Why a score changed, and what can cap it

Scores are a recent-window read

Each scoring pass reads the creator's recent content — by default the last 180 days — across every account in the footprint, plus the audience signals around it. Scores refresh every 30 days, so the window slides forward and old content stops counting.

Why it went down

The report always says. Open the score and look at the per-agent lines and the What changed notes: a new risk flag, a drop in audience quality, a run of negative comment sentiment, or a disclosure habit slipping. Every flag names the exact posts behind it.

Score caps

Some findings do more than lower the number — they put a ceiling on it while they are live, for example confirmed hate speech or a large bot-follower share. When a cap is active the report shows the cap, the finding that caused it, and the evidence. Caps lift when the finding ages out of the window or is retired as a false positive.

Missing data never counts against anyone

A platform we could not read, a private account, or an account with no recent posts does not lower the score. We only score what we can actually see.