Why a High PageSpeed Score Does Not Guarantee More Qualified Leads

A page can load in under a second and still lose every good lead that reaches it.

That sounds like a contradiction, because speed scores are usually presented as a stand-in for results. A 95 means the website is healthy. A 40 means it is costing you business. The score is real and worth improving. It is also a test of one thing, run in a lab, on one page.

On three of our published projects, the speed score and the enquiry problem turned out to be unrelated. One site had a form that silently dropped two fields. One was sending fake leads to its sales team. One had articles that crawlers could not see. No speed test showed any of it.

This article explains what the PageSpeed score measures, what it cannot see, and the order in which to check speed, conversion and lead quality.

Four Numbers That Get Mistaken for One

When someone says a website "performs well", they could mean any of four measurements. Each comes from a different place and answers a different question.

The lab score is the furthest from revenue and the easiest to get, which is why it is quoted most. The qualified lead rate is the closest and the hardest to get, which is why it is often missing from the report.

What the PageSpeed Score Actually Tests

PageSpeed Insights shows two kinds of data on one screen, and they are easy to confuse. Google's description of the tool explains that the performance score comes from Lighthouse, which loads the page once in a simulated environment. For the mobile test, that means an emulated mid-range phone on a mobile network. A score of 90 or more is rated good, 50 to 89 needs improvement, and below 50 is poor.

Three details about that score matter to anyone using it to judge a website.

1. It changes from one run to the next. Lighthouse's own notes on score variability list seven sources of variation and recommend running several tests and taking the median. They say the median of five runs is twice as stable as a single run. One test, on one afternoon, is a weak basis for a decision.
2. It depends on which test you read. On 12 October 2026, the FlowEazy homepage scored 99 on mobile and 77 on desktop in PageSpeed Insights. Same page, same day, two scores. "The score" is never one number.
3. It weights what a lab can measure. In the scoring Google documents for Lighthouse 10, Total Blocking Time is 30 percent of the score, Largest Contentful Paint 25 percent and Cumulative Layout Shift 25 percent, with First Contentful Paint and Speed Index making up the rest. Interaction to Next Paint, the Core Web Vital for responsiveness, is not in the lab score at all, because a lab test has no real person clicking.

The second kind of data describes people. Field data comes from the Chrome UX Report: real visits by Chrome users over the previous 28 days, reported at the 75th percentile. A page passes the Core Web Vitals assessment when all three metrics are good for at least 75 percent of visits. The thresholds Google publishes are 2.5 seconds or less for Largest Contentful Paint, 200 milliseconds or less for Interaction to Next Paint, and 0.1 or less for Cumulative Layout Shift. Our Core Web Vitals guide for business owners explains each one in plain words.

Google's explanation of why lab and field data differ says that where both exist, field data should come first, because it reflects what real users experienced. The PageSpeed Insights documentation makes the same point from the other side: a good lab score does not necessarily mean real users have a good experience.

There is a catch for B2B websites. A page appears in the Chrome UX Report only if it is public and popular enough, and Google does not publish the minimum. Many B2B pages, with a few hundred visits a month, show no field data. PageSpeed Insights then falls back to data for the whole site. If the site is also below the threshold, the lab score is the only number on the screen. That is how a simulated test of one page ends up being read as a verdict on a business.

Three Things the Score Could Not See

These examples come from our published case studies. In each one the speed work was real and worth doing. In each one, a separate fault was costing enquiries, and a perfect score would not have shown it.

A form that dropped half of every enquiry

FlowEazy sells Odoo ERP services. Its homepage scored 26 for mobile performance in the lab before our work and 96 or more after it. During the same project we found that the enquiry form was silently dropping the Service and Requirements fields. Sales was receiving leads with half the information missing. The page could have scored 100 and those leads would still have arrived unusable.

Leads that were not people

Sai Impression manufactures labels. Its mobile lab score moved from 55 to the 90s. Separately, bot leads with fake phone numbers were reaching the sales team, and staff were calling people who did not exist. The fix was a filter that holds suspect leads in quarantine, so a real enquiry is never deleted by mistake. Our guide on how to stop junk leads describes the method. No speed tool counts fake enquiries.

Content that was responsive and invisible

Cafy is a UK job platform built with a JavaScript framework. In real-user data from 16 October 2026, its desktop Interaction to Next Paint was 106 milliseconds, comfortably inside Google's threshold. At the same time, its blog articles stayed invisible until JavaScript ran. The lab did flag a symptom, an 8.3 second Largest Contentful Paint. It could not flag the consequence: a crawler that does not run JavaScript received a page with no article in it. The Cafy case study covers the fix, and our guide to AI crawler access problems shows how to test your own pages for the same fault.

What these examples do not show. We do not have before and after lead quality figures for these projects, so we are not claiming that speed made no difference to them. The claim is narrower. In each case, the fault that was costing enquiries sat outside anything a speed test measures.

What Speed Changes, and What It Does Not

Speed matters. The evidence that faster pages keep more visitors is reasonable. It is also narrower than the way it is usually quoted.

The study cited most often is Milliseconds Make Millions, prepared by Deloitte for Google in 2020. It followed 37 brand websites for four weeks in late 2019 and compared natural changes in mobile load time with what visitors did. For retail sites, a 0.1 second improvement went with 8.4 percent more conversions. That is the figure that gets repeated.

The lead generation chapter of the same report is less tidy, and more useful to a B2B reader. With a 0.1 second improvement, progression from the first step of a form to submitting it rose by 21.6 percent. The same chapter reports that conversion rates on lead generation sites went down as speed improved, by almost 2 percent on mobile. The report's own notes add that the results are correlations from 37 sites, that speed was measured with lab data, and that Deloitte did not audit the figures. A conversion in that study is a completed form. Nobody checked whether the people who completed it were buyers.

A cleaner test is the Vodafone case study that Google published in 2021. Vodafone split paid traffic evenly between the original version of one landing page and a faster one. The version with a 31 percent better Largest Contentful Paint produced 8 percent more sales and a 15 percent better lead to visit rate. A controlled test is stronger evidence than a correlation. It is still one consumer page, and "lead" there carries no measure of quality.

Put together, a fair reading is this:

• Speed removes a reason to leave. Visitors on a phone and a mobile connection feel it most.
• Speed does not add a reason to enquire. The offer, the proof and the form do that.
• Speed says nothing about who enquires. That depends on which searches and ads brought people to the page.
• For Google, it is one input among many. Google's guidance on page experience says Core Web Vitals are used by its ranking systems. It also says good results in those reports "doesn't guarantee that your pages will rank at the top of Google Search results", and that it tries to show the most relevant content even where page experience is weaker.

The Order to Check Things In

When enquiries are low or poor, speed is one of five things to check, and it is second on the list. Each check is quicker than the one after it, and a failure early on makes the later numbers meaningless. There is no point measuring conversion on a page that search engines cannot read.

The third check is where most of the gain usually sits, and our landing page optimization guide covers it. For the fifth, see how to measure qualified leads from organic search, which gives the stages, the fields and the formulas.

When Speed Work Is Worth Paying For

Speed work is engineering time, and it competes with everything else on the list. These are the cases where it earns its place first, and the cases where it should wait.

Google says much the same in its page experience guidance: there is more to a good page than Core Web Vitals, and chasing a perfect score only for search may not be the best use of your time.