How scoring works
Capcreds scores how powerful entities — companies, executives, public bodies, and other notable actors — affect the public. Here is how those scores are built and how an entity can respond.
Facts versus opinions
Every score rests on two clearly separated layers. The facts are the news events themselves, drawn from published reporting and linked to their original sources. The opinions are how the community evaluates the impact of those events. We never present a community rating as a fact, and we link the sourced coverage so anyone can check it.
How the community evaluates impact
For each article, members rate how the entity affected ordinary people on a five-point scale, from very harmful to very helpful. Each entity page shows the full distribution of those ratings — not just a single number — so you can see whether judgment is one-sided or contested.
Votes are weighted by how central the entity is to the article (its “focus”) and by the integrity safeguards below. The headline number is a weighted average of these community evaluations — it is not an editorial verdict from Capcreds.
Keeping scores honest
A score is only as trustworthy as the votes behind it, so several safeguards work to resist manipulation. Each person gets one vote per article for a given entity, enforced at the database level rather than in the app. Vote influence is bounded: an ordinary vote counts once, votes flagged as suspicious are automatically down-weighted, and no ordinary account can cast an oversized vote. Voting activity is logged and risk-scored so coordinated or automated campaigns can be detected and discounted. These controls shape how much a vote counts; they never let anyone — including Capcreds — set a score by hand.
Recent events count more
Scores are recency-weighted: recent coverage carries more weight and older events gradually fade. This means a score reflects how an entity is behaving now, and an entity can earn a better score over time by doing better.
Confidence
Each score carries a confidence level based on how many evaluations stand behind it. A score built on a handful of votes is shown as low confidence; confidence rises as more of the community weighs in.
If you represent a scored entity
Verified representatives can respond. Once verified, a representative’s comment is pinned to the top of the relevant discussion and clearly badged. A representative can add context and rebut coverage, but cannot edit or remove the score — the community keeps control of the rating. To request verification, open the entity’s page and choose “Represent this entity.”