Page speed: a ranking signal Google capped on purpose, twice
What Google actually turned on, in 2018
Google announced the first page speed ranking signal on its Search Central blog on 17 January 2018, with rollout beginning 9 July 2018. As TechCrunch reported the same day, the change applied to mobile search only and was built to affect "only pages that deliver the slowest experience to users." Google was equally clear about what the signal would not do: a page with the most relevant, high quality content could still rank highly even if it was not the fastest page on the results list. The company described the expected impact as a small percentage of queries, the same deliberately narrow framing it had used for HTTPS four years earlier.
Then Google folded it into something bigger, in 2021
The Page Experience update extended the same caution to a new set of metrics, Core Web Vitals, rolling out in full between May 2021 and March 2022. For years afterward, how much weight the signal actually carried depended on who at Google was asked. Chrome engineer Addy Osmani laid out the current framing in a retrospective published 2 October 2025: "a fast site won't outrank a more relevant site just because of speed," but "in cases of ties or close calls, having good Web Vitals could be the tiebreaker." Google's own John Mueller pushed back on that framing years earlier, telling a forum thread questioning whether Core Web Vitals mattered at all, as reported by Roger Montti in Search Engine Journal on 6 August 2021: "It is a ranking factor, and it's more than a tie-breaker, but it also doesn't replace relevance." Google's own page experience documentation was updated again in April 2023 to settle the ambiguity in writing rather than in forum replies: it now states Core Web Vitals "are used by our ranking systems," while, in the same update, "other page experience aspects don't directly help your website rank higher in search results." Two separate Google voices, four years apart, landed on the same conclusion from different directions: it moves rankings some, and it never overrides a page that matches the query better.
How many sites even pass, on current data
The 2025 Web Almanac, compiled by Himanshu Jariyal, Prathamesh Rasam, Humaira and Aaron T. Grogg for HTTP Archive and published 15 January 2026 from July 2025 Chrome UX Report data, found that 56% of desktop origins and 48% of mobile origins passed all three Core Web Vitals thresholds at once. Largest Contentful Paint remained the single biggest failure point on mobile, passing on only 62% of origins. Put plainly, more than half the mobile web is still failing the exact bar the 2021 update set, four years after it rolled out, which is hard to square with the idea that page speed has become a decisive ranking lever. A signal that over half the web fails and still manages to rank on page one for most queries is not a signal doing much heavy lifting.
The cost that was never about rankings at all
The number worth actually planning around predates Core Web Vitals entirely and has nothing to do with Google's algorithm. Daniel An, Google's global product lead for mobile, published the Google and SOASTA research on 28 February 2017: the average mobile landing page took about 22 seconds to fully load at the time, and 53% of mobile visitors abandoned a page that took longer than three seconds. A neural network model built on the same dataset, with 90% prediction accuracy, found bounce probability rose 113% as load time went from one second to seven. None of that's a ranking penalty. It's a visitor closing the tab before your title tag, your schema, or your ranking position gets a chance to do anything, and that makes it a bigger practical problem than the documented ranking effect sitting above it. But the two get discussed as if they were the same conversation. The same mismatch between a documented, modest signal and the inflated way it gets discussed shows up in our piece on whether on-page SEO still matters, where the leverage people assume sits in a title tag turns out to sit somewhere else entirely.
Where this actually sits next to the rest of technical SEO
Our own breakdown of Core Web Vitals and AI search goes further into the citation side of this question: the same 2025 Web Almanac pass rates above hold for Google rankings, but a 2026 study of 2 million AI citation observations found no measured effect from Core Web Vitals on which pages get cited by ChatGPT, Claude, Google AI or Gemini once content depth and domain were controlled for. Page speed is a real but bounded lever on one system and, on the only controlled evidence that exists, not a lever at all on the other. That same pattern, a real signal getting treated as a bigger deal than the data supports, is also what we found when we checked whether Reddit mentions move AI citations: a documented, measurable effect that gets discussed as if it were decisive rather than one input among several.
What we could not verify
Google has never published an updated numeric weighting for Core Web Vitals comparable to the "fewer than 1%" figure it gave HTTPS in 2014, so "tiebreaker" and "more than a tiebreaker" are the most specific public framings available, and they come from two different people at two different points rather than one settled statement. The 2017 Google and SOASTA abandonment figures predate Core Web Vitals and measure full page load time rather than the three current metrics, Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift, so treat the 53% bounce figure as a related but not directly comparable measurement to the 2025 pass rates above it. We also found no official Google statement describing whether the 2018 mobile-only speed update was formally retired or folded entirely into the 2021 Page Experience update, or continues to run as a separate check.
Sources
The 2018 mobile speed update details come from Sarah Perez, Google will make page speed a factor in mobile search ranking starting in July, TechCrunch, 17 January 2018. The Page Experience rollout and tiebreaker framing come from Addy Osmani, The History of Core Web Vitals, 2 October 2025. John Mueller's quote is via Roger Montti, Core Web Vitals Ranking Factor is More Than a Tie Breaker, Search Engine Journal, 6 August 2021. Core Web Vitals pass rates come from Himanshu Jariyal, Prathamesh Rasam, Humaira and Aaron T. Grogg, Performance, 2025 Web Almanac, HTTP Archive, published 15 January 2026 from July 2025 CrUX data. The 2017 mobile abandonment and bounce probability figures come from Daniel An, New Industry Benchmarks for Mobile Page Speed, AMP Blog, citing Google/SOASTA Research, published 28 February 2017.
Common questions
Does page speed still affect Google rankings?
Yes, but Google has capped the effect on purpose twice. The 2018 mobile speed update affected only a small share of queries and the slowest pages. The 2021 Page Experience update folded speed into Core Web Vitals, and Google's own documentation, updated in April 2023, calls it a ranking signal without claiming it overrides relevance.
Why did Google start caring about page speed at all?
Google announced the first page speed ranking signal for mobile search on 17 January 2018, rolling out from 9 July 2018. The goal was pushing the slowest pages on the web toward a baseline, not creating a lever fast sites could use to beat more relevant ones.
If the ranking effect is small, why does page speed still matter?
Because the bigger number was never about rankings. Google and SOASTA found in 2017 that 53% of mobile visitors abandon a page that takes longer than three seconds to load. Most of the real cost of a slow page is visitors leaving before any ranking signal gets a chance to apply.
Fix what actually costs you visitors
A Technical SEO Audit checks load time, Core Web Vitals, rendering and crawlability together, so you know which fixes move rankings and which ones are mostly about keeping visitors from leaving before they read a word.