How quick should your site be? Fast enough so that someone using a phone can see what you have to offer and then take the next step without any delay. While "under three seconds" is a good starting point for investigation, Google has not established a universal three-second limit for every page. You should measure the speed of the pages that people actually visit and then test the call or quote process yourself.
Which speed numbers matter?
The three aspects of a user's experience are described by Core Web Vitals; a good result is one where at least 75% of the measured visits meet the threshold.
- For Largest Contentful Paint (LCP) the figure should be two point five seconds or less, which is the time at which the biggest visible element of a page, typically the headline or the main image, appears.
- The Interaction to Next Paint (INP) should be 200 milliseconds or less; this reflects the speed at which the page responds following a tap or click.
- The cumulative layout shift (CLS) should be 0.1 or lower; this metric relates to unexpected movements such as a button moving just as a user is attempting to tap it.
These are field thresholds, not a timer which waits for each and every image, font, and script to have loaded. Although Google includes Core Web Vitals in its ranking systems, having a good score does not ensure a top result or that the page will be indexed. The page must still deal with the searcher's question.
How to read a mobile speed test
Go to PageSpeed Insights and enter the URL of a page that a customer is likely to visit—for example, a service page or one from the link in your Google Business Profile. Check the mobile option. If available, the top section displays real-user data gathered over the past 28 days. Check whether this data refers specifically to the URL in question or to the entire site since the two may tell different stories. If the site is small, it might not have any field data, which simply means that there were not enough eligible visits for that report, not because the page has failed.
The lab section involves a controlled diagnostic test; it should be used to identify possible causes, after which you should verify them on your phone. The single performance score displayed in it may vary from one test to another and does not represent the number of lost leads. Should you modify an image or script today, the lab test will immediately show this difference, whereas the rolling field data will take time to show actual visits. As Google's PageSpeed Insights explanation points out, it makes the distinction between these two views.
Match the symptom to a fix
When the headline or the main photograph is delivered late, inspect the size of the main image, the server response, and the resources that are causing the first screen to be delayed. If the taps appear stuck, check heavy scripts, chat tools, and booking widgets. If the page jumps, leave some space for the images and embeds before they load and check how the fonts look. Google's guidelines for LCP, INP, and CLS give a detailed explanation of those causes.
A narrower example was shown on our own homepage in September 2026. Following the resizing of seven of the images below the fold for mobile and the serving of responsive WebP files, two repeated local Lighthouse mobile tests resulted in a reduction in total transfer from approximately 979 KB to 381 KB. The local performance score remained at 97. This represents an actual decrease in the amount of data transferred on our test page, not proof that conversions or rankings had increased, and it does not indicate that those below-the-fold images were the page's LCP element. That is the reason why we examine what has changed before declaring a speed task to be complete.
Is it possible for a visitor to go from the page to a call or a quote?
Do it yourself on your own phone with mobile data by going to the page that a potential customer would see, finding the service area, tapping the phone number and filling in the quote form. Can you carry out these steps without having to wait for a pop-up, without losing your place when the page layout changes, or without realising that the form never acknowledges receipt? Carry out the same task on the actual service page, not just the home page, and note the one step which causes you difficulty then deal with that first.
You can't overcome a hidden contact button or a broken form by using a fast metric, and a slower metric too requires context—if the field data applies to the entire site, then you should test the problem page on its own. If you would like a second opinion on that customer journey, ask for a free Growth Audit and let us know which page is bringing in your inquiries.
The 5-Point Website Checklist
Hold your own site up against the five things every local homepage needs, speed included. Enter your email and we will send the PDF, no strings.