People asking for a product photography quote usually start with how many products they have. «It's sixty references.» The next question — how many images are needed — tends to go unanswered, or gets something vague like «about five per product». You multiply, you arrive at three hundred images, and the number frightens everyone before anyone has worked out what's inside it.
The problem is that this sum mixes three things with very different costs. A new image of a product that has never been photographed costs one thing. An image of the same piece in the next colour costs a fraction of that. A context image, with a set built and possibly a model, costs several times more and serves ten products at once. Adding them all together as if they were equivalent units is the fastest way to budget badly.
Better to do three separate calculations.
Calculation one: the base set per reference
Each reference needs a set of images that answers the questions someone asks before buying without being able to touch the product. It isn't a question of quantity: it's coverage. A well-served reference typically has:
- The hero image. Product isolated, clean framing, no ambiguity about what you're selling. It's the one that appears in the category grid and in the ads, and it's the only one a lot of people ever see.
- Two or three views. Angles that show what the hero hides: the back, the profile, the opening, the base. How many depends on the geometry of the object, not on a rule. A t-shirt needs front and back. A bottle needs the label and the pump. A bag needs the inside.
- One material close-up. Weave, grain, finish, stitching, stamp, engraving. This is the image that stands in for running your hand over the product.
- One scale image. A hand, a recognisable object, a shot of it in use. It resolves the doubt that generates the most returns.
- One functional detail, where there is one. The clasp, the fitting, the lid, what comes in the box.
That gives you somewhere between five and eight images per reference. Below five there's almost always a question left unanswered. Above eight, for most products, repetition starts creeping in — and twelve images showing the same angle half a degree apart are worth less than five that resolve five distinct doubts.
There are exceptions in both directions. Simple consumables with informative packaging and no material decision to make — a bag of coffee, a shower gel — live perfectly well on three or four. Expensive, complex or rarely bought products — furniture, equipment, technical parts — justify ten or twelve, because the cost of an unresolved doubt is proportional to the price.
Calculation two: the variants
This is where budgets go off the rails, because «variant» means very different things depending on the product, and not all of them require new photography.
The question to ask is this: does this variant change what you see?
Variants that change everything. Different colours, different prints, different materials. They need their own image, and they need a trustworthy one — showing the blue in a photo and leaving the green to the written description is a guaranteed return. But they don't need the entire base set. Usually the hero plus one or two views will do; the material close-up and the functional detail were already shot in one colour and serve all of them, as long as the finish is the same.
Variants that change nothing. Clothing sizes, capacities of the same container, units per pack. Photographing the S and the XL of the same shirt is spending money to produce two indistinguishable images. What resolves size is a decent measurements table and, in clothing, one or two images on different bodies — which is a different thing, and belongs to the third calculation.
Variants in between. The same model with gold or silver hardware, the same cabinet in oak or walnut. Here the decision depends on how much weight the variant carries in the sale. If people choose on that characteristic, it has to be visible. If it's a secondary detail, a close-up of the hardware in each finish solves it without repeating the whole series.
The practical sum is simple: full base set on the main variant, plus a reduced sub-set — two to three images — for each variant that changes the appearance. A product with six colours doesn't cost six times a single-colour product. It costs something like twice as much, because the set is already built, the lighting is already done, and what changes is swapping the piece and shooting again.
This is also why grouping variants into the same session is the decision that saves the most money in product photography. Shooting the six colours on different days turns six product swaps into six set builds — and, worse, almost guarantees the colours won't match each other in the grid.
Calculation three: the use context
Context images aren't counted per reference. They're counted per situation.
A photograph of a laid table with three pieces from the collection serves those three product pages, plus the category page, plus the homepage, plus social, plus the newsletter. A person wearing a backpack in the street serves that backpack and, if the framing helps, the jacket and the shoes they have on. Context has a different economy: it costs more to produce and pays back in more places.
What defines the cost here isn't the number of images, it's the number of sets built. Each set means location, props, surfaces, possibly a model, and time spent getting the light right. Ten images in the same set cost barely more than six. Six images in six sets cost nearly six shoots.
For most shops, two to four sets per collection or per season give enough coverage, as long as they're chosen to do work rather than to decorate:
- A real-use set, showing the product doing what it exists to do.
- A scale and proportion set, with a human body or reference objects.
- A grouping set, showing products together and suggesting combinations — useful for average order value and for people who don't know where to start.
- An atmosphere set, looser, saying who the brand is speaking to. It's the only one that can be freer, and the only one that shouldn't be first into the budget.
Clothing is the case where context stops being optional. Without a body you can't read the cut, and without reading the cut people buy blind. In that case the maths changes shape: context becomes part of the base set for every reference, and the cost per product rises structurally, because model, production, styling and fitting time all enter the equation.
How this translates into a budget
A product photography budget is time. It isn't the number of frames — it's the number of set builds, swaps and treatments.
The three blocks behave differently:
- The base set scales almost linearly. Each new reference is set time, cleaning the piece, shooting and selecting. It can be optimised by grouping products of similar geometry, but it doesn't disappear.
- The variants scale far below linearly, provided they're shot back to back. This is the line item that benefits most from planning and suffers most from fragmented sessions.
- The context is almost entirely a fixed cost per set. Deciding to do four instead of two is a real budget decision; deciding to shoot fourteen images instead of ten inside the same set is barely felt.
There's one more portion that rarely appears on the sheet and that grows with the number of images: post-production. Cut-outs, clean-up, colour matching between variants, exporting in the formats the shop, the ads and the marketplace demand. That last part multiplies — the same image in square, vertical and horizontal is the same image and is three times the work. It's worth fixing the formats before the shoot, because framing with three ratios in mind costs nothing at the time and costs hours afterwards.
Where to cut when the number doesn't fit
When the total exceeds what's available, the usual reaction is to reduce the number of images per product across the board. It's almost always the wrong choice, because it cuts the same percentage from the product that carries the shop and from the one that sells three units a year.
It makes more sense to cut in layers. Define the products that sustain the shop and give them the full set, variants included and context of their own. Give the long tail the minimum that answers the essential questions: hero, one or two views, scale. You can go back later, when one of those products starts moving.
The second layer to reduce is the number of context sets, not the care with which each one is made. Two well-produced sets are worth more than four improvised in an afternoon.
What isn't worth cutting is what sits underneath everything: the quality of the light and the reliability of the colour. Those two things are what make thirty images look like a shop rather than a folder of files.
The right question
Instead of asking how many photographs you need, ask how many questions each page has to answer, how many variants genuinely change what you see, and how many sets you can build properly. Those three answers give you a number you can defend — and, more importantly, they give you a shot list where every image knows what it's doing there.