Weakspot
Competitive intelligence from your rivals' worst reviews.
The most honest document any company produces is the one it did not write. It is the pile of one-star reviews sitting under its own product, and almost nobody reads their competitor’s.
The structural shift
Complaint data used to be unusable at scale. It is unstructured, emotional, duplicated, and full of people who are angry about the wrong thing. Reading ten thousand of them well took a team. It now takes an afternoon.
That change is quiet but it is total. A dataset that was economically unreadable is now economically trivial to read, and the market has not repriced it. Everyone is still buying survey panels and win-loss interviews, expensive instruments for capturing what people will admit to a stranger, while the unfiltered version sits in public.
The wedge
Not sentiment. Sentiment analysis on reviews is a solved and worthless product: it tells you a score you already knew.
The wedge is the specific, repeated, unfixed complaint. One angry review is noise. The same structural complaint appearing across two years of a competitor’s releases is a roadmap they have chosen not to build. That is the gap you can walk into.
The framework: the complaint ladder
Complaints sort into four rungs, and only one of them is worth anything.
- Expectation. The buyer misunderstood the product. Tells you about their marketing, not their product.
- Execution. It broke. Tells you about their quarter. It will be fixed.
- Design. It works exactly as intended and the intention is wrong. This is the rung that matters. Design complaints survive releases, because fixing them requires the company to disagree with itself.
- Identity. The customer is not who the product is for. Tells you where the segment boundary actually sits, which is usually not where the pricing page says it is.
Rung three is the only durable signal. Everything else decays.
My position
The interesting product is not a review scraper. It is the thing that separates rung three from the other three rungs reliably, and then watches whether it persists across releases. Persistence is the whole moat. A complaint that survives two shipping cycles is a strategic choice, not a bug.
The prediction
Within a few years, reading competitors’ complaint corpora becomes a standard line in product diligence, the way traffic estimates became standard in the 2010s. The teams that get there first will not be the ones with the best model. They will be the ones with the best taxonomy.