Shiny red ball

There is a loop that developers fall into and it goes like this.

You build in private for a while. The thing gets good. Genuinely good, better than it was. Eventually you show it. The response is thin: some polite interest, nobody buying, a couple of people who say it looks great and then do not reply. And you conclude, reasonably, that it is not ready.

So you go back and shine the ball. Six more weeks. Show it again. Same result.

What is actually being rejected

Almost never the polish.

When somebody does not buy, it is usually one of a small number of things. They do not have the problem. They have it and it is not expensive enough to act on. They have it, it is expensive, and they already do something about it that is good enough. They would buy but not from a company that might not exist next year. Or they would buy and you never actually asked.

Only one of those is about quality, and barely: the "good enough" one, which is a comparison rather than a standard. The others are about who you picked and what you said.

Polishing addresses none of them. It is a real improvement to the thing, aimed at a problem that was not raised.

Why polish is the default response

Because it is the only response entirely within your control, and because it is the part you are good at.

Shipping a rounder corner is a certainty. Asking somebody why they did not buy is not. They might not reply, and if they do the answer might be that the premise is wrong, which is a much worse afternoon than a refactor.

So the loop persists, and from the inside it feels like diligence. Six weeks of hard, skilled, genuinely useful work, none of which touches the reason nobody bought.

Doing it the other way round

Show it earlier than is comfortable and find out which of the five it is.

The good news is that you can usually establish this in a week. Five conversations, each one asking about the problem before the product. Do they have it. What does it cost them now. What do they do about it today. If the answer to the first is no, all the polish in the world is irrelevant and you have learned that for the price of five emails rather than six weeks.

If the answer is yes and they still do not buy, then ask the second question, which almost nobody asks: what would have to be true for you to pay for this? People answer it. The answer is rarely "make it nicer".

The thing to notice

If you have shown it twice and gone back to building twice, the third round is not going to work either. It is the same experiment run again with a nicer finish, and you already have the result.

Stop shining it. Go and find out which of the five it is. That question has an answer, and the answer is not in the codebase.