Why does Search Console say I rank higher than I do?

The reported figure is the topmost result, averaged across queries

Two mechanics sit inside that average, and the second one surprises people.

Google defines it precisely: "The position value shown in the Performance report is the topmost position occupied by a link to your property or page in search results, averaged across all queries in which your property appeared."

Topmost is doing work there. If your site appears three times on one results page, only the best of the three is recorded. The lower two vanish from the number entirely.

Google supplies its own worked example, which is worth reading slowly: "Query 1 returns your property at positions 2, 4, and 6: its position is counted as 2 (the topmost position). Query 2 returns your property at positions 3, 5, and 9: its position is counted as 3 (the topmost position). The average position across these two queries is (2 + 3)/2 = 2.5."

Six appearances, two recorded, and an average of 2.5 for a site that was often much further down. Nothing is being hidden. The metric is simply answering a narrower question than the one you asked.

Deep rankings are under-recorded, which lifts the average

This is the part Google states plainly without spelling out the consequence, and the consequence matters.

The rule: "A link must get an impression for its position to be recorded."

Now reason from it, as reasoning rather than as something Google asserts. An impression requires somebody to have seen the result. Picture a result sitting on the third page of results. The search happens, nobody goes past page one, and the result is never seen. No impression means no recorded position for that query.

So the worst placements are the least likely to be counted. Queries where you ranked badly contribute nothing, while queries where you ranked well contribute every time. The average is computed over a sample that systematically favours your better days.

That is not a flaw to complain about. It is a reason to treat the figure as a trend line rather than a report card, which is exactly what Google recommends: "We recommend that you monitor change in position over time, particularly sudden position changes, as well as absolute position."

The same words mean different things in different places

A position number is not a distance down the page, and the phrase itself is used for something else in advertising.

Within search, the number describes an element rather than a pixel. Google is explicit that "only elements containing at least one non-query-refinement link occupy a position", So a carousel of query refinements does not push your result down the count. Google offers its own warning, which is worth keeping: "a position number can mean different things in different situations, and so you should not make simple assumptions."

The advertising sense of the phrase is a separate thing entirely, and on one platform it no longer exists. Microsoft documents its ads metric and its retirement in the same breath: "AveragePosition The average position of the ad on a webpage. This column is deprecated as of March 31st, 2021. Going forward, performance reports will return average position of 0 (zero)."

Worth knowing because the phrase travels between conversations. If somebody reports an average position of zero from an advertising report, that is a retired column returning its placeholder, not a ranking achievement.

The query view has limits of its own

Filtering to one query is the right move, and it is worth knowing what it costs before you reconcile totals.

Some searches never appear. Google explains why: "Some queries are omitted from the report to protect user privacy. These are called anonymized queries."

Those omitted searches are not simply missing. They sit in the charts but not in the table, and applying a filter removes them. Google states the behaviour: anonymized queries are "included in chart totals, unless a query filter is applied".

Read the consequence. The moment you filter by query, the totals drop, because a group of searches that was quietly counted stops being counted. So add up your filtered views and they will not reconcile with the unfiltered number. That is by design rather than by fault.

There is a second limit on top of that: "Due to internal limitations, Search Console stores and shows only the most important data rows. Not all queries beyond anonymized queries are shown in the table."

So the query list is a selection, not a census. For a small site that hardly matters. For a site with a long tail of rare searches, most of that tail is simply absent. An audit built on the visible rows will describe a narrower business than the real one.

None of this makes the report untrustworthy. It makes it a sample with published rules, which is a better thing to have than a number with none.

What to check, in order

Five checks, and all of them are free.

• Stop testing by searching. Your own search carries your history and location, and Google names both as reasons the number will differ.
• Read the figure as a direction rather than a place. Compare this month to last month rather than to what you saw on your phone this morning.
• Filter to a single query before drawing conclusions. An average across every search your site appeared for is not a ranking for the one you care about.
• Check whether the page appears more than once for a query. Only the topmost appearance is recorded, so the number can improve when a second listing disappears.
• Treat a missing position as information. Google records nothing when there was no impression, so absence usually means nobody scrolled far enough to see you.

One piece of arithmetic to settle the most common argument, using Google's own example shape. Suppose a page ranks 3rd for one query and 30th for another, and both get impressions. The reported average is 16.5, a figure that describes neither search and matches nothing you could ever see on a screen.

Now suppose the 30th placement stops earning impressions because nobody goes to page three. The recorded average improves to 3, and the page has not moved at all. An average that rises while nothing improves is the normal behaviour of this metric, not a glitch, and it is the single best reason to look at query-level data before celebrating.

Reading these reports so the numbers answer the question you actually asked is part of what we do on search.

Terms used on this page

• Average position

Sources

• What are impressions, position, and clicks? (Google Search Console Help)
• Report attributes and performance statistics (Microsoft Advertising, Microsoft Learn)
• Performance report: Dimensions and data groupings (Google Search Console Help)

Last reviewed 2026-09-12.