A fast score can hide the slow click

On 13 October 2026, Chrome added experimental Core Web Vitals to the Chrome UX Report. The new fields count soft navigations: the moment an app changes the page and the address without a full reload. They cover largest contentful paint, interaction to next paint, and cumulative layout shift. Chrome's release notes for the September dataset, published that day, say the stable scores are unchanged. Search Console is not switching yet. Chrome aims to take the new metrics out of the experimental stage in May 2027, if the feedback supports it, and Search Console follows after that.

If your site loads a fresh page on every click, Annie Sullivan of Chrome, as reported by Search Engine Journal the same day, says you should not expect your metrics to move. About 70 percent of origins in the report have little or no soft-navigation traffic. Between 66 and 75 percent, depending on phone or desktop, have soft navigations on fewer than 5 percent of page views. Only 7.0 percent of Android origins and 10.6 percent of desktop origins have soft navigations on more than half of page views.

The sites that will see a swing are the ones built as a single-page app, where JavaScript swaps the content and updates the address. Among the top 1,000 sites, Search Engine Journal reports that 35 percent on Android and 42 percent on desktop already take more than half their page views that way.

The swing can look like a gift. Chrome's own release notes, for all origins together, put a good Core Web Vitals result at 55.9 percent without soft navigations and 59.5 percent with them. Good largest contentful paint moves from 68.3 percent of origins to 72.1 percent. Layout shift moves from 81.7 percent to 85.1 percent. Interaction to next paint moves from 85.4 percent to 87.4 percent. The dataset covers 18,376,044 origins.

For the heaviest single-page apps the gap is wider, and this is where a careless reading starts. Search Engine Journal, citing Sullivan, says that among Android origins where 75 to 100 percent of page views are soft navigations, the share with a good largest-contentful-paint score is 26 percent on the stable metric and 87 percent on the experimental one. Sullivan's point, in that report, is that this does not mean single-page apps are faster than ordinary sites. Follow-up clicks are faster on any site. The old report charged later clicks back to the first load. The new one scores each view on its own, so a site with many later clicks starts to look like a multi-page site with many later clicks.

Interaction to next paint is the number to watch if your app feels sticky after the first screen. The experimental version scores each view, instead of keeping the slowest interaction of the whole visit on the landing address. Search Engine Journal says that roughly doubles the Android pass rate for sites that live on soft navigations, and that those sites still lag ordinary sites. Heavy JavaScript during a page change still shows up.

None of this is a reason to rebuild a company website as a single-page app so the chart looks kinder. The stable number in Search Console is the one a buyer can see today. The experimental number can still move month to month. Chrome said so. A May 2027 target is a target, not a date you can invoice against. The November release is expected to add experimental first-contentful-paint and navigation types. Wait for those before you treat a toggle in CrUX Vis as a verdict.

What you can do now is simpler, and it is the work a care plan is for.

Open the pages a customer actually uses: the service page, the price conversation, the form. On a phone and on a desk. If a click swaps the screen without a reload, write down which click feels late. That late click is the one the new metric is trying to see. It is also the one a buyer feels, whether Search Console has caught up or not.

Website care can include a check of those pages, the certificate, and the updates the site depends on. The written plan names what the care includes, and the price. We can look after a site we did not build, after a first look at the code and the host. Do not send passwords in the form. Messages on WhatsApp or email are answered Monday to Friday, 08:00 to 17:00 Western Indonesia Time. After hours, the reply is the next business morning.

Search work is separate. It is work on the pages you already have, so a person can find what you sell. The plan names the work and the price. We do not promise a rank, and we do not promise a date when a search engine will recrawl. A speed chart and a rank chart answer different questions. Mixing them is how a team rebuilds the wrong thing.

A company website still has to say what you sell, on a phone and on a computer. The plan names the pages, the price, and who owns the code before we start. We do not promise a load time for every country. If the job is an app rather than a site, the plan names what the app does. A single-page website and a mobile app are not the same document.

If someone hands you a screenshot of a greener score this month, ask which field it is. Stable, or the experimental soft-navigation field. If they cannot say, the screenshot is not a brief.

Tell us what you need. We reply by the next business day. You agree the written plan before we start. After handover, faults in the agreed work are fixed for at least 3 months. New features are not part of that cover.

Sources: Chrome UX Report release notes, 13 October 2026, https://developer.chrome.com/docs/crux/release-notes. Matt G. Southern, Search Engine Journal, 13 October 2026, reporting Annie Sullivan's Chrome post, https://www.searchenginejournal.com/google-experimental-soft-navigation-metrics-crux/592820/