Analysis | 14 October 2026
Google's latest CrUX update introduces experimental measurements for website journeys that continue without a full page reload. Existing Core Web Vitals scores remain unchanged. For South African businesses with app-like websites, the useful next step is to check how those journeys are measured before changing performance targets. Google's announcement
What the new measurements cover
Announced on 13 October, the October Chrome User Experience Report release adds experimental versions of Largest Contentful Paint (LCP), Interaction to Next Paint (INP) and Cumulative Layout Shift (CLS). They combine full page loads with soft navigations. The CrUX API can attribute results to the destination URL, subject to eligibility. Release details
A soft navigation happens when an application changes the content and URL using JavaScript, without loading a new document. Chrome's definition requires a user action, a visible URL change and a visible paint, meaning content is drawn on screen. The measurement APIs are enabled from Chrome 151, with support limited to Chromium-based browsers. Chrome's technical guidance
The experimental data is available through the CrUX API, History API, BigQuery and CrUX Vis. Google is targeting May 2027 for graduation, subject to feedback; Search Console will switch after graduation. Availability and timing
Why a score can change without a faster website
In everyday terms, LCP concerns loading performance, INP concerns responsiveness, and CLS measures visual stability. Each describes an aspect of the experience worth examining as customers move between pages. Core Web Vitals definitions
Under the existing method, LCP records the initial load, while INP and CLS can reflect interactions and shifts across the application's page lifecycle. The soft-navigation method starts measurements for each recognised view, including LCP for newly displayed content. Different boundaries change what is counted, even if the website's code stays the same. How the methodologies differ
Who can use the data
CrUX contains measurements from eligible real-world Chrome users. It does not represent every customer or browser. A business should treat it as one source of performance evidence alongside its own analytics and testing. CrUX overview
South African websites can qualify through the global dataset, which includes experiences from all countries. Pages must be publicly discoverable and sufficiently popular. Google does not disclose the minimum visitor threshold, so a smaller business may find that its website or individual pages have no results. Eligibility criteria
Check coverage whenever data is missing. A missing result cannot establish whether a website performs well. Ask your developer which important URLs are represented and which need another way of assessing performance.
For teams building reports, the CrUX API is free within a limit of 150 queries per minute per Google Cloud project. The History API shares that quota, and Google does not offer paid increases. Development and reporting work may still carry a supplier cost. API limits, shared History API quota
How to review your website
Start with one commercially important journey. For an online shop, that could be moving from a category to a product page. For a service business, it could be moving from a service description to an enquiry page.
Ask the developer to establish whether each transition uses a full reload or a soft navigation. Record the starting URL, destination and action taken. Do not assume that every section of the same website behaves identically.
Then compare the existing and experimental measurements for eligible pages over the same reporting period. Label each series clearly. Chrome's developer guidance recommends considering both methodologies so that teams can maintain historical and cross-browser comparisons. Keep the previous baseline available when introducing a new dashboard. Measurement guidance
For each score change, record whether the calculation, website code or audience changed. Ask your reporting provider to explain any break in the trend. Avoid joining the old and new series into a single line that makes a reporting change look like a successful optimisation.
If the two methods tell different stories, investigate the underlying journey before commissioning fixes. Ask which navigation contributed the result and whether the same issue can be reproduced. Keep the initial arrival visible in the review, because a customer coming directly to a page may have a different experience from someone already browsing.
Keep SEO and conversion reporting honest
Google recommends good Core Web Vitals as part of a good page experience for Search. That supports continued performance work, but it does not establish a ranking gain for an individual website. Google Search guidance
If a supplier presents a better experimental score, ask whether the website changed or only the measurement changed. A revised calculation is insufficient evidence that customers now experience faster pages or interactions. Likewise, performance scores alone cannot demonstrate additional enquiries or sales.
Track the relevant business outcome separately, such as completed enquiries from the journey being reviewed. Note concurrent changes to advertising, offers or page content before attributing movement to technical work.
This week's practical decision is modest: ask the person responsible for your website to identify its navigation behaviour and data coverage. Use that answer to choose one useful investigation, with a clear baseline and a result you can check.