A 3D product configurator shows a customer a picture of exactly the version they have chosen, in the size, colour and finish they picked. For businesses selling products with a lot of options it is often treated as a nice extra. Past a certain number of options it becomes the only practical way to show customers what they are buying, because product photography, and even 3D product rendering done one image at a time, cannot keep up.
Here is the sum that shows why. Most businesses selling made-to-order products have done it at least once, usually just before booking a photographer, and usually on the back of an envelope.
Say you make 300 products. Each comes in four styles and 25 colours, and customers can choose between two configurations. That is 60,000 different things a customer could order, and in principle 60,000 pictures.
Now add a new trim next season, in ten colours. You have not added ten pictures. You have multiplied by ten. It is now 600,000.
Options multiply, they do not add
This is the part that catches people out. Adding a product adds a handful of pictures. Adding an option multiplies every picture you already needed, because every existing combination now comes in each version of the new choice.
Even at five minutes a picture, which is quick for a studio or for a properly lit 3D render, 600,000 pictures is 50,000 hours of work. That is roughly twenty-nine years for one person working full time, before anybody reshoots anything because a colour has been discontinued.
Nobody does that, so businesses quietly do something else instead.
What usually happens instead
Look at almost any website selling configurable products and you will see the same few workarounds:
- One photograph of the best-selling version, a row of colour swatches underneath it, and the words “for illustration only”
- Photographs of a few popular combinations, and nothing at all for the rest
- A base photograph with the chosen colour painted over it by the website
- A sales team emailing pictures from the brochure, or posting out samples
Each of these is a reasonable compromise, and each one costs you something. The customer has to imagine the thing they are about to pay for. The ones who cannot imagine it do not buy, or buy it and send it back, or ring you to ask, which is a sale your website was supposed to make without you.
The painted-over photograph is the most tempting, and it has a particular weakness. Colour is not a flat sticker on the front of a product. It is on inside surfaces, seen at an angle, in the shadow of another part. A picture that recolours the front and leaves the inside the old colour is not nearly right. It is wrong in a way customers notice, and as we have said before, a picture that is nearly your product is worse than no picture at all.
The signs you have crossed the line
There is no exact point where photography stops working, but the symptoms are easy to recognise:
- The photography schedule is permanently behind the product list
- “Colour shown for illustration only” appears on more pages than not
- Someone has argued against a new option because of what it would do to the photography
- Customers regularly ask to see a particular combination before they order
- Returns or complaints include the phrase “it was not what I expected”
If two or three of those sound familiar, the answer is probably not a bigger photography budget.
Product photography, 3D product rendering or a configurator: three options compared
Once you can see the numbers, there are really three options on the table.
- Product photography. Nothing beats it for authenticity, texture and showing a product in real use. It needs a physical sample of every version you want to show, and every combination is another shot. Right for small ranges and for lifestyle images whatever the size of your range.
- 3D rendering by a CGI studio. No physical samples needed, and changing a colour is easier than reshooting it. But each image is still produced by an artist, one scene at a time, so the cost still grows with the number of pictures. It moves the problem rather than removing it.
- Automatic 3D rendering from your product data. Each product is described once, and every combination is drawn by a computer when it is needed. The cost grows with the number of kinds of product you make, not the number of pictures. This is the idea behind a 3D product configurator.
The rest of this post is about the third option, because it is the one that changes the arithmetic.
How a 3D product configurator works
If you make a product to order, you already know exactly what it looks like, in millimetres. Somebody in the business has to, or it could not be made. The measurements, the materials and the way the parts fit together are all written down somewhere.
That is enough to draw it. Rather than photographing every version, you describe each product once as a set of measurements and parts, and a computer draws whichever version the customer asks for. A new colour or finish stops being a photo shoot and becomes a setting. Adding ten trim colours means adding ten small material samples, not 540,000 pictures.
This is not AI image generation, and the difference matters. Nothing is invented or guessed. Every edge in the picture comes from a real measurement, which is exactly why it can be trusted where a generated image cannot.
The speed is what surprises people. In a proof of concept we built recently, each high quality image took about a second and a half to produce on an ordinary server with no graphics card, and came out as a small, web-ready file. At that rate the whole 600,000 in the example above is around ten days of computer time on a single machine, and less if you spread it across several. Render time and storage were never the real constraint.
In practice you would not produce all of them anyway. You draw the popular versions in advance, so listing pages, search engines and social media previews have a picture straight away, and let the rest be drawn the first time somebody asks for one. The same drawing can also run live on the product page, as a 3D product configurator, so the picture changes as the customer picks their options.
Where the real work is
If the computing is the easy part, the effort goes into two things, and both are about information rather than technology.
The measurements. Stock and ordering systems usually hold codes, prices and descriptions, not shapes. The dimensions that matter tend to live with the engineering or design team, often in a spreadsheet they keep for their own purposes. The best outcome is reading that file exactly as it is. Asking a busy engineering team to maintain a second copy, in a format the website would prefer, is how the data goes out of date.
The materials. To draw a surface convincingly you need a flat, evenly lit sample of it that repeats without a visible join, one for each colour or finish. That is not the same thing as a photograph of a finished product, and it is not something most marketing teams already hold. Expect to commission these or have them made from what you have, and expect it to take longer than the software.
What it will not do
- It is not lifestyle photography. It produces clean, accurate product pictures. You will still want real photographs of the product in use, in a real setting, to sell the feeling.
- It suits some products far better than others. Anything built from flat panels and straight parts is a good fit. Curves, soft fabrics and glass are all possible but much harder to make convincing, so it is worth asking early whether yours need them.
- It is only as right as the data. A wrong measurement produces a wrong picture, confidently. Somebody has to own the numbers.
- A new kind of product still needs a developer. New colours and finishes are data your own team can add. A product built differently from anything before it needs describing once.
- It is not instant to set up. Most of the time goes into gathering the measurements and the material samples, not into the drawing.
Do you need a 3D product configurator?
If your range is small and rarely changes, choose photography. Book the photographer.
If you need images of a handful of new products before samples exist, a CGI studio is a sensible middle route.
If you are adding options faster than you can photograph them, or you have stopped adding options because of the photography, the arithmetic has already made the decision for you. The pictures you need are sitting in your product data. They just have not been drawn yet.
If you would like to talk through whether your products would suit this, get in touch.