I showed five people and they all said it looked good - how do you get real feedback instead of politeness?

The highest-value thing, and the most uncomfortable: watch them use it in silence.

Give them a task: something concrete, phrased in their language rather than yours - and then say nothing. Do not help. Do not explain the button they cannot find. Do not say oh that bit is not finished. Sit on your hands.

What you learn in ten minutes of this exceeds anything anybody says afterwards, because you are observing rather than asking. Where they hesitate, what they click that you did not expect, the point at which they stop reading, the thing they assume it does that it does not.

It is genuinely painful and that discomfort is the signal. Every urge to jump in is a place the product is not clear.

Two practical notes:

Say out loud that you cannot help, at the start, so the silence is not rude. Something like: I am going to be quiet because I need to see where it is confusing, and being confused is useful to me.

Ask them to narrate. What are you looking for, what did you expect that to do. It feels unnatural for about a minute and then people forget.

Four people watched properly is worth forty people asked what they think.

You identified both mistakes: you explained first, and you asked hypothetical questions. Fix those two and the next four conversations are worth more than the first five.

Ask about the past, not the future. People are unreliable about what they would do and quite reliable about what they did.

Bad: would you use this. Would you pay for this. Do you like it.

Good: when did you last have this problem. What did you do about it. How long did that take. What did you try before. How much did that cost you.

If somebody cannot remember the last time they had the problem, that is your answer regardless of how positive they are about the product.

Do not explain it first. The moment you describe what it is for, you have told them what to say and removed your only chance to see whether it is comprehensible. Put it in front of them and say as little as possible.

Ask about their problem before showing anything. Ten minutes on how they currently do this thing, before they have seen your solution and started being nice about it.

And a rule that changes the tone completely: tell them you are trying to find out whether to stop. People are far more honest when the encouraging answer is not obviously the kind one.

And the unambiguous version, when you are ready for it: ask for something that costs them.

Encouragement is free, so it is abundant. The signals worth trusting all have a price attached:

Money. Even a small amount, even a pre-order, even a deposit. The gap between people who say they would pay and people who take out a card is enormous and it is the whole game.

Time. Will you spend twenty minutes setting it up with your real data. A yes here is worth a hundred compliments.

Their data or their work. Will you use it for the actual thing this week rather than a test.

An introduction. Who else has this problem, and would you introduce me. People do not put their reputation behind something they think is pointless.

Any of those getting a no, politely, tells you more than five yeses to would you use this.

On your two who said they would use it and have not: go back and ask what happened. Not to pressure them, genuinely to find out. The answer is usually something specific and useful, and it is a question people answer honestly because the moment has passed and there is nothing to be polite about.