An index of what people are asking for

How the index is built

1. 01

   Collect

   Public text where people describe a problem in their own words: app-store reviews, product and developer forums, public discussion threads — in English and Vietnamese.

2. 02

   Extract

   Each post is reduced to the problem it states, with how strongly it is felt. Posts that state no problem are dropped.

3. 03

   Group

   Posts describing the same problem are grouped by meaning, however differently they are worded, and counted by distinct author.

4. 04

   Score

   Demand, pain, momentum and competition are computed by a fixed formula. No language model sets or adjusts a score.

The full formula, thresholds and sources are on the scoring methodology page.

Editorial standards

• A page asks to be indexed only once at least 8 different people have described its problem. Below that it stays readable but out of search.
• Every number links back to the quotes that produced it, each with its source, date and a link to the original post.
• Quotes are shown without the author's username. Pages that turn out not to be about a real business problem are taken off lists and out of search.
• App pages collect reviews rated three stars or below, on purpose: they are a record of complaints, never an app's store rating.
• Scores are research signals, not financial instruments or investment advice. Nothing here can be bought, sold or traded.