Core Web Vitals are a tiebreaker, not the prize

After more than a year of warning, the page experience update is now fully live. Google announced it back in 2020, set the gradual rollout to begin in mid-June 2021, said it would not reach its full weight until the end of August, and has now confirmed that the rollout completed in early September. So the thing the field spent a year preparing for has arrived, and the headline is simple: page experience is officially part of how Google ranks pages. The centre of that update is the set of metrics Google calls Core Web Vitals — Largest Contentful Paint, First Input Delay, and Cumulative Layout Shift — which together try to put numbers on how a page feels to load and use. They do not sit alone; page experience also folds in whether a page is mobile-friendly, served over HTTPS, and free of intrusive interstitials, the pop-ups and overlays that block content. But the Core Web Vitals are the new and demanding part, the piece that sent everyone to their dashboards. Two boundaries on the change are worth fixing in mind from the start, because they shape everything that follows. The first is that, for now, this is a mobile signal: the page experience factor that just rolled out applies to mobile search, not desktop, a desktop version being something Google has said will come later. The second is proportion, which is the heart of this note: the update makes page experience a ranking factor, but a deliberately modest one, and the months of anticipation produced an expectation of impact far larger than what Google promised or what the data delivered. Understanding what actually shipped — a real but small mobile signal, layered on top of everything that already decided rankings — is the difference between responding sensibly and overreacting to a change that was loud in the trade press and quiet in the results.

It is worth knowing what the three Core Web Vitals measure, because the names are opaque and the fixes follow directly from the meaning. Largest Contentful Paint is about loading: it marks the moment the largest piece of content visible without scrolling — typically a hero image or a big block of text — has finished rendering, and Google counts under 2.5 seconds as good. It is the metric that answers the visitor’s first impatient question, has the page loaded yet. First Input Delay is about interactivity: it measures the lag between a visitor’s first interaction, such as tapping a button, and the moment the browser is actually able to respond, with under 100 milliseconds counted as good. A poor score here is the feeling of tapping a page that has visibly loaded but sits frozen because the browser is still busy running scripts. Cumulative Layout Shift is about visual stability: it quantifies how much the content jumps around as things load in, and a good score is below 0.1. Everyone has felt a bad one — you go to tap a link and an image loads above it, shoving the link down so your finger lands on an ad instead. Two things about how these are judged matter as much as the definitions. They are assessed at the 75th percentile, meaning a good score requires the experience to be good for at least three-quarters of visits, not just on average. And they are measured on field data — anonymised readings from real Chrome users collected in the Chrome User Experience Report — rather than on a single lab test. That last point explains a common confusion: the number a speed tool shows you from one run on a fast connection can differ from the field score Google actually uses, because the field data reflects the full spread of real devices and networks your visitors bring, including the mid-range phones and patchy connections that a developer’s machine never simulates.

Here is the part the year of anticipation tended to bury: Core Web Vitals are a small ranking factor, and Google said so plainly before the rollout and the independent data confirmed it after. Studies that tracked the same set of ranking URLs from before the page experience rollout to after it completed found that the relationship between good scores and ranking position barely changed in shape — the pages that ranked well still ranked well, the correlation between performance and position stayed roughly where it had been, and the update did not reshuffle results in the way the alarm suggested it would. The consistent finding was that content relevance and quality remained the dominant factors, exactly as Google had been saying. The right way to hold this is as a tiebreaker. When two pages are genuinely close on the things that decide most of the result — how well they answer the query, how trustworthy and complete they are — page experience can be the thing that gives one of them the edge. That makes it most valuable precisely where competition is fiercest and pages are otherwise hard to separate: in a crowded niche where a dozen results target the same query with comparable content, being the fast and stable one can move you a few positions. But the same framing tells you its limit. A tiebreaker only operates when there is a tie to break; it does not lift a page that is losing on relevance, and no amount of performance tuning will carry thin or off-topic content past something that answers the question better. There is opportunity in the numbers — surveys at the time found only a small fraction of pages were passing all three metrics, which means clearing the bar genuinely distinguishes you from most of the web — but the opportunity is to stop being disadvantaged, not to buy an advantage you have not earned on substance. Read correctly, the update rewards sites that are already good and merely slow, and asks very little of the rankings beyond that.

The practical programme that follows from all this is short, which is the point. First, measure where it counts: open the Core Web Vitals report in Search Console, look at the mobile data specifically, and let it tell you which pages and which metrics are actually failing on real-user field data rather than guessing from a single speed test. Then fix the cheap, high-impact things first, because for most sites a handful of unglamorous changes clear the bar. To steady Cumulative Layout Shift, set explicit width and height attributes on images, embeds, and ad slots so the browser reserves their space and nothing slides as the page fills in. To improve Largest Contentful Paint, find the largest above-the-fold element — usually an image — compress and right-size it for the device, and make sure it is not lazy-loaded, since the one thing you want to load fast is the thing that defines the metric. To help First Input Delay, trim or defer the heavy JavaScript that ties up the main thread, so the browser is free to respond when a visitor first taps. Those three moves clear most failures. The harder discipline is knowing when to stop. Once a page reliably passes all three metrics on field data, you have captured essentially the entire page experience benefit, and the hours you would spend chasing a near-perfect score are hours your content and relevance would repay far better. The whole stance this update rewards is one of proportion: treat performance as a floor you clear and then leave, not a trophy you keep polishing, and put the freed time back into the substance that actually decides where you rank. Holding that proportion — clearing the threshold, then returning to what matters — is the measured way the AC Group has worked for 27 years.