Moderator accuracy measures how often a moderator's decision, approve, remove, or escalate, matches the correct outcome under the platform's policy. It's usually shown as a percentage, based on comparing a sample of decisions against a trusted reference set or a senior reviewer's judgment.

‍

‍Why it matters

Moderation calls aren't always black and white. Context, intent, and culture can lead experienced moderators to different answers on borderline cases, so accuracy is how a platform checks how consistent its enforcement really is. Under-enforcement leaves harmful content visible, and over-enforcement removes legitimate content and costs user trust. A drop in accuracy in one policy area is usually a sign the guideline needs clarifying, not that moderators should be blamed one by one.

‍

‍How it's measured

Teams pull a sample of each moderator's decisions, compare them to a gold standard set or a senior reviewer's call, and score each as a match or a miss. Results are tracked by policy area and over time, and anyone or any area below the agreed threshold gets retraining or clearer guidelines. Looking at accuracy by severity matters too, since a miss on a serious violation costs more than a miss on a minor one.