Experiment in setting Core Web Vitals benchmark for publishers

Google’s Core Web Vitals are seen as the baseline for search performance. But publishers using the user experience standards to judge how their websites are likely to perform have had to contend with changes to the definitions recently. If you are struggling to make sense of how you are doing in relation to other publishers, this experiment from the University of Missouri describes an attempt to lay out an interesting Core Web Vitals benchmark.

Takeaways

Successful SEO starts with user experience, according to Ryan Restivo, RJI Fellow at the University of Missouri. Google’s Core Web Vitals set the standard for how easy a page is to interact with; users are 24% less likely to abandon pages that meet Core Web Vitals thresholds.

The speed that a web page loads sits at the center of any performance considerations. Google’s own studies have found that users on the mobile web have an attention span of between four and eight seconds; they will lose patience even sooner on web pages that they are not familiar with.

In that context, Restivo highlights the importance of making sure your pages load quickly.

The worst thing to experience with a slow web connection is an article taking a long time to load… Every second that an article doesn’t load is a chance for the user to leave.

Performance experiment

As part of his work as an RJI Fellow, Restivo ran an experiment to test the performance of the homepages and article published by a group of independent US news providers. Using Google’s tools he tested more than 90,000 home pages and 70,000 article pages.

Restivo’s objective was to measure page performance to determine if the publishers – members of the LION publishers association – were meeting Core Web Vitals thresholds. He wanted to know if it was possible to build a baseline so that publishers could understand their performance against a common Core Web Vitals benchmark.

Core Web Vitals looks at a range of metrics to score performance out of 100 – across all tests just 0.06% of homepages returned a scored of 100 and 0.94% of article pages. Only six organizations out of received at least one 100 rating.

The results

Of the three key page speed performance metrics, Restivo focused on LCP (Largest Contentful Paint). He explained that the name might sound intimidating, but it is really just the time it takes for the page to load the largest block of text, image or video.

An LCP score of 2.5 seconds or less is considered good by Google; over 2.5 seconds needs improvement; and 4 seconds or more is labeled poor. Less than 5% of the homepages that Restivo tested returned an LCP score of 2.5 seconds or less. Article pages did better, with 14% scoring 2.5 seconds or less.

Restivo also looked at Google’s accessibility measures – site best practice, SEO and use of PWAs (progressive web apps) – in trying to build a picture of publisher performance. His results offer advice on testing pages and links to testing tools and resources that suggest how to improve site performance. He said: