rank tracking

What a “position” actually means

Google’s Search Console defines position as the topmost spot your link held for a query, averaged across every query where it earned an impression. If your page shows at 2, 4, and 6 in one result set, it counts as 2, and a link that never gets an impression records no position at all. That is why your Search Console average position (real, blended, your-own-users data) rarely matches a third-party rank tracker, which probes the SERP from a fixed location and device on a clean, logged-out session.

Neither number is “wrong.” They answer different questions. Search Console tells you where real users saw you; a rank tracker gives you a controlled, repeatable snapshot you can compare day to day. Expect the two to disagree, and expect any tracker to drift a few spots from what you see yourself, because results are personalized by location, search history, and device, and positions past the top few are volatile.

Good trackers now record more than a single blue-link number: featured snippets, local packs, and increasingly AI Overviews and ChatGPT or Perplexity mentions, tracked per keyword, per city, and per device. Two habits make the data usable: tag keywords by intent so you watch commercial terms separately from informational ones, and bucket rankings into bands like top 3, 4-10, and 11-20 so you measure share of visibility instead of one vanity number. Tracking a small competitor set alongside your own helps a drop read as “we slipped” versus “the whole SERP moved.”