Anonymized queries in Google Search Console

In Google Search Console, anonymized queries are the searches Google strips out of your query report because too few people made them. Their clicks and impressions still count in your page and site totals, just not against any named search term. On most sites they account for close to half of all clicks, which is why the queries you can see never add up to the totals at the top.

So when your query list comes up short, nothing is broken. You are looking at the deliberate gap Google leaves to protect the people behind rare searches. This sits underneath the broader split between page-level and query-level data: anonymized queries are the mechanism that makes those two views disagree.

TL;DR:

• Google hides queries that aren't searched by more than a few dozen people over a two-to-three-month window. The clicks stay in your totals; the search terms disappear.
• Across Ahrefs' April 2025 study of 22 billion clicks, anonymized queries averaged about 47% of all clicks, and 45% to 80% is the common per-site range.
• You cannot filter your way to them. Any query filter drops anonymized queries from the totals too, so they vanish from every filtered view.
• You can still measure how much you are missing: page clicks minus the clicks of your named queries equals your anonymized share. The Bulk Data Export gives the exact figure.

What counts as an anonymized query?

An anonymized query is one that too few people searched for Google to name it safely. The threshold is specific: Google omits queries that "aren't issued by more than a few dozen users over a two-to-three month period". Below that line, the search term is removed from your report, while its clicks and impressions stay counted in your totals.

The privacy logic is straightforward. A search rare enough that only a handful of people made it could, in theory, identify one of them, especially paired with the page they landed on. So Google drops the term and keeps the traffic. The threshold is a rolling one, measured over the trailing few months, so the same query can sit inside your report one month and fall out of it the next as its search volume drifts. The glossary entry for anonymized queries keeps the short version; the rest of this page is what the short version leaves out.

Why can't you filter your way to the hidden queries?

Because filtering removes them completely. Google keeps anonymized queries in your chart totals, in its own words, "unless a query filter is applied (for example, 'queries containing' or 'queries not containing' a given string)". The moment you filter by query, the anonymized clicks drop out of both the table and the total. No string you type will surface a term Google has decided to hide, because the filter runs against data the term is no longer in.

This is the trap most people fall into when they go hunting for missing clicks. They filter, the numbers shrink, and they conclude the data is thin. It is the filtering that thinned it. One distinction saves you here: a page filter keeps anonymized clicks in the total, but a query filter strips them out. That difference is exactly what makes the next step, measuring your hidden share, possible.

How much of your Search Console data is anonymized?

Usually about half, and sometimes far more. Ahrefs analyzed 22 billion clicks across 887,534 properties in April 2025 and found anonymized queries averaged 46.77% of clicks, with the most common per-site range running from 45% to 80%. The share has held near 46% every year they measured it, back to 2022.

The figure swings by page, though, and the reason is the long tail. A page that ranks for a few head terms loses almost nothing to anonymization, so its query list nearly reconciles. A page that ranks for a long tail of rare, specific searches loses most of its query data, because rare searches are precisely what the threshold removes. The more long-tail a page is, the less its query list tells you about its real volume. One honest caveat on the 47%: Ahrefs measured it before AI Overviews and AI Mode rolled out, and those surfaces add gaps of their own, so the hidden portion is unlikely to be shrinking.

How do you measure your own hidden-query share?

Subtract your named queries from your page total. Take a single page's total clicks from the page view, then add up the clicks of every query attributed to that page in the query view. The difference is your anonymized clicks. Because a page filter keeps anonymized clicks in the total while the query rows exclude them, that gap is the hidden portion, sitting in plain sight.

One caveat keeps this honest. In the Search Console interface the query table caps at 1,000 rows, so a page with thousands of distinct named queries will show fewer named clicks than it truly has, and the subtraction will overstate anonymization. For the vast majority of pages, which rank for far fewer than a thousand named queries, the estimate is close. This is the discipline QueryScope is built around: it reads volume only from the page level, where every click is counted, and never sums the query rows into a total they were never meant to fill. You can see the same gap on a live account in 30 days of Search Console on a real SaaS, where the named-query list ran well short of the page-level totals.

Does the Bulk Data Export recover anonymized queries?

No. The BigQuery Bulk Data Export carries the same privacy filter as the interface. Every row has an is_anonymized_query flag, and Google's schema is blunt about what happens when it is set: "The query field will be null when it's true to protect the privacy of users making the query." You get the row, the clicks, and the impressions. You never get the words.

That sounds like a dead end, but it is the most precise measurement you have. In the export you can sum the clicks of every row where is_anonymized_query is true, grouped by page, and read your exact hidden volume with no 1,000-row cap and no subtraction estimate. The export answers "how much" down to the click, and still refuses to answer "which." That is the privacy threshold working as designed, not a limit you can engineer around.

What should you do about anonymized queries?

Stop trying to recover them, and read around them instead. A few habits follow directly:

• Take volume and trends from the page view. It counts every click, anonymized or not, so it is the only honest basis for "how much" and "did this move."
• Treat the query view as a sample of intent, not a tally of traffic. It tells you what a page ranks for and why, never how much, and it is how you spot an intent mismatch.
• Don't go hunting for the missing clicks. They are real, counted, and deliberately unnamed. Time spent recovering them is time lost.
• Store your own history. Search Console keeps only a 16-month window with no delta export, so the totals you trust today disappear unless you snapshot them first.

None of this makes anonymized queries a problem to solve. They are a fixed property of the data, like the 16-month wall or the one-to-three-day lag. Once you read each view for what it actually measures, the gap stops looking like missing data and starts looking like what it is: the half of your search traffic Google will tell you the size of, but not the name. That distinction is the whole job, and it is what QueryScope holds for you in the terminal, so you read the right number from the right view without keeping the books by hand.

Sources

• Google Search Console Help, Performance report: anonymized queries and query filtering (anonymized queries are included in chart totals unless a query filter is applied).
• Google Search Central, A deep dive into Search Console performance data filtering and limits (the few-dozen-users, two-to-three-month threshold; rows omitted from tables but counted in totals).
• Google Search Console Help, Bulk data export table reference (the is_anonymized_query field; the query field is null when it is true).
• Ahrefs, Anonymized Queries Make Up Nearly Half of Google Search Console Traffic (April 2025, 46.77% average across 22 billion clicks and 887,534 properties).