The review wasn't about the product

The most candid writing about your market is not about you. It is sitting in reviews of the products your buyers considered instead, written by people who had no idea you existed and no reason to be tactful.

Most founders skim these for competitive intelligence: what the rival does badly, what to put on a comparison page. That is the least valuable thing in them. The valuable part is that a review is a record of the gap between what somebody expected and what they got, and the expectation is the part that tells you about the market.

Most reviews are not about the product

Take an illustrative three-star review, of the kind that exists in every category:

“Does what it says. Took me a while to work out what it was for, and in the end I went back to the spreadsheet.”

Read as a product review, that is mild and unhelpful. Nothing is broken. The reviewer is not even angry. A competitor analysis would file it under neutral and move on.

Read as market signal, it contains two specific findings, and neither of them is a complaint about the product. That is the reframe the whole method rests on: reviews record the collision between an expectation and a thing, and the expectation was formed out there in the market, before anybody arrived.

The two findings hiding in one review

Finding one: a positioning gap. “Took me a while to work out what it was for” is not a usability note. It says the product failed to communicate its own purpose to somebody who had already decided to try it. That person was as motivated as a visitor ever gets, and they still could not place the thing. If a category is producing that sentence repeatedly, buyers are arriving without a clear frame for what these products are, which is a market-level fact and an opportunity for whoever states it plainly first.

Finding two: a named alternative. “I went back to the spreadsheet” identifies what the product was actually competing with, and it is not another company. It is a habit that is already installed, already understood, free, and has nobody marketing it. That alternative does not appear on anyone’s competitor grid, because it is not a competitor in the sense a grid understands. It is still what won.

We wrote about that specific blind spot in the piece on competitor lists made of companies. Reviews are the cheapest place to find the evidence for it, because people name what they went back to without being asked.

Why the star rating is the least useful part

The rating is the first thing you see and the last thing worth reading.

A star rating compresses a specific person, with a specific expectation, in a specific situation, into one number, and then averages it with hundreds of others whose expectations were completely different. Two products with the same average can be failing entirely different people for entirely different reasons. The average tells you neither.

Worse, the rating invites the wrong question. “Are they rated better than us?” is a scoreboard question. “What did people expect that they did not get?” is a market question, and only the second one changes what you do on Monday.

The three-star reviews are usually the richest. One star is often an incident, a billing problem or a bad support day. Five stars are often enthusiasm without detail. Three stars is where somebody liked it enough to be fair and was disappointed enough to be specific.

Reading a set without fooling yourself

One review is an anecdote. The method only works across a set, and there are a few disciplines that keep it honest.

Read competitors, not yourself. Your own reviews are written by people who already chose you. The market is in the reviews of the products chosen instead.

Collect phrases, not verdicts. Write down the actual words people use for the problem. Those phrases are what buyers type into a search box, and they are almost never the words on your own homepage.

Count what repeats, and notice what repeats across different products. A complaint that appears about one product is about that product. The same complaint about four products in the category is a fact about the category, and that is the one worth acting on.

Log every alternative anyone names. Including the non-products. Spreadsheets, a person they hired, an existing habit, doing nothing. That list is a better map of your competition than any research report.

Do not skip the positive ones. A five-star review that says what finally made it click is describing the moment the value became legible, and that moment is usually missing from your own page.

What this method cannot tell you

Reviews are evidence, not a survey, and the difference matters.

People who write reviews are not a representative sample. They skew toward the annoyed and the delighted, and the large quiet middle is absent, which is the same sampling problem that makes your own feedback folder misleading. A review tells you why one person left. It cannot tell you how many others did, and treating a vivid complaint as a measure of frequency is the most common way this method goes wrong.

Reviews also age. A complaint from two years ago may describe something that has since been fixed, and a category can reposition itself underneath a body of old text.

So read them for what they are good at: naming the expectation, naming the alternative, and supplying the vocabulary buyers actually use. For the question of how much, they are the wrong instrument, and anybody selling you a sentiment percentage off a review corpus is answering a question the corpus cannot support.

Reading a set of these properly takes an afternoon per competitor, which is the only real reason it does not get done. MarkLens reads them alongside search behaviour and what sits beside you on a shelf, weights them against each other, and states which findings are actually sourced and which are inference. Where the evidence is thin, it says so rather than turning a handful of reviews into a confident number.