Here's the short answer. If you're asking "do Core Web Vitals affect rankings," yes, Google uses them in its ranking systems, and it says so plainly in its own documentation. But if you're hoping that fixing your CWV scores will produce a predictable jump in rankings, the evidence doesn't back that up. The studies that looked for a strong relationship between passing scores and better positions found something modest and inconsistent, not a lever you can pull with confidence.
Call this one mixed. That Core Web Vitals work as a core web vitals ranking factor isn't in question, since Google says so directly. The size of its effect, and whether you can count on seeing it move your rankings, is a different question, and the research doesn't settle it. This matters because a lot of marketing leads are weighing whether to send engineering time at page speed work versus something else. And the honest answer is that it depends on what problem you're actually solving, not on a general rule either way.
What does Google actually say about Core Web Vitals?
Start with what Google has confirmed, because this is the part people get wrong in both directions.
That Google's ranking systems use Core Web Vitals at all isn't in dispute. Google's Search Central documentation, last updated in September 2026, states this directly, and that single sentence settles the basic question. If someone tells you CWV play no role at all in rankings, that's not accurate.
But the same documentation also says good scores don't guarantee top rankings, that chasing a perfect score purely for SEO reasons may not be the best use of your time, and that Google still wants to show the most relevant content even when the page experience isn't great. So the confirmed role and the practical weight of that role are two separate things, and conflating them is where a lot of advice online goes sideways.
You'll also hear CWV described as "just a tiebreaker." That's not quite right either. Back in 2021, Google's John Mueller told a Reddit thread that Core Web Vitals are "a ranking factor, and it's more than a tie-breaker, but it also doesn't replace relevance." That's a more useful way to think about it than either extreme. It's not nothing, and it's not the decider, and it definitely doesn't override whether your page actually answers the question someone searched.
Mueller wasn't done three years later. He added more context on LinkedIn, saying Google has "been pretty clear that Core Web Vitals are not giant factors in ranking," and that he doubted anyone would see a big ranking drop just because of them. Put those two comments together and you get a consistent picture: CWV matter, they're not confined to literal ties between otherwise-identical pages, and they're not large enough on their own to explain a big swing in your rankings.
One more thing worth knowing. Google's ranking systems generally evaluate pages individually, while also making some assessments at the site level. That doesn't mean every single URL gets its own isolated CWV score that feeds directly into a formula. Google hasn't published the exact mechanics here, so be careful about treating a domain-wide performance number as proof of how any one page is being ranked.
What do CWV correlation studies show?
Beyond what Google has said, a handful of independent studies have tried to measure whether passing Core Web Vitals actually lines up with better rankings. None of them is a clean experiment. Every CWV correlation study here is observational, meaning the researchers looked at pages that were already ranking and checked for patterns, rather than changing one site's performance and watching what happened. Here's what they found.
Start with Moz. It ran one of the more detailed looks at this, pulling results from the top 20 positions for 10,000 keywords across mobile and desktop. In August 2021, pages that passed all three CWV thresholds averaged a 0.39-position advantage over pages that failed at least one. That sounds like a real finding until you notice the comparison point: the same gap was 0.38 positions back in May, before most sites had time to react to the rollout. The advantage barely moved. That's the detail that keeps this from being a strong ranking-impact story. One thing did stand out, though: pages failing all three tests saw a 1.15-position decline, compared to about 0.2 positions for the wider sample, which suggests the worst performers may have felt something, even if the overall relationship stayed flat.
Then there's SISTRIX. It looked at it from a different angle, tracking visibility index scores for domains before and after the CWV rollout. Domains that met all the requirements grew their visibility by 3.7% on average, versus 2.7% for the full set of monitored domains, a one-point gap. SISTRIX itself was upfront that several other Google updates happened during the same window and that correlation here doesn't prove causation. Worth remembering too: a visibility index isn't the same unit as a ranking position, so you can't line these numbers up against Moz's directly.
Advanced Web Ranking took a broader approach. It analyzed close to 3 million pages from top-20 results and found the picture was uneven across the three metrics. Loading performance showed some association with better positions. The old interactivity metric showed basically no clear relationship. Visual stability showed only a slight trend with small differences between positions. That's a useful reminder that "Core Web Vitals" isn't one thing: treating all three metrics as equally predictive isn't supported by what this study actually found.
Then there's Ahrefs, and this one gets misquoted a lot. It looked at 5.2 million pages using Chrome UX Report data. Ahrefs didn't conclude there's no correlation between CWV and rankings. The researcher who ran it said he wanted to assess ranking impact but didn't think he could draw a reliable conclusion because of the timing and other concurrent changes. What the study actually showed was measurement coverage: roughly a third of sites and about a fifth of individual pages passed the thresholds at the time, and plenty of already-ranking pages simply didn't have enough traffic data to be measured at all.
Why can a ranking factor have a weak visible correlation?
If Google says it uses the signal, why do these studies keep finding such a small or inconsistent relationship? A few reasons, and they're worth sitting with before you draw a conclusion either way.
Relevance usually wins. Google's own documentation says it will still show the most relevant content even when the experience is sub-par. When you study thousands of search results at once, you're mixing together pages with wildly different levels of relevance, backlink profiles, and content quality. A small performance signal gets buried inside all that noise.
Correlation can show up without CWV actually driving it. Sites that invest in SEO tend to invest in a lot of things at once, better content, cleaner structure, and faster pages, all around the same time. A site with good CWV scores might be ranking well mostly because of everything else it's doing right, not because of the page speed itself. Both the Moz and Ahrefs researchers flagged exactly this.
A small result doesn't prove zero effect either. This cuts both ways. If a study finds a weak aggregate relationship, that doesn't mean the signal does nothing, especially in a sample where a thousand other ranking factors are also moving around. The confirmed mechanism is Google's own statement that it uses CWV. A quiet correlation study doesn't override that.
The available data skews toward popular pages. Chrome UX Report data requires enough real-user traffic to generate a reading. Lower-traffic and newer pages often don't show up in these studies at all, which means the sample is already tilted toward sites that are already doing fine.
And the studies aren't measuring the same thing. Moz looked at average ranking position. SISTRIX looked at a visibility index. Advanced Web Ranking looked at performance patterns across positions. None of these numbers can be stacked on top of each other to produce one combined "CWV effect size," because they're answering different questions with different units.
This isn't unique to Core Web Vitals, either. Studies of author bylines run into the same correlation problem, finding a weak signal once you account for everything else on the page. Dwell time research hits the same wall. So do arguments over keyword density, for exactly the same reason Moz and Ahrefs flagged here. A single signal rarely explains much on its own, in a page that's doing dozens of things at once.
What do the studies fail to prove?
It helps to be precise about what this body of research doesn't establish, because the gap between what's claimed online and what's actually been shown is where most of the bad advice comes from.
None of these studies proves a causal effect from fixing one specific page. They're all observational snapshots, not before-and-after experiments on a controlled set of pages. None of them gives you an expected number of positions you'd gain by improving your scores, and none quantifies how that supposed benefit might differ by industry, query type, or device. And none of them, including Ahrefs' large one, was designed or able to isolate CWV's contribution from everything else happening on the web at the same time.
A fair reading of any CWV correlation study from this period also has to account for timing. Most of this research used the older measure set, including a metric called First Input Delay that Google has since replaced with Interaction to Next Paint. Studies built on that older framework can't tell you much about how the current system behaves. Treat the 2021 and 2022 findings as historical evidence about an earlier rollout period, not a live read on today's ranking mechanics.
When should marketing teams prioritize CWV?
This is the practical question, and the honest answer is that it depends on the problem you're actually trying to solve, not on treating a core web vitals ranking factor as a guaranteed lever.
Some pages are just hard to use. If important pages are genuinely frustrating to load or interact with, and your Search Console field data backs that up, that's worth fixing regardless of what it does for rankings. You have a confirmed user problem and a confirmed, if unquantified, ranking consideration stacked on top of it.
If a shared template is creating a bad experience across a lot of commercially important pages at once, fixing the template addresses a broad problem with one change. The ranking return is still unknown, but the user-experience case alone often justifies it.
If your content is already competitive and relevant, and performance is the clear remaining weakness, it's worth evaluating. Google's own guidance says a good experience can help when a lot of equally useful results are available. That's not a guaranteed tiebreak, but it's not nothing either.
If your field performance is already solid and the project is really about squeezing out a slightly better lab score, that's usually a lower priority for SEO purposes specifically. Google has said directly that chasing a perfect score for SEO alone often isn't the best use of time.
Pages that don't satisfy the query are a different problem entirely. If they lack useful content in the first place, don't treat CWV work as a substitute for fixing that. Relevance comes first in Google's own stated priorities, and no amount of page speed work changes that.
Sudden ranking drops deserve a closer look before you assume CWV caused them. If your rankings dropped and CWV is your only suspected cause, investigate before committing a big project to it. Mueller's 2024 comment specifically cautions against assuming CWV alone explains a meaningful decline.
It also helps to separate the different cases you might be making. A user-experience case asks whether the site is materially frustrating for real visitors. A commercial case asks whether a credible test shows better completion or conversion. An SEO case asks whether there's evidence of improved visibility beyond everything else that changed at the same time. These are three different claims, and conflating them is how a page-speed sprint ends up justified by a number it was never going to produce.
What would change this verdict?
The honest position here is "we know Google uses it, we don't know how much it moves things for you specifically." Here's what would actually move that needle.
A recent, transparent study using the current measure set, not the retired First Input Delay metric, would tell you more than anything from 2021 or 2022. URL-level field data matched carefully to device type and actual ranking positions would help separate signal from noise. A defined before-and-after intervention, where you can see what else changed and what didn't, would get closer to a real causal read than any cross-sectional snapshot. And effect sizes with some sense of uncertainty attached, rather than just a big sample size, would tell you whether what's being measured is reliable or coincidental.
Until something like that exists, the responsible move is to treat Core Web Vitals as confirmed to matter and unconfirmed in exactly how much. Before you greenlight the next performance sprint, ask what problem it solves if rankings don't move at all. If there's a real answer, that work holds up on its own. If the only answer is "it should help SEO," you're betting on a number the research hasn't actually produced.
