Published in full, not as extracts.
A three-page teaser proves nothing. These are complete reports on real companies, with the source table intact, so you can check the work rather than take a view on the design. They are demonstrations — the verification column is unsigned, and each report says so on its own last page.
| No. | Report | Product | Read |
|---|---|---|---|
| 001 | UK meal kit market analysis 2026: why the leader is shrinkingSix competitors. The market leader is shrinking on purpose, and the discount ladder tells you where the real fight is. | Competitive Landscape | Open |
| 002 | Do smart ring subscriptions still make sense? 87 buyers answered87 coded buyer comments. The threat is not another ring — it is a $99 screenless band nobody is defending against. | Competitive Landscape | Open |
| 003 | Did our paywall cause the slowdown? What Duolingo's data showsShorter, written to be presented. What the leader's reversal says about your free tier. | Short briefFormat retired — kept as a sample of shorter work | Open |
Three different categories on purpose. The method does not change with the industry — only the query set does.
What every report contains
The bottom lines, first
Four ranked, falsifiable statements on page one, before any exposition.
Ranked actions, second
With effect, effort and owner against each, on page two.
The full source table
Every figure, where it came from, when it was accessed, and who checked it. Estimates marked as estimates, with the method and range stated.
The gaps, declared
Where coverage was thin, the report says so on the page rather than quietly routing around it.