background. Human shoots drift between sessions — different day, different assistant, slightly different light — and that drift is visible in a catalogue grid even when each individual image is excellent.

The cost argument is straightforward. A studio day with a stylist and retouching is a real budget line, and it has to be repeated every time packaging changes or a variant is added. Regenerating from a new photo does not carry that cost, which changes what is worth photographing at all. Long-tail SKUs that never justified a shoot can now have proper images.

What AI does not replace is art direction. A hero campaign image — the one with an idea in it — still comes from a person who understands the brand. A generator will reliably produce a good photograph of your product; it will not decide that your product should be photographed underwater because that is the story this season. Treating it as a replacement for that judgement produces catalogues that are technically clean and completely forgettable.

There is also a hard limit worth stating plainly: the tool works from a photograph of a real product, not from a description. That constraint is deliberate. It is what keeps the output honest about colour, proportion and label text — and it means you still need the sample in your hand.

Motion is becoming the default, not the upgrade

Static product images are no longer the whole job. Amazon now recommends at least six additional images and one video per listing, and paid social placements are built for motion first — a still frame in a video slot is competing at a disadvantage before anyone reads the caption. The marketplaces have moved faster than most catalogues have adapted.

The barrier has been production. A short product clip traditionally meant a second shoot, a turntable or a gimbal, and an edit — a different budget and a different vendor from the stills. Generating motion from the same product image collapses that into one step, and more importantly keeps the product identical between your still and your video. A listing where the photograph and the clip disagree about colour is worse than a listing with no clip.

The useful moves are unglamorous: a slow orbit that shows the product in the round, a push-in that lands on the detail the description is about, a lift that shows weight and scale. None of them are cinematically ambitious. All of them answer a question the still could not.

Source: Amazon Seller Central — Product image guide — recommends six additional images plus one video per product

A workflow that holds up at catalogue scale

The difference between a catalogue that looks professional and one that looks assembled is almost entirely process. Photograph every product the same way on input — same distance, same orientation, same rough lighting — and the generated output will be consistent without further effort. Vary the input and no preset will save you.

Decide the preset before you start, not per product. The temptation is to pick the nicest background for each item, which produces a grid that reads as a marketplace of different sellers. Pick one surface for the main image across the whole catalogue and vary only in the secondary frames.

Then run in batches and review as a grid, not one at a time. Problems that are invisible in a single image — a drift in background tone, an inconsistent crop, one product photographed noticeably larger than its neighbours — are obvious the moment you see twenty-five thumbnails together. That review step takes minutes and is the one most often skipped.

• Standardise the input photo before standardising the output
• Choose one main-image preset for the entire catalogue
• Generate in batches of up to 25 and review as a grid
• Keep the white packshot for image one; spend images two to five on scale, texture and context
• Regenerate rather than retouch when packaging changes