On one of the two properties we measured, 55% of Search Console impressions came from section anchors rather than from pages. They produced four clicks.
Search Console logs a jump-to-section link as its own page row, so a single appearance in results can be counted five or six times.
Here’s what that looks like on two sites, what it does to the numbers you read, and how to remove it.
What We Measured
Ninety days, two properties on the same account, page-level export.
| Property A | Property B | |
|---|---|---|
| Anchor URLs | 73 across 13 pages | 148 across 25 pages |
| Their share of impressions | 1.3% | 55.5% |
| Their clicks | 0 | 4 |
| Median anchors per page | 5 | 6 |
| Reported average position | 40.4 | 6.0 |
| Position with anchors removed | 40.9 | 9.1 |
The share differs by a factor of forty between two sites on the same account. That’s the part worth sitting with, because it means you cannot assume your own figure from anyone else’s.
Why One Appearance Becomes Six Rows
It happens because a result can carry links to sections within the page, and each of those destinations gets reported as a row of its own.
The page ranked once. The report shows the page, plus an entry for each section link beneath it, and every one of them carries the impression.
The count matches the mechanism
Median anchors per affected page came out at five and six across the two sites. That’s the same order as the number of section links a result tends to display, which is what you’d expect if each one is being logged.
What It Does to Your Numbers
Three figures move, and one of them moves a lot.
- Impressions inflate. On property B the reported total was more than double the number of real appearances.
- Average position improves: anchors sit high, so they pull the site average toward a number nobody experiences.
- Click-through rate collapses. The denominator grows and the numerator does not.
Property B reported an average position of 6.0. With the anchor rows removed it was 9.1, three places worse and considerably more honest.
One page can dominate a whole property
The worst single case had seven anchors and produced 2,715,695 impressions, which was 52% of everything that property reported.
A number that large from one page is the tell. If a single URL family is carrying half your volume, check what it’s made of before you celebrate.
How to Strip It
You need the page export and one filter.
- Export the pages report. The interface aggregates, so the anchors are easier to see in the raw rows.
- Drop any URL containing a hash: that single rule removes the duplication.
- Recompute impressions, position and CTR. Compare the two versions and keep the second.
Do this before any reporting cycle where average position is a headline number, because that’s the figure the anchors distort most.
Why It Varies So Much
It swings that far because the effect follows page shape rather than site quality. Between our two properties the share went from 1.3% to 55.5%.
Anchors appear on pages split into short titled sections, and the effect concentrates wherever one page attracts high-volume queries. A site with a few such pages sees a rounding error; a site with one popular page of that shape sees half its report.
What This Does Not Prove
It doesn’t explain why Google emits these rows, only that it does and that they carry impressions. Nothing here says the underlying rankings are wrong, and the pages themselves are ranking normally.
Two properties is also two, not a sample. The forty-fold spread between them is the reason to measure your own rather than borrow either figure.
What we corrected
An earlier read of this data treated the anchors as evidence that page sections outrank their parent pages. The sections did sit higher on 38 of 38 pages, but 2.9 million impressions producing four clicks is duplication rather than a ranking insight.
Frequently Asked Questions
Why does Search Console show URLs with a hash
They’re section anchors on your own pages. When a result displays links to sections within a page, each destination can be reported as its own row carrying the same impression.
Are anchor impressions fake
They’re real appearances counted more than once. The page did appear; the report lists it five or six times over, once per section link shown beneath it.
Should I remove anchors from my pages
No. They help readers and the ranking is unaffected. The problem is in the reporting, so fix it in the report rather than in the page.
How much of my Search Console data is affected
Anywhere between almost none and more than half. We measured 1.3% on one property and 55.5% on another, so the only reliable answer is your own export.
Does anchor duplication affect clicks too
Barely. Anchor rows produced 0 and 4 clicks across the two properties, so clicks stay close to accurate while impressions and position do not.
Filter the Hash Before Your Next Report
Pull the page export, drop every URL containing a hash, and recompute. If your averages move, they were never describing what you thought they were.