Google began rolling out the page experience update last week. The rollout is gradual, and Google says it will not be complete until the end of August.
Since the announcement a year ago, the SEO community has shown steady concern about three acronyms: LCP, FID and CLS.
Our assessment is more measured. For most small sites, we do not anticipate any significant ranking change. This update is nevertheless one of Google's most useful initiatives for small business websites in a long time, but not for the reason causing concern.
What does the page experience update change?
Google has used some page experience signals for years: is the site mobile-friendly, does it use HTTPS, does it overload the page with pop-ups? This update combines those signals and adds Core Web Vitals, three measurements of how a page behaves for a real person loading it.
- Largest Contentful Paint (LCP): the time it takes for the main element of the page to appear. Google's "good" threshold is 2.5 seconds.
- First Input Delay (FID): the time the page takes to respond to the first tap or click. The "good" threshold is under 100 milliseconds.
- Cumulative Layout Shift (CLS): the extent to which elements move during loading. The "good" threshold is 0.1 or less.
Google's guidance is clear: great content still outweighs a fast page with weak content. Page experience serves mainly as a tiebreaker between pages that are otherwise comparable. Its documentation on page experience in Google Search explains this in a far more measured way than the many blog posts on the subject.
One side effect matters to news publishers: pages no longer need to be AMP to appear in the Top Stories carousel. For a small business, this change has little impact, but it indicates Google's direction: it prefers to measure how a page behaves rather than impose a format.
Which data should you check first?
The real data in Search Console. This point is often misunderstood.
There are two types of speed measurements. Lab data comes from tools such as PageSpeed Insights or Lighthouse: one simulated visit, on a simulated phone, over a simulated connection. Field data comes from real Chrome users who visited your site over the last 28 days. Google bases its ranking on field data.
As a result, some business owners worry about a PageSpeed score of 42 shown in red, while the Core Web Vitals report in Search Console indicates that most real visits are fine. The reverse also happens: a lab score of 90, while real users on older phones in rural areas have a poor experience.
We recommend starting with the Search Console report. It sorts your URLs into three categories (good, needs improvement and poor) based on actual visits. If your site is small and receives little traffic, there may not be enough data to display a result. This is normal. In that case, lab tools are the only source available, and they should serve as a guide rather than a verdict.
How do you fix Core Web Vitals on a small WordPress site?
Most of the small business sites we have worked on over the years run on WordPress, and they tend to fail for the same handful of reasons. There is no miracle plugin, but there is a list of common causes.
Layout shift: the simplest fix
CLS is the measurement visitors perceive most directly. A person is about to tap "Book now," an image loads above it, the button moves, and they tap an ad or another link. This situation is very frustrating and familiar to most users.
The fixes are generally simple. Give images and videos a width and height so the browser reserves the necessary space. Do not insert a banner at the top of the page after it loads. Load web fonts in a way that prevents text from reflowing during loading. Cookie banners that push content down are a frequent cause; display them as an overlay at the bottom of the screen instead.
LCP often depends on a single large photo
On a tourism or restaurant site, the largest element is almost always the hero photo. Very often, it is the full-size file straight from the camera, which the browser must shrink on every visit. Resize it to its actual display size and compress it. Convert it to WebP if your setup allows; plugins exist for this purpose. Also avoid applying lazy loading to the only image visible on the first screen, which many optimization plugins do by default.
This measure alone moves many sites from "poor" to "good".
Hosting also plays a role. A cheap shared host that takes a second and a half just to begin responding makes the 2.5-second target very difficult to reach, regardless of the image optimizations made.
Slow response is related to scripts
FID is less often a problem on simple sites, but when it is, the cause is excessive JavaScript: page builders, a slider no one watches, three analytics tools, a chat widget, an Instagram feed. Each seems light, but together they tie up the processor of a mid-range phone, and the page does not respond when tapped.
The solution is mainly to remove elements. Take out anything that does not provide real value. Often, no one on the team even remembers adding some of these elements.
Does speed benefit visitors or rankings more?
Visitors, first. In our assessment, the ranking effect will be limited, as Google itself has indicated. It is unlikely that a site will move from page three to page one simply by fixing its CLS.
The effect on visitors, however, is real and immediate. A site that loads quickly and whose elements remain stable holds attention: visitors read the page, find the phone number and book. That has always been the goal. It is even more important for businesses in rural Quebec, where many visitors are on a phone with a weak signal, parked by a lake, trying to find out whether you are open today.
The other benefit of this update is that it gives site owners and developers free tools and shared metrics. It becomes much easier to agree on priorities, and to check a month later whether a fix had an effect, when everyone is looking at the same report.
Before hiring anyone for a technical SEO review, we therefore recommend consulting Search Console. The tool is free and will first tell you whether a problem exists.