How to shoot a clothing catalogue without a studio

It is worth itemising, because “we need a shoot” hides at least five different things and you probably do not need all of them.

• A controlled light source. Soft, consistent, repeatable. This is the part people think they are buying and it is the easiest to replace.
• A person to put the clothes on. Booked by the half day, which is why shoots get batched into marathon sessions.
• Someone with taste. Styling, framing, knowing that the collar needs fixing before the shutter. Genuinely hard to replace.
• Retouching. Usually billed separately and usually where the timeline actually goes.
• A deadline. Underrated. A booked day is the thing that makes the photographs happen at all.

A small shop adding twenty products a month cannot run that cycle monthly, so in practice the catalogue goes up with whatever the supplier sent, and the new items look nothing like the ones from the last shoot.

You need a clean photograph of the actual product. Flat on a table, hung on a plain wall, laid on a sheet — it does not need to be beautiful, it needs to be accurate: the real colour, the real print, the whole garment in frame, no deep shadows swallowing a sleeve.

This is the input everything else is built from, whether that is a retoucher or a generation tool, and it is the one step no software removes. Shoot it on a phone by a window. Overcast daylight through a north-facing window is, genuinely, close to the light a softbox is imitating.

What makes a flat-lay usable

• The whole garment in frame, not cropped at the hem.
• Shot square-on, not at an angle, so the shape is not distorted.
• Even light across the piece — no hard shadow across half of it.
• Colour that matches the real thing. Photograph it next to something white and you have a reference to correct against.

Here is the thing most people get backwards. A customer almost never looks at one of your photographs in isolation. They look at a category page: thirty products in a grid, at thumbnail size, in one glance.

At that size, individual image quality barely registers. What registers is whether the grid looks like one catalogue. Same person, same light, same distance, same crop. Thirty products on one model reads as a brand; thirty products on eight different models, three backgrounds and two lighting setups reads as a marketplace, even when every single photograph is good.

So when you are judging any approach — a photographer, a tool, a friend with a camera — generate or shoot six products and look at them as a grid. Not one at a time. The difference is invisible in a single frame and obvious in a row.

Clothing needs a body. Flat-lays sell basics acceptably and sell anything structured badly, because the customer cannot tell how it hangs.

Three honest options, with their real trade-offs:

• Hire one. Best result, highest cost, hardest to repeat. The repeat problem is the real one: the same model in six months, available, at the same rate, is not a given, and switching models mid-catalogue is visible.
• Mannequin or invisible-mannequin. Cheap and consistent forever. Reads as a catalogue from 2012, and badly on anything draped.
• A generated model. Consistent by construction and available at 2am. Cannot be a real person your customers recognise, and in some markets you will need to say it is generated.

That third option is what Softbox does, and we have written honestly about when it is the wrong choice. If you need a face customers already know, none of this helps you.

Campaign imagery. The six pictures at the top of your homepage that have a concept behind them — a location, a mood, a story about who buys this. That is art direction, not photography, and nothing automated does it, because the hard part is deciding what the picture should be, not making it.

Also: anything where the texture is the product. Heavy knitwear, leather grain, embroidery, anything sheer. Get those shot properly once and reuse them forever.

If you are starting from nothing this week:

• Photograph everything flat, by a window, in one sitting. One session, one light, so the inputs are consistent before anything else happens.
• Decide your model once — hired, mannequin or generated — and write down what they look like. This is the decision you are least likely to revisit and most likely to regret rushing.
• Pick one background and keep it. Changing it later is cheap; mixing it is not.
• Do six products end to end before doing sixty. Put them in a grid and look at them together.
• Only then scale up.

The expensive mistake is not choosing the wrong tool. It is doing two hundred products with a setup you have not looked at as a grid.