Discussion starter: an SEO measurement question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.
- Search Console
- SEO Analytics
Rank tracker positions versus Search Console average position: which do you trust for decisions, and when do they diverge most?
Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.
1 reply
-
adminAdmin
The short answer is to use Search Console to decide whether something changed for real searchers, and a rank tracker to see what a specific results page looks like and who you are up against. They are not two estimates of one number. A tracker records a scripted search per keyword, from the location and device you chose. Search Console records the topmost position your site held in each real search where it appeared, then averages across those searches. So the tracker also leads where Search Console is thin: rankings below country level, low-volume queries and positions beyond page one. For decisions that involve money, neither position figure should lead: clicks, impressions and conversions should, with position as the explanation.
The big gaps come from five situations: comparing a tracked keyword with a page-level or site-level average, results inside SERP elements that Search Console counts as a single position, rankings that vary by location, device or personalisation, positions beyond page one, and reporting changes such as the end of the num=100 parameter.
What each tool actually records
Aspect Rank tracker Search Console What is recorded A scripted search per keyword per run The topmost position of your site or page in each real search where it appeared, averaged Queries The keywords you chose to track Every query that produced an impression, with rare queries withheld from query tables for privacy Location and device Fixed by your settings Every searcher, unless you filter by country and device Personalisation Typically a clean, logged-out session Whatever real searchers saw Timing A snapshot at run time Every search across the day SERP features Counted by the tool’s own rules Each element takes one position; links in an AI Overview share its position Depth As deep as the tool checks Only where your result appeared on a results page someone loaded Competitors Visible Not shown Google’s help on how impressions, position and clicks are counted sets out the rules that matter here. Position is the topmost position occupied by a link to your property or page, typically counted top to bottom on the primary side of the page and then on the secondary side. Each element in the results occupies a single position however many links it contains, and all links in an AI Overview are assigned the AI Overview’s position. Position in AI Mode follows the same methodology as a results page. An impression generally counts when your result appears on the current page of results, whether or not it is scrolled into view, as long as the searcher does not need to click to see it; inside carousels the item typically has to be scrolled into view or expanded. The same page notes that your own search can show a different position from the average because of variables such as search history and location.
One more rule catches people out. According to the Performance report help, data grouped by query, country, device or date is aggregated by property, while data grouped by page or search appearance is aggregated by page. The same query can show position 3 at property level and 8 on the row for the URL you care about, because another of your URLs ranked higher in some searches.
Where the gaps get big
1. A page or site average against one tracked keyword
A page-level average blends every query the page appeared for, and because it is averaged over searches, high-impression queries pull hardest. Gaining visibility for new long-tail queries at deep positions makes the average worse while traffic grows.
Hypothetical example:
A guide ranks second for its main query in your tracker. In Search Console it has 10,000 impressions at position 2 for that query and 5,000 impressions across long-tail queries averaging position 30, so its page-level average is (10,000 × 2 + 5,000 × 30) ÷ 15,000, about 11.3. You expand the guide and it starts appearing for new queries: 10,000 more impressions at an average of position 25. The page average becomes (20,000 + 150,000 + 250,000) ÷ 25,000 = 16.8. The tracker still says 2, clicks are up, and a dashboard shows average position worsening by more than five places. Both tools are right; the dashboard asked the wrong question.
2. SERP elements that hold many links
If you are cited in an AI Overview at the top of the page and also rank sixth organically, Search Console records the AI Overview’s position for that impression because it is your topmost, while an organic-only tracker reports sixth. Trackers differ on whether they count features as positions at all, so read your tool’s documentation before comparing. As of September 2026, Search Console’s “Search generative AI performance” report is available to eligible sites worldwide after its global rollout at the end of August 2026. It shows impressions, not clicks, for AI Overviews and AI Mode, which helps separate that effect from organic movement.
3. Location, device and personalisation
A tracker set to one city reports one local results page; Search Console averages every searcher, and it does not break results down below country level. For local-intent queries the spread between suburbs can be wide. The same applies to device: if the tracker runs desktop and most searches for the query happen on mobile, the two numbers describe different pages. Filter Search Console to the tracker’s country and device before comparing anything. If the gap persists on a local query, a grid of tracked locations tells you more than either single figure.
4. Positions beyond page one
Search Console only records a position when your result appeared on a page of results someone loaded, which for deep positions means someone went to that page. That creates a selection effect. This is an inference from the counting rules, not something Google documents: if your rank for a query varies between 6 and 14 across locations, the searches where you appeared on page one generate impressions for everyone who searched, while the searches where you sat on page two generate impressions only for the few who clicked through. The average is pulled towards your better placements, and a tracker sampling one location might say 12. For deep positions, treat Search Console as a sparse and flattering sample, and the tracker as the steadier view if your tool checks that deep.
5. Reporting and collection changes
When Google stopped supporting the num=100 results parameter in September 2025, rank trackers were disrupted, and many sites saw desktop impressions fall and average position improve in Search Console, while clicks were affected far less, according to Brodie Clark’s analysis. Anyone comparing September 2026 with September 2025 is straddling that change. A year-on-year position improvement across that boundary may be a reporting artefact, so check it against clicks before reporting it.
Which to trust for which decision
Decision Lean on Why Is this query group gaining or losing real visibility? Search Console, filtered by query group, country and device It measures what searchers saw; read clicks and impressions before position Did a release hurt a template? Search Console page groups, with a tracker sample as an early warning The tracker runs on fixed queries; Search Console’s 24-hour view shows hourly, preliminary data Where do we stand against named competitors? Rank tracker Search Console shows only your own site Which SERP features appear, and are we in them? Tracker for presence; Search Console’s AI report for AI impressions Different views of the same results page Local rankings by suburb Tracker with location settings Search Console stops at country Low-volume, high-value queries Tracker, cross-checked against Search Console when data appears Search Console may withhold rare queries or show too few impressions to read Forecasts and executive reporting Clicks, conversions and revenue Position from either source is context, not the outcome On speed, the gap between the two has narrowed. The Performance report’s 24-hour view plots hourly data points, including preliminary data that can still change, and the Search Analytics API has offered hourly data since April 2025. A daily tracker is still useful as an early warning on a fixed set of queries, but it is no longer the only fast signal.
Making the two sources useful together
- Compare like with like. One query, the tracker’s country and device, the same dates. For a stable head query the two should land close together. If they do not, check the tracker’s location and device settings, how it counts SERP features, and whether you appear in an AI Overview or carousel for that query.
- Treat the gap as a diagnostic, not an error. A widening gap on local queries points to location variation. A Search Console position far better than the tracker’s organic rank points to feature placements.
- Report position by query group, not as a site-wide average. A site-wide average moves with your query mix as much as with your rankings.
- Annotate measurement changes, such as the num=100 change or new search types and filters in Search Console, so nobody reads a reporting change as a ranking change.
A first comparison to run
Take your ten most valuable query groups. For each, put the tracker’s position next to Search Console’s position for the same country, device and dates, with clicks alongside, and label every gap with one of the five causes above. A gap you cannot label is the one worth investigating. A position change with no matching change in clicks does not belong in a forecast.
Need help with this on your own site?
Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.
Add a reply Cancel reply
Replies are open to approved members. Log in or apply to join.