Before you rebuild it, find out whether anyone saw it

Silence feels like judgement. It usually is not. And the difference matters enormously, because one of these problems is solved by months of building and the other is made worse by it.

Run the free check

A product that nobody has encountered has produced no evidence about its own quality, so silence cannot be read as rejection. Before rebuilding, establish how many people genuinely arrived, from where, and what they did. If the answer is that almost nobody arrived, the finding is about distribution and rebuilding will change nothing, because the new version will also be seen by nobody.

The trap, described plainly

You launched. Nothing happened. Rebuilding feels like the responsible response: clearly something was wrong, so fix it. It also happens to be the thing you enjoy and know how to do, which is why the instinct is so strong.

But consider what you would need in order to know that the product was the problem. You would need people to have used it and not come back, or to have arrived and left, or to have told you what was missing. Absent all of that, there is nothing to learn from, and a rebuild is a response to information you do not have.

The cost is not just the months. It is that the rebuilt version launches into exactly the same silence, because nothing about how people find it has changed, and after the second silence most builders stop.

The three numbers that settle it

You need only three figures to know which problem you have, and you can gather all of them today.

  • How many people who are not you visited in the last thirty days. Exclude your own traffic first, otherwise the number is fiction.
  • Where they came from. Direct visits from people you told personally are a different thing entirely from visits from strangers who found you.
  • What they did. Did they touch the product at all, or did they land on the page and leave within seconds.

Reading the three numbers

The combinations point clearly.

Almost nobody visited: you have no product evidence at all. The finding is entirely about distribution and a rebuild would be a guess dressed as a decision.

People visited but left within seconds without touching anything: the page is not explaining itself quickly enough. This is a messaging problem, not usually a product problem, and it is a comparatively cheap fix, most often the first sentence on the page.

People used the product and did not return: now you have genuine product evidence, and it is worth taking seriously. Even here, ask whether they had a reason to return, because some perfectly good tools are used once and that is correct.

People used it, returned, and there are simply not many of them: the product is working and the constraint is volume. Do not touch the product. Go and get more people.

The conversation worth having before any rebuild

If you have even five people who used the product, talk to all five, and ask about their experience rather than about features they would like.

The questions that produce useful answers are about what happened, not about what would be nice. What were you trying to do when you found this. What did you expect it to do. Where did you get confused. What did you do instead.

Feature requests, on the other hand, are a poor guide, because people ask for what they can imagine rather than what would change their behaviour. Five conversations about actual experience are worth more than fifty survey responses about desired features, and they take an afternoon.

When the product genuinely is the problem

It does happen, and it is worth naming the signals so you do not talk yourself out of a real finding.

The clearest one is people arriving with the right intent, understanding what the product does, trying it properly, and then not returning, repeatedly, across a reasonable number of users. That is real, and it is a signal about the product rather than about the page.

Another is people who say they want it, then do not use it. This is common and it means the problem you are solving is genuine but not painful enough for anybody to change their habits over.

In both cases the honest response is to change the product, and the rebuild is warranted. What makes it warranted is that you have observations behind it.

The order that saves the most time

For most builders reading this, the sequence that avoids the wasted quarter is straightforward.

Do the distribution work first, because it is a known method with a known shape and it costs weeks rather than months. Get readable, get indexed, buy the domain, write the pages, get listed, open the AI crawlers. Give it a quarter of continuous effort.

Then look at what the visitors do. At that point you will have the three numbers with real values in them, and the product question becomes answerable rather than speculative.

The order matters because distribution work is never wasted. Even if you eventually change the product substantially, the domain, the indexed pages, the listings and the citations all carry over. A rebuild done first, on no evidence, frequently is wasted.

Part of a larger guide

This page is one part of Nobody is using it. The other parts:

Questions people ask

How do I know if my product is bad or just unknown?
Look at three numbers: how many people who are not you visited in the last thirty days, where they came from, and what they did on the site. If almost nobody arrived, you have no product evidence at all and the finding is about distribution. Product conclusions require people to have actually used the thing.
Should I pivot if nobody is using my app?
Not until people have encountered it. A pivot based on silence is a guess, and the pivoted version launches into the same silence because nothing about how people find it has changed. Do a quarter of distribution work first, then read the behaviour of the visitors who arrive.
What does it mean when people visit and leave immediately?
Usually that the page is not making a promise the visitor understands fast enough. This is a messaging problem rather than a product problem, and the fix is often the first sentence: say plainly what the thing does and who it is for, above the fold, before any decorative element.
How many users do I need before I can judge the product?
Enough that a pattern is visible rather than a coincidence. More useful than a threshold is the shape of the evidence: people arriving with the right intent, understanding the product, trying it properly, and then not returning, repeatedly. That is a product signal. A handful of visits is not.
What should I ask the few users I do have?
Ask about what happened rather than what they would like. What were you trying to do when you found this, what did you expect it to do, where did you get confused, what did you do instead. Feature requests are a poor guide because people ask for what they can imagine rather than what would change their behaviour.
Is distribution work wasted if I end up changing the product?
Very little of it. The domain, the indexed pages, the directory listings and the citations all carry over to a changed product. A rebuild done first, on no evidence, frequently is wasted, which is why the distribution quarter comes before the product decision rather than after it.

Get the evidence before you spend the quarter

The free Get Found Check shows whether your product is reachable at all right now, which is the question you have to answer before any product decision means anything.

Run the free check Have Licheo do it for you