Rank Tracker now cross-checks each incoming ranking against a keyword's recent history before recording it. If a result looks like an outlier — a sharp spike or drop that doesn't match the recent trend — it retries the fetch automatically before locking in a value.
This reduces the chance that a bad data center reading or a Google A/B test lands in your client's report as a real ranking movement. Keyword positions should now hold steady unless there's a genuine change worth reporting on.