Ask Where It Broke

The demo ends.

They smile.

Then they say, "Looks great."

You close the call with a warm little glow. Two days later, they do not buy. Now the praise feels less like evidence and more like a curtain someone pulled across the useful part of the conversation.

So you blame the follow-up. Or the price. Or the buyer's lack of urgency. Anything except the question that made honesty awkward.

"What do you think?" sounds open. It is not. It asks another person to judge the whole thing, judge you for making it, and deliver the verdict to your face. Most decent people will reach for the softest answer in the room.

Then you take their kindness back to the desk and spend another week polishing the work they never understood well enough to reject.

Praise Is Cheap Data

General questions produce general answers. Did you like it? Would you use it? Does this make sense? Each one invites a tidy opinion after the fact.

But buyers do not experience your work as one tidy opinion. They meet it moment by moment. A headline makes them lean in. A label makes them pause. A claim asks for proof. A button makes them wonder what happens next.

The useful truth lives inside those small reactions. That is why the Nielsen Norman Group calls thinking aloud a core usability method. You learn by watching someone attempt the task while saying what they notice, expect, and misunderstand. The value is not a polished review. It is the break happening in front of you.

Praise tells you the person wants the conversation to end well. A visible hesitation tells you where the work stopped carrying them.

The Friendly Room Is Rigged

Your favorite reviewers know too much. They know what you meant. They know the feature is unfinished. They know which rough edge to forgive. Their context quietly repairs the work while they look at it.

Then a stranger arrives with none of that mercy. They do not translate the vague headline. They do not hunt for the hidden promise. They do not reward the clever structure that took you three nights to perfect. They simply miss the point and move on.

This is why a friendly audience can make bad work feel ready. The room is efficient at protecting your feelings and useless at exposing the fault.

The UK Government Digital Service guidance on moderated usability testing makes the distinction plain: give people realistic tasks and watch what they do without leading them. You are testing the thing, not the person. Rescue them too early and you erase the evidence you came to collect.

That will feel slow. Good. Efficiency is not the prize here. Effectiveness is. Ten quick compliments can leave you blind. One quiet minute watching someone fail to find the next step can save the whole offer.

Use a Break Report

Put one real task in front of one person who did not help build it. Do not explain the clever parts. Do not tell them where to click. Let the work make its own case.

When they slow down, ask three exact questions.

1. What did you expect here? The gap between expectation and reality shows you which promise the work accidentally made.
2. What made you pause? A pause catches friction before the person turns it into a polite explanation.
3. What would you do next? Their next move reveals whether the path is clear or whether they are guessing to please you.

Write the answer as a Break Report: the task, the exact moment, what they expected, what happened instead, and what they did next. No score. No defense. No committee trying to decide whether the person was right.

This is observational evidence, not a vote. Digital.gov's current usability guidance recommends task scenarios, open follow-ups, and non-leading questions to expose where problems appear. You are not asking one person to design the answer. You are asking reality to locate the damage.

Fix the First Break

Here is where smart builders get greedy. They collect every comment, build a grand redesign, and vanish for a month. The feedback becomes another excuse to avoid contact with the market.

Do less. Fix the earliest break that blocks the task. Then put the new version in front of someone else.

The earliest break matters because everything after it is contaminated. If the buyer misunderstands the promise, their opinion of the features cannot rescue you. If they cannot find the next step, asking whether the checkout feels trustworthy is theater.

You may discover that the product is not the problem. The promise is vague. The proof arrives late. The task begins in the wrong place. That can sting because you worked hard on the part nobody reached.

Take the sting. It is cheaper than another month of applause from people who were never going to buy.

Make Truth Easy

The next demo ends. This time, you do not ask whether they liked it. You give them a task and keep your hands off the wheel.

They pause at the second screen. You feel the old urge to explain. You let the silence stay.

Then they tell you what they expected. There it is: one small break, clean enough to fix, real enough to matter.

The room may feel less kind. The work will get better. And the next time someone says "looks great," you will know praise is the greeting, not the evidence.