If your PageSpeed Insights score changed, compare the reports before paying for fixes or celebrating an improvement. Check the page address, mobile or desktop setting, test conditions and Lighthouse version. Then look for evidence that visitors experienced a change, too.
There is a timely reason to do this. Lighthouse 13.6 changed part of the testing process, and its arrival in PageSpeed Insights was confirmed on November 2, 2026. For a business owner reviewing an agency report or a repair quote, that creates a practical question: did the website change, did the measurement change, or did both?
What changed in PageSpeed Insights?
PageSpeed Insights uses Lighthouse to run a controlled test of a page. The official Lighthouse 13.6 release notes, dated October 27, 2026, describe improvements to its loading simulation. In plain English, the tool changed how it estimates the time a page would take to load under the conditions being tested.
Google performance advocate Barry Pollard confirmed on November 2 that the version was on PageSpeed Insights. He explained that two simulated loading measurements became more accurate and that sites relying heavily on JavaScript were more likely to be affected. That does not promise a higher score for every website.
The practical implication is that a score can move without proving the site itself became faster or slower. If an agency made changes during the same period, part of the difference could come from the test. Ask for supporting evidence before assigning all the credit, or blame, to the work.
Compare the same page under the same conditions
Start by asking for the original reports, including their dates. A cropped screenshot of a score leaves out details you need to judge the result. For example, a mobile test of your booking page and a desktop test of your homepage cannot show whether the booking page improved.
Use this short comparison checklist:
- Match the exact page address. Check that both reports tested the same destination, including any redirect to another page.
- Match mobile with mobile, or desktop with desktop. These are separate tests with different conditions.
- Check the tool, version and environment. Record the Lighthouse version shown in each report. A test inside someone’s Chrome browser may use a different setup from PageSpeed Insights.
- Repeat the lab test. Keep the settings consistent and review several results instead of selecting the best run. Google notes that test conditions and other factors can produce variation.
Now compare the timing of website changes with the timing of the reports. A new booking widget, a changed template or a hosting change belongs in that record. If neither the settings nor the versions match, treat the comparison as inconclusive until you have a better baseline.
This is where a focused SEO audit can help: establish what was measured and which pages deserve attention before building a repair list.
Read the real visitor data separately
The large performance score is based on a lab test: a controlled run used to investigate a page. PageSpeed Insights also shows field data, which comes from real Chrome visits collected through the Chrome User Experience Report. Google’s explanation of PageSpeed Insights makes this distinction clear.
The field section covers the previous 28 days. It is a rolling view, so it will not instantly show the full effect of a fix made this morning. Check the dates and whether the report describes the specific page or a broader collection of pages.
If Google has too few samples for one URL, it can show origin data instead, covering pages on that website origin. If even that has too little data, no field results may appear. A missing result does not establish that the page performs well.
Imagine a service business with a busy homepage and a rarely visited quote page. Website-wide field data may provide useful context, but it cannot prove the quote page’s new form loads well. Check that page directly and test the customer journey, including opening the form and submitting an inquiry.
Use the evidence to choose your next step
A changed score deserves a proportionate response. The useful question is whether there is a repeatable problem on a page that matters to the business. The decision guide below can help you turn a report into a sensible next action.
| What you see | Your next step |
|---|---|
| Different test versions or settings | Establish a comparable baseline before judging the work. The score difference alone cannot isolate a website change. |
| A repeatable problem on an important page | Request a focused diagnosis, especially if visitor data or customer reports support the concern. |
| A better score, with no visitor evidence yet | Check the customer journey and allow time for field data. Keep claims about rankings and leads separate. |
| No page-level field data | Check whether the report uses origin data. Test the page directly; missing data does not mean it passes. |
A practical decision aid, not a guarantee of a site’s condition. Compare saved reports and investigate page-specific evidence.
Keep the customer journey in view as you investigate. If people report that a booking form freezes, check it even when the lab score looks good. If the score improves but bookings stay flat, review traffic and tracking before claiming the improvement produced more customers.
For example, a business might add an appointment widget just before its mobile score drops. That timing gives you something specific to investigate. It does not yet prove the widget caused the problem or that removing it would be worth the loss of online booking.
For broader background, MDM’s mobile optimization guide covers general improvements. Here, the first job is to establish whether the before-and-after evidence is trustworthy.
When is a website speed audit worth paying for?
An audit is easier to justify when it answers a defined business question. That might be why a quote page feels slow on customers’ phones, or whether a recent redesign introduced a repeatable loading problem. Ask the provider to explain which page they will investigate and what evidence they will deliver.
Ask for a diagnosis you can use
A useful review should connect the symptom to a likely cause, identify the work needed and explain how the result will be checked. Ask for priorities that reflect the pages customers use, rather than a long list of warnings with no clear order.
If the evidence points to a loading bottleneck, technical SEO support can investigate it further. If a shared page template or interaction needs changing, web design and development may be part of the solution. The scope should follow the diagnosis.
Agree on how success will be measured
Use comparable lab tests to check the immediate change, then review real visitor data over an appropriate period as it becomes available. Include practical checks such as whether customers can complete the booking or inquiry process. Agree on these measures before the work begins.
Google describes Core Web Vitals as measures of real-world loading, responsiveness and visual stability. A Lighthouse score alone cannot demonstrate a ranking gain, more leads or higher revenue. Be cautious about a proposal that treats a perfect score as the whole outcome.
Get a clear answer before buying fixes
If your PageSpeed Insights score changed and the explanation still feels unclear, bring the reports and the page that matters most to your business. That gives the review a concrete starting point: what changed, whether customers are affected and what deserves attention first.
Motivated Digital Marketing can help scope that investigation through an SEO audit and connect confirmed issues with the right technical or development work. Talk with MDM about your website performance to decide what to check next.
Sources and review date
Reviewed November 9, 2026. The release date and the public confirmation that it was on PageSpeed Insights are separate events. Check the version in each report rather than assuming every tool uses the same one.