SERPs.ioWe Fact-Checked the 10 Most-Quoted SEO Statistics
Walk into any SEO conference, open any agency pitch deck, skim any "Why SEO Matters" blog post, and you will run into the same ten statistics. They show up in LinkedIn carousels, client proposals, keynote slides, and onboarding decks. Many of them are wrong. A few were never really right. Almost all of them are older than the iPhone X.
We call these zombie stats. They keep shambling through the industry because nobody ever pulls up the original source and asks if it still holds. This article does exactly that. For each of the ten most-quoted SEO statistics, we traced the origin, evaluated the methodology, and checked whether the number is still defensible in 2026. Some hold up. Most do not.
The scorecard
Here is what we found, at a glance.
| Stat | Real source | Year | Still true? | 2026 replacement |
|---|---|---|---|---|
| 75% of users never scroll past page one | Untraceable, popularized by HubSpot | c. 2010 | Wrong framing | 99% of clicks happen on page one |
| 53% of mobile users abandon sites over 3s | Google/SOASTA via Think with Google | 2017 | Partially true | Use Core Web Vitals thresholds instead |
| Content with images gets 94% more views | Jeff Bullas infographic, citing dead Skyword data | 2012 | Unverifiable | Drop the stat entirely |
| B2B buyers are 57% through before sales | CEB (now Gartner) | 2011 | Obsolete | Gartner: buyers spend 17% of time with reps |
| First page results average 1,447 words | Backlinko | 2016, updated 2020 | Correlation only, not causal | Word count is not a ranking factor |
| Top 10 results have 3.8x more backlinks than #2-10 | Backlinko/Ahrefs | 2020 | Directionally true | Still holds, caveat for AI Overviews |
| 60% of Google searches end without a click | SparkToro + Datos | 2024 | Accurate | Use this one. It is well-sourced. |
| 90.63% of pages get no Google traffic | Ahrefs | 2020 | Outdated low | 96.55% (Ahrefs, 2023) |
| 8 in 10 marketers say SEO is effective | HubSpot (varies by report) | 2022-2024 | Rephrased annually | Cite the specific year or skip it |
| Position 1 gets 27-39.8% CTR | Backlinko, Advanced Web Ranking | 2019-2023 | Broken by AIOs | Position 1 CTR dropped 58% with AIOs |
Now let us audit each one.
1. "75% of users never scroll past the first page of search results"
The origin
This is the undead king of SEO statistics. You will find it cited to HubSpot, imForza, Junto, Search Engine Journal, and at least a dozen marketing blogs that picked it up from each other.
The earliest traceable appearance is a 2010 HubSpot slideshare deck called "120 Awesome Marketing Stats, Charts and Graphs," which attributed the number to marketshare.hitslink.com (HubSpot on SlideShare). That underlying source is a dead link. The imForza blog post "8 SEO Stats That Are Hard to Ignore" also lists the number without citing any study at all (imForza).
There is no controlled study behind this number. It is marketing folklore.
Is it still true?
The underlying idea, that almost nobody clicks through to page two, is correct. Chitika's 2013 clickstream study found that 91.5% of traffic went to page one results, with just 4.8% to page two (Chitika Insights). Backlinko's more recent analysis of 4 million searches showed that only 0.63% of searchers click anything on page two (Backlinko).
So the behavior is real. The specific "75%" number is just something somebody typed into a deck in 2010.
2026 replacement
Say "only 0.63% of searches result in a click on page two of Google" and cite Backlinko. It is a more dramatic number, and it is sourced.
The famous "75% never scroll past page one" stat has no study behind it. It was typed into a 2010 HubSpot SlideShare citing a now-dead link. The real number is sharper: only 0.63% of searches result in a click on page two.
2. "53% of mobile users abandon sites that take over 3 seconds to load"
The origin
This one is real and traceable. It comes from a 2017 Think with Google article called "Find Out How You Stack Up to New Industry Benchmarks for Mobile Page Speed" (Think with Google). The data was collected by SOASTA Research (later acquired by Akamai) in 2016 and published in early 2017.
The study pulled from 11,800 global mobile web homepage domains tested on a simulated 3G connection.
Is it still true?
Partially. The 53% figure was specific to one dataset on one connection profile almost a decade ago. Since then, Google has moved on to Core Web Vitals (LCP, CLS, INP) as the industry-standard speed benchmarks (web.dev). The 3-second heuristic is reasonable but oversimplified. A 3-second LCP on a complex e-commerce page is far more forgivable than a 3-second LCP on a blog post.
More importantly, most sites are not on 3G anymore. The threshold users actually tolerate varies by device, connection, content type, and perceived value.
2026 replacement
Use Core Web Vitals language. "Pages with a Largest Contentful Paint over 2.5 seconds fail Google's good threshold" is more precise and still cites Google.
3. "Content with images gets 94% more views than without"
The origin
This one is an SEO urban legend. It traces to a Jeff Bullas infographic from 2012 or 2013 that cites "research by Skyword" (Jeff Bullas). Skyword is a real company. The actual underlying research post on their site is no longer live, and the Internet Archive captures do not include the methodology or sample size.
We could not find a primary source. Skyword's 2011 analysis claimed that articles with relevant images received more views than articles without them, categorized by topic, but the "94%" figure does not appear cleanly in any surviving primary document.
Is it still true?
Impossible to say, because nobody can reproduce or audit the original study. Even if the 94% was accurate in 2011, it was based on what looks like a self-reported customer dataset of a single content platform. Generalizing that to all content in 2026 is nonsense. Modern content discovery is dominated by AI Overviews, social algorithms, and video, none of which behave the way blog content did in 2011.
2026 replacement
Drop it. If you need to argue for images, cite modern eye-tracking or scroll-depth studies from Nielsen Norman Group instead (NN/g).
4. "B2B buyers are 57% of the way through the decision before contacting sales"
The origin
This came from a 2011 CEB (Corporate Executive Board) study, conducted in partnership with Google, based on interviews with 1,500 B2B customers. CEB was acquired by Gartner in 2017, and the study has been cited in Gartner and LinkedIn sales content ever since (LinkedIn Business).
Is it still true?
No. This stat is 15 years old, and buyer behavior has changed dramatically in both directions. Gartner's more recent research finds that B2B buyers spend only 17% of their total buying time meeting with potential suppliers at all, across the entire journey, not just the pre-sales phase (Gartner). Forrester has quoted figures closer to 70% and 80% for the self-directed portion of the journey.
The "57%" number has also been challenged by researchers like Bob Apollo, who points out that the CEB study measured buyer self-reports of information-gathering, not actual decision progress (Inflexion-Point).
2026 replacement
Cite Gartner's 17% figure for time-with-sales, or Gartner's finding that 75% of B2B buyers prefer a rep-free buying experience for first purchases (Gartner).
5. "The average first-page result on Google is 1,447-1,890 words"
The origin
The 1,890-word figure came from serpIQ's 2012 analysis, which is long dead. The 1,447-word figure comes from Backlinko's 2020 update of their "We Analyzed 11.8 Million Google Search Results" study (Backlinko).
Is it still true?
This is the stat SEOs most love to misquote. Backlinko themselves, in the same study, said: "despite the fact that long-form content tends to be best for link building, we found no direct relationship between word count and rankings." The 1,447-word figure is a descriptive average of what already ranks on page one. It is not evidence that writing longer content makes you rank.
Google has been explicit about this for years. John Mueller has repeatedly stated that word count is not a ranking factor (Search Engine Journal). Longer content tends to rank because thorough coverage is a signal of topical depth, not because Google counts your words.
Post-AI Overviews, the correlation has weakened further. AIOs frequently pull content from short, focused answers that are well-structured, not 2,000-word essays.
2026 replacement
"Write the shortest piece that fully answers the query, with the depth your topic requires." For more on this, see our content structure guide for AI citations.
6. "The top 10 results have X backlinks on average"
The origin
The most-cited version of this stat comes from Backlinko's analysis of 11.8 million search results, which found that the #1 result has 3.8x more backlinks than positions #2-10 (Backlinko). Similar findings have appeared in Ahrefs' ranking correlation studies over the years.
Is it still true?
Directionally, yes. Backlinks remain a ranking factor, and pages that rank for competitive queries tend to have more authority signals than pages that do not. Ahrefs' more recent data confirms the relationship still holds (Ahrefs).
The caveats in 2026:
- AI Overviews are now pulling citations from pages that sometimes have very few backlinks, because LLMs value structure and freshness differently than ranking algorithms do.
- For AI search specifically, brand mentions are starting to matter more than backlinks in some contexts (see our brand mentions vs backlinks analysis).
- Correlation is still not causation. Positions and backlinks are both downstream of the same thing: quality and time in market.
2026 replacement
Use Backlinko's 3.8x figure if you need a headline number. Just don't pretend it means "build 100 links and you rank #1."
7. "60% of Google searches end without a click"
The origin
This one is well-sourced. Rand Fishkin at SparkToro, working with Datos (a Semrush company) clickstream data, has been publishing zero-click search studies since 2019, with a major 2024 update (SparkToro).
The exact finding: 58.5% of US Google searches and 59.7% of EU searches end in zero clicks, based on a statistically significant sample from September 2022 through May 2024.
Is it still true?
Yes, and it is probably understated. The SparkToro study was largely pre-AI Overview rollout. Other 2025 data suggests zero-click rates are higher now. Semrush found that Google AI Mode produces a 93% zero-click rate, nearly three times traditional search (Semrush).
For a full breakdown, see our zero-click searches deep dive.
2026 replacement
Keep using this one. It is the rare SEO stat that has aged well and been updated responsibly. Just cite the 2024 SparkToro version, not a third-hand blog repeating the 2019 number.
8. "90.63% of pages get no Google traffic"
The origin
Tim Soulo and the Ahrefs team published this in 2017 and updated it in 2020, based on a crawl of roughly 1 billion pages. The original claim was that 90.63% of pages got zero organic traffic from Google (Ahrefs, 2020 archive).
Is it still true?
No. It is actually worse. Ahrefs updated the study in 2023 using an expanded dataset of 14 billion pages. The figure jumped to 96.55% (Ahrefs). Only 3.45% of pages get any meaningful Google traffic.
With AI Overviews now suppressing clicks to non-cited pages, the practical number in 2026 is almost certainly higher still, though no one has published a clean 2026 update yet.
2026 replacement
"96.55% of pages on the internet get no traffic from Google" (Ahrefs, 2023). Same study, updated number.
The 90.63% zero-traffic stat is outdated. Ahrefs' 2023 refresh of 14 billion pages found 96.55% get no Google traffic, and that was before AI Overviews started suppressing clicks to non-cited pages.
9. "8 in 10 marketers say SEO is effective" (or some variant)
The origin
This one is a shapeshifter. Every year, HubSpot's State of Marketing Report publishes a slightly different version of this claim. In various reports it has been phrased as "8 in 10 marketers say SEO is effective," "49% of marketers say organic search has the best ROI," or "61% of marketers say improving SEO is a top priority" (HubSpot).
HubSpot's 2025 State of Marketing report, for example, found that 19% of marketers plan to build an SEO strategy for generative AI in search (HubSpot 2025).
Is it still true?
Something like this is probably true every year. The problem is that people quote specific percentages without attaching them to a specific survey year. "8 in 10 marketers" is not a finding, it is a vibe.
2026 replacement
If you want to say marketers believe SEO works, say it generally. If you want a number, cite the specific HubSpot report and year, like: "49% of marketers said organic search drove the best ROI in HubSpot's 2024 State of Marketing Report." Don't pass a four-year-old percentage off as current.
10. "Ranking #1 on Google gets 27-39.8% CTR"
The origin
The 27.6% number comes from Backlinko's 2019 CTR study, updated in 2023 (Backlinko). The 39.8% figure comes from Advanced Web Ranking's ongoing CTR dataset, which aggregates Google Search Console data from partner sites.
Is it still true?
Not anymore. This is the stat that has aged worst in the post-AI Overview era.
Ahrefs' December 2025 analysis found that AI Overviews reduce the click-through rate for position one by 58%, up from 34.5% in April 2025 (Ahrefs). For keywords that trigger AI Overviews specifically, the CTR for the top organic result fell from 0.073% in December 2023 to just 0.016% in December 2025.
Position 1 CTR dropped 58% once AI Overviews arrived. On AIO keywords, the top organic result fell from 0.073% in Dec 2023 to 0.016% in Dec 2025. The 27.6% Backlinko number is describing a different internet.
In other words, position one still gets more clicks than position ten, but the absolute numbers are dramatically lower than what Backlinko's pre-AIO dataset captured. Anyone quoting 27.6% CTR for an informational keyword in 2026 is quoting a different internet. This is exactly why serps.io tracks SERP features and AI citations alongside rankings: point-in-time CTR numbers go stale the moment a new SERP layout ships.
For the full picture, see our breakdown of AI Overview CTR impact.
2026 replacement
Use a modern CTR curve that segments by SERP feature presence. For an informational query triggering an AI Overview, expect 58% lower CTR than the old curves predicted. For commercial queries without AIOs, the old Backlinko curve still directionally works. And consider whether CTR is even the right metric anymore, given the death of the ten blue links.
Frequently asked questions
What is the most-quoted fake SEO statistic?
"75% of users never scroll past the first page of search results." It appears in thousands of agency decks, pitch decks, and blog posts, but its earliest traceable appearance is a 2010 HubSpot SlideShare citing a now-dead link. There is no controlled study behind the specific 75% number. The behavior it describes is real, just not that number.
Is the "75% don't scroll past page one" stat true?
The spirit is correct, the number is made up. Backlinko's analysis of 4 million searches found that only 0.63% of searches result in a click on page two, which is a much sharper version of the same idea. If you need a citation, use the Backlinko figure instead of the 2010 HubSpot slide.
What is the actual click-through rate for position 1 on Google in 2026?
It depends heavily on whether the query triggers an AI Overview. Ahrefs' December 2025 data shows AI Overviews cut position 1 CTR by 58%, and on AIO-triggering keywords the top organic result gets around 0.016% CTR. For commercial queries without AIOs, the older Backlinko curve of roughly 27% directionally still works, but anyone quoting a single flat CTR number for position 1 is oversimplifying.
How often should SEO statistics be updated?
At minimum, any number you cite should be rechecked annually, and any stat that predates the May 2024 AI Overviews rollout should be treated as suspect for click-through, zero-click, and SERP behavior claims. Backlink and word count studies age more slowly, but their causal claims were never strong to begin with. Always include the year the study was run, not just the year the blog post citing it was published.
Where do the "zombie stats" come from?
Most of them originate in legitimate studies from 2010 to 2017 that got stripped of their context as they were quoted and re-quoted across marketing blogs. The original HubSpot, CEB, SOASTA, and Skyword sources were often real, but the specific numbers drifted, methodologies were lost, and the underlying pages went offline. What remains is a game of telephone where the stat outlived its source.
How do I fact-check an SEO statistic I see quoted?
Trace it to the primary source, not the blog that cited it. Look for the original study, the sample size, the year, and the methodology. If the primary source is a dead link, an inaccessible PDF, or "an infographic," treat the stat as folklore. Tools like the Internet Archive help recover dead sources, and a quick check of the study year against major industry shifts (AI Overviews, mobile-first indexing, HCU) will tell you whether the number still describes the current internet.
How to use statistics responsibly
A few rules we try to follow whenever we publish a number on this blog.
Always cite the primary source. Not "HubSpot says X" with a link to a random blog post citing HubSpot. The original study, the specific report, the actual PDF. If the primary source is dead, admit it and flag the number as folklore.
Include the year. "53% of mobile users abandon" is meaningless without "in 2017 on a simulated 3G connection." Behavior changes. Context changes. Benchmarks change.
Check when the study was last updated. The difference between "90.63% of pages get no traffic" (2020) and "96.55% of pages get no traffic" (2023) is a decision that probably affects your content strategy. Use the current version.
Challenge single-number claims. Any statistic that reduces a complex system (buyer behavior, CTR, engagement) to one percentage is approximate at best. Real systems have distributions, segments, and edge cases. Headlines use averages. Reality uses percentiles.
Be especially suspicious of pre-2023 SEO numbers. The introduction of AI Overviews, AI Mode, and LLM-driven search has reshaped almost every downstream metric. Pre-2023 data on CTR, click distribution, and zero-click behavior is describing a different product.
If you cannot find the primary source, stop quoting the stat. This is the hardest one. It means dropping some of the most quotable numbers in your deck. But a memorable-but-wrong number is worse than a boring-but-true one.
For our own ongoing work, we maintain a running list of verified AI search stats in our 2026 AI search statistics roundup. Every number there is traced to a primary source and dated. That is the standard we think the rest of the industry should adopt.
The takeaway
Of the ten statistics we audited, one is still accurate and well-sourced (SparkToro's zero-click number), two are directionally correct with caveats (CTR by position, backlink counts), three are outdated but have valid modern replacements (mobile speed abandonment, pages with no traffic, and B2B buyer journey), and four are essentially folklore that should be retired (75% never scroll past page one, 94% more views with images, specific word count recommendations, and any "X% of marketers say SEO is effective" stat without a dated report).
The common thread: the SEO industry has been running on 2011-2017 data in a world that has been remade by AI search. If your strategy document leans on any of these numbers, update the citations. If a vendor pitches you on any of them, ask for the primary source. And if you publish them, at minimum, flag the year.
The numbers that actually matter in 2026 are different. AIO-adjusted CTR curves. Zero-click rates by query type. Citation frequency by LLM. The percentage of traffic going to the open web. Those are the new benchmarks. The old ones are ghosts.