Lighthouse scores do not affect Google Search / SEO

Lighthouse scores do not affect Google Search / SEO

So there's that, it has been said. Your Lighthouse score doesn't matter. Let's nuance this a bit.

However, there's nothing to nuance to John Mueller's tweet. For those who don't know: John Mueller is Google's Search Advocate (Google Webmaster Trends Analyst is one of his other titles).

Those who have done a PageSpeed Insights test before might have seen both Lab data and Field data. But then there is Core Web Vitals and the overall pagespeed score as well.

The above is an example of a PageSpeed Insights report,including the overall performance score, Field data, Core Web Vitals status and Lab data. Let's nuance this a bit.

Pagespeed Insights is running Lighthouse

In case you didn't know, when you're doing a PageSpeed Insights test, it's actually just a Lighthouse test that is being run. Wonder what the difference is when you would run Lighthouse yourself within your Chrome DevTools? The circumstances.

PageSpeed Insights will test the given URL on a mid tier device (Motorola Moto G4) on a mobile network (3G).

Lighthouse score and lab data

Such test will then result in numbers per metric, and often quite some recommendations. Important to know is that your overall score is purely based on the individual lab data metrics. There is no direct relation between the recommendations and your overall performance score.

Field data and Core Web Vitals

That's where field data comes in, which can be recognized by the coloured charts. But only when the tested webpage or your domain has a sufficient amount of visitors.

Although this isn't something that Google is publicly telling us, I found that you'll need at least 400 pageviews via Google Chrome, per device type to be able to get any field data at all.

Real user experiences

The most important part here is that Core Web Vitals data is Field data. And while Field data might already reveal it: Field data is based on real user experiences, as tracked by Chrome User Experience report (CrUX for short).

Pagespeed impact on SEO and ranking

This brings us to the answer of the question if pagespeed will impact SEO: yes it will, but only via real user experiences. So Field data is the one you would want to look at, and specifically the "this page passes / does not pass the Core Web Vitals assessment".

See if your domain's overall data is passing Core Web Vitals

And lab data is not based on real user experiences. It is based on a simulated laboratorium test (hence lab data).

Optimize for real users

This makes sense though when looking at copywriting. In the end, we should write texts for real users. Just like we should optimize pagespeed and performance for real users as well.

This does mean that when you're audience is usually browsing under more difficult circumstances (older devices, warm weather, bad connectivity), you would have a harder job to optimize for your audience. This applies to copywriting as well: you would want to adjust your texts to the language level of biggest part of your audience.

Why does Google only say this now?

News has reached Google that the Lighthouse or PageSpeed Insights score is being faked or gamed more often. I Shopify performance engineer discovered this and his tweet got picked up by John Mueller. Gaming the pagespeed score is equal to deceiving yourself, or even worse: your stakeholders who will eventually find out anyway.