AdminScoring config

Trust

BetterScore configuration

The weights, the confidence curve and the recency window. Changing anything here moves every product on the platform, so every change is logged with a rationale and announced before it takes effect.

Changes here are platform-wide

A weight change recalculates every leaderboard on the next scoring pass. Run a preview first — it reports how many products would change rank and by how much, without writing anything.

Weights total

100%

Recency half-life

120 days

Evidence older than this counts for less

Last change

28d ago

Problem resolution weight 18% → 20%

Signal weights

Must sum to 100%. The published version of this table is what buyers see on the methodology page.

User satisfaction

Verified rating, recommendation rate, sentiment

%

Problem resolution

Resolved rate, response speed, user confirmation

%

Review trust

Verified users, depth, anomaly checks

%

Use-case fit

Primary-use votes, workflow match

%

Community activity

Helpful replies, Q&A, founder participation

%

Product momentum

Recent improvements, sentiment trend, release activity

%

Popularity

Follows, saves, likes, qualified profile visits

%
Saving schedules the change for the next scoring pass and publishes it to the methodology page.

Confidence curve

How much a product's own signals count before it has enough evidence.

TierReviews in the jobDamping factorPublic label
Provisional0–9Provisional
Low10–19Low
Medium20–39Medium
High40+High

Recency

Hard invariants

Configuration history

ChangeDetailActorWhen
score.weights.updatedProblem resolution weight 18% → 20%Harriet Vance28d ago