October 21, 2026
You can judge a service-business site’s speed without learning to build websites. You need a plain definition of “fast enough,” and you need the record of what real visitors already experienced.
Why seconds matter
People leave when a page makes them wait. The BBC found it lost an extra 10% of users for every additional second its site took to load. Google cites that finding in Why does speed matter? On a local service site the loss is easy to miss. The person who was about to call goes back to the list and tries the next business. Nobody writes to tell you which second they wouldn’t spend.
Google’s three yardsticks in plain words
Google writes the yardsticks down, with the line for a good result, in Web Vitals. There are three. Each one is a question a customer is already asking, whether or not they know the name.
Is the main content on screen within 2.5 seconds? The name of that measure is Largest Contentful Paint, or LCP. It marks when the largest thing in view has actually appeared. The time is counted from the moment the page starts loading. On a service site, the large thing is often the first screen a customer is trying to read: what you do, where you work, and how to reach you. Until that is on screen, they don’t have a page yet. They have a wait.
Does the page respond to a tap within 200 milliseconds? The name is Interaction to Next Paint, or INP. It is the gap between a tap and the next thing the page paints to show it noticed. Tapping Call, opening a menu, or moving to the next box in a form are all this kind of moment. A page can look finished and still fail here, if it sits frozen after the tap.
Is the layout shift 0.1 or less? The name is Cumulative Layout Shift, or CLS. It scores how much the page jumps while it is still settling. Google’s line for good is 0.1 or less. A lower score means a steadier page. The jump a customer feels is the words, the photo, or the button moving after they have already started to read or to aim a thumb.
Google does not score you on your quickest visit, and it does not score a test run from a desk. It judges each measure at the 75th percentile of real visits. Put the visits in order from fastest to slowest. The visit at the 75th percentile is the one that has to clear the line. Mobile and desktop are judged separately. A pass on the office computer does not speak for the phones.
Check it yourself in five minutes
Open PageSpeed Insights and enter your address. If the site is already in Google Search Console, open its Core Web Vitals report. You are looking for LCP, INP, and CLS.
What you will see first is often a lab result: a simulated visit, set up for the test, on a controlled device and a controlled connection. Read past it. Find the real-visitor result. Google calls that field data. It is the record of phones and computers that actually opened the page. The Web Vitals page says a lab measurement is not a substitute for field measurement. Treat the lab result as a hint about what might be heavy. Treat the field result as the answer to whether you pass.
Sometimes the field result says there have not been enough real visits to judge the page. That is not a pass, and it is not a failure. It means the record is too thin to use. Keep the lab hints. Don’t decide the site is fine because the simulated run looked calm.
Usual culprits on local service sites
The same kinds of things slow local service sites.
Oversized photos. A picture placed on the page at the full size it had on the camera has to travel down the connection before the main content can appear. It can still look right on a phone after it is saved smaller. Until that happens, the customer is waiting on the file.
Heavy page builders, and plugins and widgets stacked on them. A builder often loads a large kit of its own. Widgets added on top — chat, reviews, popups, extra forms — each add more code for the browser to handle. That work happens before the hours, the services, and the phone number get their turn.
Banners that push content down while loading. A bar that shows up after the first words are already on screen — a promotion, an announcement, a notice — shoves those words and the buttons downward. The shove is layout shift. It is also how a thumb that was aimed at Call lands on something else.
What to ask whoever runs your site
You don’t have to make the change yourself. You do need a straight answer from whoever looks after the site. Ask:
- “Do we pass all three on mobile?”
- “What’s slowing the main content?”
All three means LCP, INP, and CLS. On mobile means the phone result, judged on its own. If field data exists, the answer should come from that, not from a lab run and not from how the site feels on the office computer.
The main content is what the customer came to see and to use. A useful answer names the measure that is failing, and it points at a photo, at the builder and the widgets stacked on it, or at a banner that arrives late. If the answer is only a promise to look at speed in general, open the field result on the same call and read the three measures from the page.
How the same site should behave in a hand, on a weak signal, is in Your website should work on a truck’s phone signal.