Five site fixes before the page experience update

Google confirmed last month that the page experience update starts rolling out in mid-June and will not be fully live until the end of August. That is a slow, gradual rollout by design, and Google has been unusually clear that it is a tiebreaker rather than a wrecking ball: relevant content still wins over a faster page that answers the question worse.

So we are not rebuilding anyone’s site in the next three weeks. We are doing five things, in this order, on every local business site we look after. Here they are, plus the ones we are ignoring on purpose.

What the update actually measures

Page experience bundles a few signals that already existed with three newer ones, the Core Web Vitals:

• Largest Contentful Paint, how long until the main thing on screen appears. Good is under 2.5 seconds.
• First Input Delay, how long the page takes to respond the first time someone taps. Good is under 100 milliseconds.
• Cumulative Layout Shift, how much the page jumps around while it loads. Good is under 0.1.
• Plus the ones you should already have: HTTPS, a mobile friendly page, and no intrusive interstitials.

Two details matter for how you check yourself. Google uses field data, meaning what real visitors on real phones experienced over a rolling 28 day window, not the score a testing tool gives you on your laptop. And the report that matters is the Core Web Vitals report in Search Console, because it groups your URLs and tells you which template is the problem rather than which page.

The five fixes

1. Fix the hero, then stop. On local business sites the largest element is almost always the top image or the background video behind the headline. Compress it properly, serve it at the size it is actually displayed, preload it, and delete the rotating slider if there is one. Nobody reads slide three. This single change moves Largest Contentful Paint more than everything else combined.

2. Reserve space for anything that loads late. Layout shift on local sites comes from four culprits: images without width and height set, embedded booking widgets and forms, sticky call bars and offer banners that appear a second in, and web fonts swapping. Give every one of them a fixed box before it arrives. A booking iframe that pushes the page down as someone reaches for the button is costing you calls regardless of what Google thinks of it.

3. Audit the third party scripts. Every site we inherit has a chat widget, a review widget, a heatmap tool somebody installed in 2019, two analytics tags and a tag manager container with things in it nobody can name. Each one competes for the same main thread that has to respond to the first tap. Remove what is unused, defer what is not needed immediately, and load the chat widget after interaction rather than on page load.

4. Kill the popup that fires on arrival. Intrusive interstitials are part of this update, and the on-load offer modal is the most common one we find on local sites. Fire it on exit intent or after real scrolling instead. You will lose almost no signups and you will stop paying a penalty for the privilege.

5. Check the mobile page, on a mobile phone, on cellular. Not the emulator. A mid range phone on a normal connection, standing outside the building. Every site is fast on the office wifi. This is also the quickest way to find the things the reports do not catch: a tap target nobody can hit, a form field that zooms, a phone number that is not a link.

What we are skipping

We are not chasing a green score in a lab tool for its own sake. A site can score in the nineties on a desktop test and still fail on field data, and a site can sit at 70 and pass everything that ships in June.

We are not moving anyone off their platform because of this. Migration risk is a real cost and the ranking benefit is, by Google’s own description, a small one.

We are not rewriting content because of page experience. If a page is losing to a competitor with worse speed, the problem is the page, not the milliseconds.

Measure it so you can actually tell

Here is where most owners get stuck in August, when someone asks whether the update helped or hurt. Get three things in place before mid-June so you have a before.

Take a screenshot of the Core Web Vitals report in Search Console today, with the counts of good, needs improvement and poor URLs. Do the same with the mobile usability report. Those numbers move slowly, so today’s version is your baseline.

In analytics, build a view of organic sessions by landing page for the last 90 days, and save it. If you tagged a Google Analytics 4 property alongside your existing one last year, keep both running through the rollout. A property with three months of history is more useful in August than a perfect setup started in July.

Most importantly, keep counting calls and booked appointments, not just sessions. A gradual rollout across ten weeks means you will never see a clean before and after in a traffic chart. What you can see is whether the phone rang less.

What is coming next week

Google I/O ran last week and most of the search talk was about MUM, a model Google says understands questions across formats and languages. Nothing to do there yet. The one that matters more for anyone running ads is the Marketing Livestream on Thursday, which is where Google usually announces the ads and measurement changes for the year. We expect more automation and more of the measurement rework that the tracking changes on iPhones have forced.

Either way, the preparation is the same and it is worth doing this week: conversion tracking that fires once, on the right action, with clean names, and the site fast enough that a click has somewhere good to land.

What to do this month

• Open the Core Web Vitals report in Search Console and note which template, not which page, is failing.
• Compress, resize and preload the hero image, and delete the slider.
• Set width and height on images and reserve space for the booking widget, the form and any sticky bar.
• Remove or defer third party scripts you cannot justify, and stop the popup firing on load.
• Load your own site on a real phone on cellular and book an appointment through it.
• Save a baseline of Search Console and analytics before mid-June.

None of this is glamorous and all of it pays whether or not the ranking effect shows up. Fast, stable pages book more appointments, which is the only reason we do website work at all, and it sits underneath everything else in local search. If you would rather hand the list to someone, that is how we work with owners.