Performance · 2026/09/28

Google documented ad metrics. That is not a cue to put ads on a corporate site

Mid-September Chrome added experimental metrics to CrUX: how many ads in view, how much of the viewport they cover, how much network and CPU they burn. Google then listed that line under “understanding page experience in search results.” Search Engine Land asked John Mueller. He said the metrics are there because you can measure them and talk about the site. Not everything is a ranking factor.

CrUX is field data from real Chrome sessions, not a lab score from one run. These four numbers are for pages that actually run ads. Chrome also writes that aggregate reports tend to include origins with an ads.txt file that lists authorised sellers. ads.txt is how the ad industry marks who may sell a slot. A brochure site that does not sell inventory should not have that file. On that kind of site the panel is often blank for this row. Blank means the experiment could not measure you. It does not mean search deducted points.

You do not need to hunt in Search Console for the old page-experience chart. That report is already gone. To see field data for your own site, the person who maintains it opens the CrUX lookup, or the field data in PageSpeed Insights. If the ad-metric row is missing, skip it. Spend the time on LCP, INP, and whether the form moves after the guest taps Submit.

What most small-firm sites should still watch is load, and whether a tap works. A phone that will not open the page, or a form that does nothing when tapped, loses more business than ad density. Chat widgets, analytics, and webfonts all drag INP: the guest taps Submit, the page is busy with other scripts, the score looks fine, and the mail never leaves. A publisher that sells ads can use CrUX to argue for one fewer overlay. An ordinary company site does not need AdSense because this document grew a sentence, and it should not invent a fake ads.txt for an experiment.

The experience signals closer to search are still the familiar Core Web Vitals, including INP, not this ad row. If the lab score looks tidy and Submit stalls on a phone, trust the tap you just made. In our GA4 window from late August into September, active users were few and the home page bounced a lot. A form that arrives is closer to actual work. Ad density on a site with nothing to sell can stay blank. If you add AdSense because of this document, you add another ad script on the site and the form is more likely to slow down. A company site with no inventory was not the audience for this row. When speed needs work, the person who maintains the site starts with the scripts that stall a tap, such as the chat widget and analytics, not with the ad row.

In practice the maintainer does three checks. Open the contact page on a phone, tap Submit once, and see whether the button stalls and whether the mail reaches the named inbox. Then read successful forms in GA4. Do not stop at bounce. If a chat widget sits on the button, remove it or load it later. Speed and structure: SEO/GEO.

Two easy misreads

Someone forwards the help-doc screenshot: ads are now mentioned under page experience, so should the company site grow a row of ads? Mueller’s answer already blocks that. The metric is in the document. It does not mean a brochure site with no ads should invent ads. The reply is: we have no slot to sell, so this row can stay blank.

The other scene is a contact page with chat and analytics, where Submit stalls on a phone. In that GA4 window our home page bounced a lot. A form that arrives is closer to work than an invented ad-density score. The person who maintains the site gets that mail through first.

Check whether the form stalls Structure

Sources: Chrome on CrUX ad metrics; SEL / Mueller.