Countries with Slow Internet Can Affect Core Web Vitals Scores

In a Google Office Hours Hangout Google’s John Mueller answered  if the Core Web Vitals scoring would give consideration to sites whose users come from countries with predominantly slow Internet.

Core Web Vitals Lab & Field Testing

Testing core web vitals with Lighthouse gives what’s called Lab Data. Lab data is an estimated score using a simulated environment.

Field data, which is what’s reported in PageSpeed Insights, is based on real-world data. The real world data is gathered from actual site visitors who have opted into anonymously providing the site download speed information.

• Lab data = Simulated
• Field data = Actual scores from real people visiting the website

The person asking the question noted that they had updated their site and achieved a score of 100 in the lab testing. But that when the field data was updated they saw that their field scores (based on actual site visitors) as reported in Google Search Console actually went down and were significantly lower than the simulated scores predicted through Lighthouse.

The person asking the question confirmed that he knows the difference between lab and field data and that field data takes time to update.

The heart of the question is, why does the simulated lab data reflect the changes made to improve the core web vitals score but that Google Search Console shows the metrics getting worse.

Mueller began his response by reminding the person asking the question that there is a reporting delay with field data.

The person asking the question reiterated that they understood this and that it had been taken into account, that they had waited over 28 days to see the updated scores and the scores were worse.

Mueller next responded that it’s hard to say about this specific case without looking at the details and this is the point where the discussion became very interesting.

Mueller:

“One thing that I would try to do there is to try to figure out which part of Core Web Vitals is affected by that, if it’s like Largest Contentful Pain or if it’s CLS.

And based on that, try to figure out where it might be coming from.”

Next Mueller lowered expectations on the accuracy of the lab data.

Mueller continued:

“One of the things that generally happens with the lab versus field data is that with the lab data it’s basically an assumption.

It’s an approximation of what our systems think might happen in the field. Because there are just so many unknowns out there that depends a little bit on your users, where they’re coming from, what kind of devices they have… all of that.

Which means that you can use the lab data to incrementally improve but you don’t necessarily see a clear connection between the lab results and the field results. I don’t know if that’s something that might be playing a role there…”

The follow up question inquired if there were considerations for slower Internet speeds in different countries. He wanted to know if there were different scores for countries with slow mobile connections.

He next pointed out that this might put his site at a disadvantage with countries that have a preponderance of users in developed countries with faster Internet connection.

The background to that question is that Core Web Vitals metrics are measured from actual users. If a user is on a low quality cell phone on a low quality data connection then the core web vitals scores will be lower than someone from another country with a faster Internet connection and a higher quality mobile phone.

Mueller responded:

“I don’t know what the final setup there will be.

It is something where we have country information in Chrome User Experience Report data. So it is something where we’d be able to figure out where users are primarily coming from.

But the general idea is still kind of that users should be able to have a good experience.

And if the bulk of your users sees a slow experience, regardless of why, then essentially that’s what will apply there.

So that’s at least from what I know, that’s kind of the general standpoint there.

It’s like if 90% of your users are coming from locations that are slow and essentially 90% of your users have this …sub-optimal experience with your site, then that’s kind of what will be taken into account.”

Block Slow Countries for Better Core Web Vitals Scores?

The idea has taken hold that blocking certain countries might help the Core Web Vitals scores. Someone asked Martin Splitt about in January 2021.

This is how Martin responded about blocking countries for core web vitals:

No. That’s thinking that is laser focused on the Core Web Vitals and that’s really, really risky.”

A, Because people from these countries, if they want to access your website, they will through a proxy or what’s called a “VPN” which really is mostly a proxy for most cases.

And then the speed is even slower, so not helping.

Users From Slow Countries and Core Web Vitals

As John Mueller understands it, he said that there will be no allowances made for sites whose users come from countries where the Internet connection is slower.

The implication and reason for the question is that this seems to put a website at a disadvantage compared to websites with users predominantly from countries with fast Internet.

But in a way that presumption kind of needs a deeper think.

The companies with mostly fast Internet users are not necessarily competing with websites whose user base is predominantly in slower Internet countries, since their user base is from the fast countries.

Expanding the user base to countries with fast Internet is a matter of country code domains (.uk, .in, .jp, etc.), language and other factors like overt hints about International Targeting (and “un-targeting”) in the Google Search Console.

The Core Web Vitals ranking influence is a very small thing and not going to have as much influence as these other more important signals that help Google understand the geographical and language relevance of a website.

Citation

Watch Mueller answer the question at approximately the 2 minute mark in the video: