What a product clinic can measure, and what it can't

Ask a designer how they know a shape will sell and you may get an answer that sounds like evidence: we put it in front of buyers and they preferred it. The useful part of that sentence is hidden in two small words. Which buyers, and preferred it to what?

A product clinic is the tool behind those claims. It’s a real instrument with real limits, and knowing how it works makes launch presentations easier to read.

A clinic is a room, a model and a recruited group

A clinic shows a vehicle, or a stand-in for one, to people recruited to match a target buyer. What they see might be a full-size model, a mock-up with no interior, a clay under studio lights, or images on a screen. The setting is controlled and participants are usually asked to keep what they saw confidential, which is why the details rarely surface in public.

The design of the session matters more than the venue. A clinic can show one design to a group, which is called monadic testing, or several designs to the same group, which is comparative. Comparative work ranks alternatives against each other well. Monadic work is closer to how a buyer actually meets a car: alone, with nothing beside it for scale.

Neither is wrong; they answer different questions, and a presentation that cites “customer research” without saying which kind was run has told you very little.

Controls are how a team finds out whether the test works

Good clinics include options that aren’t meant to win. Sometimes those are deliberately wrong: a design the team expects to fail, included to check that the session can tell a strong answer from a weak one. If the control scores as well as the intended design, the test isn’t measuring what the team thinks it is.

That’s a modest kind of rigour. It checks the instrument. It doesn’t check the future.

Ranking is something people do well; prediction is not

Clinics are genuinely good at the job they’re built for. Given these options, which do you prefer, and why? That’s a ranking task, and people are reasonably consistent at it, especially when the options are shown at full size.

They’re much weaker at predicting use. A session captures a reaction in a room, in a single sitting, with a vehicle that doesn’t move and a participant who knows they’re being watched. It doesn’t capture what happens on a wet Tuesday with two children, a tight parking bay and a phone that needs charging.

That gap isn’t a flaw in the method so much as the edge of it. Reaction and use are different data. A preference stated in a clinic is a statement about first contact: how something looks, how the seat feels when you first sit in it, whether the controls make sense before you’ve learned them.

The familiar has an advantage it didn’t earn

Participants bring brand and price expectations into the room with them. Show the same shape with a different badge and the answers move. Show it at a different implied price and the answers move again. Nobody in the session is being dishonest; they’re doing what buyers do, reading a car against everything they already know.

Novelty, meanwhile, tends to be penalised. A design that doesn’t resemble what participants already own asks them to imagine living with something unfamiliar, and imagination is expensive in a room with a fixed agenda.

Clinics happen before design freeze, which is exactly why they’re worth running: the answer can still change the car. But it also means the design that survives is the one that has already been softened towards the average of the room. A shape that tests badly against a familiar alternative may be a shape ahead of its time, or simply a bad shape. The clinic can’t tell you which.

A clinic can tell you which of two cars people prefer. It can’t tell you what either one is like to live with.

How to read a customer research claim at a launch

When a launch presentation says the design was shaped by customer research, the follow-up questions are specific and slightly dull.

• What was shown: a full model, a clay, images, or a car people could sit in?
• What was it compared with, and was one of the alternatives designed to lose?
• Was the question about preference, or about use?
• Who was in the room, and what did they already drive?

A clinic is a legitimate tool with a narrow job: it tells a team which of the options in front of it is more appealing to the people it recruited, under conditions it controls. That’s genuinely useful, and it’s cheaper than building the wrong car.

What it can’t do is stand in for the market. The market is what happens after the car is on sale, in weather, in traffic, at whatever price the badge will bear, to people who never sat in a clinic and never will. Human factors has its own standards literature, which is a hint that measuring people is a technical problem rather than a matter of taste, and a technical problem has a measurement range. A clinic sits inside that range.

The honest version of a research claim is narrow: this shape beat that shape with these people, on this day, under these conditions. Everything wider than that is a story told afterwards, usually by someone who wasn’t in the room.