Speed Testing Tools Don’t Reflect Google’s Ability to Pre-render & Cache AMP Pages
When displaying AMP pages, Google is able to pre-render and cache them directly from the search results, which saves the time typically spent retrieving the initial HTML to get the rendered version. However, this isn’t something that is taken into account in the testing tools.
Confirm PageSpeed Insights & Structured Data Testing Tool IP Addresses with Test Page
PageSpeed Insights and the Structured Data Testing Tool use different IP addresses than Googlebot. There isn’t a list of IP addresses that they currently use, so check this using a test page and analyse the requests coming from these tools.
Speed is Critical For Google to Index Content Quickly
In order to index content quickly, for example news articles, Google need to be able to crawl the pages quickly. This includes being able to access the server quickly, with pages also loading quickly.
PageSpeed Insights Lab Tests & Aggregate CrUX Report Data Are Used for Assessing Site Speed
Some of the sources that Google uses to assess site speed are PageSpeed Insights lab tests and aggregrate data from the Chrome User Experience Report.
Speed Metrics Important for UX Are Different to Metrics Important for Crawling & Indexing
While there is some overlap, speed metrics that are important for UX are completely different to those used for crawling and indexing. For the latter, Google needs to be able to request the HTML pages as quickly as possible and server response time is also important.
Use Lighthouse Metrics to Understand How Users Experience Site Rather Than Search Engine Algorithms
Google uses a variety of metrics to determine site speed, but a lot of the metrics in Lighthouse are more likely to impact how users judge speed rather than Google’s search algorithms. John recommends testing site speed with users rather than trying to determine if Google sees it as too slow.
Speed is a Minor Ranking Factor Compared to Content Quality or Relevance
Improving site speed alone probably won’t change rankings or visibility for a website because it is a very small factor compared to the relevance and quality of a page’s content with regards to rankings.
Lighthouse Analysis Uses Medium-end Smartphone on 3G Connection
Lighthouse analysis uses an average smartphone on a 3G connection to measure speed. This might account for discrepancies between Lighthouse and independent tests which might be carried out over a much faster Wi-Fi connection.
Server Location Shouldn’t Impact Google’s Page Speed Measurement
The geographic location of a web server shouldn’t impact Google’s measurement of page speed. They have labs data to artificially measure the speed a page would take, and field data to measure the speed of pages for real users.