Before anything is designed, find out what exists, why it fails, and what one sentence explains it. The output is a diagnosis with sources and a thesis.
Facts
- Every claim carries its source. A link, a store listing, a public filing, a page anyone can open. If it cannot be checked in a minute in front of whoever decides, it is an opinion and is labeled as one.
- Measure what exists before imagining what could. What the client uses today, what it costs, who uses it and how much. "There is no app" is a finding; "the app is bad" is a guess until it is shown.
- Numbers come with their source and their date. Downloads, ratings, prices, sizes and revenue age fast. A number without a date cannot be compared with the next one.
- Assumptions are written next to the figures they shape, never folded into them.
The competition
- Read what their users say, not the score. Ratings say how many are unhappy; reviews say why. The repeated complaint is the product gap.
- Look at what the best competitor already proved for free. What works at their scale is a pattern; what their users keep asking for is the opening.
- Separate who pays from who uses. Many products fail because the person who benefits is not the one who buys.
What changes the conversation
- Look for the fact that reframes the pitch. Sometimes the current tool costs so little that any comparison on price loses; then the argument is about category, not cost. Find that before the conversation, not during it.
- Note what already died and why. A previous attempt that failed for lack of a payer or a maintainer is the most useful finding in the folder.
Before ruling anything out
- Research before saying no. An idea that sounds blocked is often blocked only by an assumption. One search tends to remove the doubt and sometimes brings the best decision of the project.
- If a real limit remains, state it once with the closest path that works.
The thesis
Close with one sentence: why what exists fails, and what the product resolves instead. A good thesis moves the problem away from the obvious feature. A karaoke room does not fail for lack of songs; it fails because nobody wants to go first. Every feature that follows has to trace back to it; kero-scope takes it from here.
Output
A diagnosis in docs/research.md inside the project folder, creating the folder when the project does not exist yet: the thesis at the top, then findings with their sources and dates, the competition read through its users, and the fact that changes the conversation.
Research stops when the thesis holds and new sources only repeat what is known. Say how many sources were read and which kind (stores, reviews, filings, vendor pages); a diagnosis built only from vendor pages is labeled as opinion.