Please stop asking your customers what they want

There is a sentence that gets said in meetings, usually by the most reasonable-sounding person in the room, and it quietly ruins things. “Let’s just ask our customers what they want.”
It sounds humble. Customer-focused. Grown-up, even. Who could argue? Arguing against it feels like arguing against listening, and nobody wants to be the person against listening.
So let me be that person for a minute, because this well-meaning little sentence has flattened more good ideas than any villain ever did. Not because listening to customers is bad. Because asking them what they want is the wrong question, aimed at the wrong faculty, and it gives you confident answers that are almost always wrong.

 

People don’t know what they want. They know how they feel.

Here’s the uncomfortable bit of psychology underneath all this. Humans are brilliant at telling you what’s wrong with something in front of them, and genuinely hopeless at imagining what doesn’t exist yet.
Ask someone “what do you want from a bakery” and they’ll say something sensible and useless like “good bread, fair prices, friendly service.” Thanks. That describes every bakery that has ever existed and points you toward nothing. Nobody in the history of the world walked into a market researcher’s office and described, unprompted, the thing they’d actually fall in love with next year. They can’t. It isn’t in the room yet, and imagination-on-demand is not a thing people are good at.
What people can do, superbly, is react. Put something real in front of them, and watch. Watch where they hesitate, what they reach for, what makes them lean in, what they quietly ignore while telling you they love it. That’s the gold. Not what they say they want in the abstract, but how they behave when something specific is actually there. The mouth lies to be polite. The behaviour doesn’t bother.

 

This isn’t just a hunch. People have actually studied it.

Marketers even have a polite name for the whole mess: the “say-do gap,” the stubborn distance between what people tell you and what they go on to do. And whenever someone bothers to check the two against each other, the talking loses to the doing.
There’s a lovely example of this. A team of behavioural scientists at Warwick University once tested it properly: they took what shoppers said they’d buy, took what shoppers actually bought (validated against a full year of real sales across hundreds of supermarkets), and compared the two. Watching what people did turned out to predict real behaviour far better than asking them ever could. Not a little better. Meaningfully, embarrassingly better.
And then there’s the finding that still makes me grin, because it’s so sneaky. A marketing professor called Vicki Morwitz, over at NYU, discovered something back in the nineties that should unsettle anyone who loves a quick survey: the mere act of asking people whether they intend to buy something makes more of them actually buy it. She and her colleagues asked people about their intentions to buy a car, or a computer, and that alone nudged the numbers up. It even has a name now, the “mere-measurement effect.” Which means your innocent little questionnaire was never innocent. It doesn’t just measure the room, it changes it, and then hands you the altered number with a straight face.
Pile on two things psychologists keep bumping into, our habit of answering to look good rather than to be accurate, and our talent for inventing tidy reasons after the fact for decisions our gut already made, and the survey stops being a window onto your customer. It becomes a funhouse mirror, and you’re trying to steer by it.

 

“What would you like?” gets you a committee. And committees make beige.

There’s a second problem, and it’s worse. When you ask everyone what they want and then dutifully try to give it to them, you don’t get a great product. You get an average one. You get the design equivalent of a meal cooked to please a table of forty people with different diets. Technically edible. Loved by no one.
Every strong brand you can think of is strong precisely because someone, at some point, did not put it to a vote. Somebody decided. They had a point of view, they backed it, and they accepted that some people wouldn’t like it. The stuff that gets designed by consensus, by carefully averaging everybody’s stated preferences, comes out smooth and shapeless and instantly forgettable, because the sharp edges people actually remember are the exact things a survey sands off.
Asking “what do you want” and obeying the answer isn’t leadership. It’s outsourcing your judgment to people who were never supposed to have your job, and who, quite reasonably, put less thought into the answer than you’ll put into acting on it.

 

So what do you do instead? You watch, and you decide.

None of this means ignore your customers. It means pay attention to the right signal. Stop interviewing their opinions and start observing their behaviour. What do they actually click, buy, abandon, come back for, complain about at 11pm? What problem do they keep working around because no one solved it properly? That’s data with a pulse. It tells you the truth their questionnaire answers were too polite to.
And then, the part that takes nerve: you use your own judgment. You’re the one who’s supposed to have it. Customers can tell you where it hurts. It’s your job to prescribe the cure, not to hand them the prescription pad and ask them to fill it in. Henry Ford almost certainly never said the line about faster horses, but it survives because it’s true in spirit: people describe their problems in the language of what already exists. Your job is to hear the problem underneath the request and answer that instead.

 

The version that actually works

So here’s the swap. Instead of “what do you want,” ask better questions and then trust yourself with the answers. What’s the most annoying part of doing this today? What did you try before us, and why did it let you down? What would have to be true for this to be a no-brainer? Those get you real problems instead of imaginary solutions.
Then you go away, and you decide. You build the thing you believe is right, you put it in front of them, and you watch what they do, not what they say. You’ll be wrong sometimes. That’s fine, being wrong fast on something real beats being vaguely right forever on a survey.
The businesses that win were almost never the ones taking the most orders from their audience. They were the ones who understood their audience well enough to give them something they didn’t know how to ask for, and were confident enough to be occasionally, usefully, unpopular.

 

Listen closely. Just stop taking dictation.