Categories organize the market. They never receive a rank or score.
Public decision standard
A recommendation is only useful if you can inspect why it exists.
Niched publishes the limitation, source, date, confidence, and condition that could reverse a verdict. The method is public. The operational collection system is not the product.
Revenue is context. The buyer decision is the product.
Revenue databaseHow much is this company making?
A present-tense, supply-side snapshot published about a company.
NichedShould I build on this, and does that choice still hold?
Demand-side judgment using job fit, dated evidence, history, caveats, and change.
No single source is required to rank. Missing evidence is unknown, never negative evidence. When a product lacks a factor, that factor is excluded from its composite and the gap is stated on the verdict rather than redistributed into a number that looks complete.
Visible scorecard
4 weighted factors. No hidden precision.
Weights can change with the buyer's constraints. Missing evidence lowers confidence and is excluded from the composite instead of silently becoming zero. A factor earns a weight only when an approved source can populate it for at least 80% of the products in a niche; below that it is published as context.
Only a validated buyer decision receives a scored shortlist.
A filter narrows the same decision unless evidence shows that the winner changes.
The same product can win one niche and lose another. Niched publishes no global software leaderboard.
- 01Job fit · 40%
Does the product satisfy the buyer's exact job and constraints?
Sourced from: Documented capabilities, limits, and regional availability; First-party answers to the niche's constraint questions.
- 02Adoption · 20%
Is the product actually integrated into real codebases and businesses?
Sourced from: Official client SDK downloads; Stars and contributors; Verified revenue.
- 03Current momentum · 20%
Is the product improving or losing important ground?
Sourced from: Dated changelog and release cadence; Release frequency and commit recency; Pricing and packaging changes over time.
- 04Durability and risk · 20%
Can a buyer reasonably keep depending on it?
Sourced from: Ownership, funding posture, and stated support commitments; Maintenance activity and issue responsiveness.
What Niched does not score, and why
Niched has no approved source for this. The review platforms that hold the data prohibit the collection it would require, so this factor carries no weight rather than scoring every product neutral and pretending the rubric is complete.
Blocked sources: G2, Capterra. Under evaluation: Hacker News search API, Stack Overflow API, Reddit official API, First-party structured buyer reviews.
Carrying a weight Niched cannot populate would mean either silently redistributing it or scoring every product neutral. Both publish a number that is not real, so the factor is named here instead.
Different evidence answers different questions.
Facts may refresh automatically when the source and timestamp stay attached. Buyer-facing judgments always require human review.
Verified revenue context for participating products. Absence is never a penalty.
Official client SDK downloads. Ranked above repository stars for adoption because it measures integration and covers closed-source products.
Release cadence, repository activity, issues, and dated changelogs. Stars are a fallback adoption signal only, and are open-source biased.
Pricing, packaging, integrations, documentation, and public claims.
Launch timing and early attention, never durable quality on its own.
Useful first-party context, clearly labeled until independently supported.
Migration and outcome evidence, published only with consent or in aggregate.
Nothing removed. Nothing hidden.
No material correction has been recorded in this baseline. When a verdict changes, the previous claim, new claim, date, initiator, and reason remain public.
0 recorded correctionsPay for monitoring, never placement.
Buyers and founders can pay for alerts, history, and competitive tracking. Payment never changes rankings, evidence requirements, or editorial conclusions.
0 paid placementsRead the commercial disclosure policyPublication rules
The rules that cannot bend.
- Start with the buyer's job.
- Keep proof beside every claim, with source and date.
- Label estimated data and leave missing data missing.
- Let automation collect facts and humans publish judgments.
- Explain every material ranking movement.
- Score only inside a niche. Categories and products carry no portable score.
- Publish only after enough comparable products clear review.
- Keep corrections visible permanently.