Transparent discovery
How the rankings work
Rankings describe activity on StartupReader. They do not establish product quality, business performance or the strength of a moat.
Signals and freshness
The stored trending score combines upvotes, profile views and the reach of embedded badges. Time decay gives newer profiles a chance to surface. The score is refreshed by the ranking job; a weekly label indicates the current calendar week, rather than a full reset of all signals.
score = (upvotes × 1 + views × 0.02 + badge reach × 0.5) / (age in hours + 2)^1.6Badge reach uses this week’s impression counts from sites whose link to the profile has been verified. Each site contributes log(1 + impressions), with impressions capped at 5,000 per site. A nofollow link counts equally; badge links do not buy preferential website links.
Limits of the signals
Views and upvotes are activity signals, rather than a unique-reader census. They can include repeated activity, and legacy views may include automated traffic. A score does not prove customer adoption. Where scores tie, the ranking list uses upvote count as a secondary sort.
Promotion and editorial work
Paid placement contributes nothing to the ranking formula. Promoted profiles are labelled wherever placement is sold. Payment does not change verification, article review, claim labels or source evidence.
Compare the product’s audience and approach, read the underlying sources and check the open questions on its profile before drawing a conclusion.