Seven Seconds Used to Be Acceptable. What Counts as Fast Now?
Nobody tells you when the standard changes.
There is no announcement, no email from Google, no notice from your web host. Your website does not get slower. The bar simply moves underneath it, and one day the site that felt perfectly acceptable three years ago is quietly losing you customers to a competitor whose site loads in two seconds.
This is the most under-appreciated fact in web performance: “fast” is not a fixed number. It is a moving target set by what your customers have grown used to. And it has been moving in one direction, down, for thirty years.
Here is the documented arc, with the sources, and what it means for what you should be aiming at now.
1993: Ten Seconds Was the Outer Limit of Human Attention
The baseline comes from cognitive psychology, not from marketing.
In 1993, Jakob Nielsen published a synthesis of human-computer interaction research, drawing on Robert Miller’s 1968 work and on Card, Moran and Newell’s 1991 research, that established three response-time limits for any interactive system:
- 0.1 second, the response feels instantaneous. The user perceives no delay at all.
- 1.0 second, slow enough to notice, but the user’s flow of thought stays uninterrupted.
- 10 seconds, the absolute limit of a user’s attention. Past this, they stop waiting and go do something else.
It is important to understand why this research is still valid despite its age. These are not technology benchmarks that expire when hardware improves. They are measurements of cognitive constants, how long a human brain will hold a thread of attention while waiting for a machine. That number does not change when broadband arrives. It is the outer boundary of human patience, and in 1993 it sat at ten seconds.
That is our starting line. Ten seconds was the point at which you had definitively lost the visitor.
2024: The Average Acceptable Wait Had Fallen to Eight Seconds
Thirty-one years later, someone asked consumers directly.
Forbes Advisor commissioned a survey conducted by OnePoll of 2,000 U.S. adults in 2024. The finding: the average acceptable homepage load time was approximately eight seconds.
Note what happened. Nielsen’s ten seconds was a hard ceiling, the point of total attention collapse. The 2024 figure of eight seconds is an acceptability threshold, which is a lower bar being measured. Consumers were no longer reporting that ten seconds is where they give up; they were reporting that eight seconds is already past the line of what they consider reasonable.
The same survey produced a second number that matters more to a business owner than the first: when a page loads too slowly or is unresponsive, 48% of searchers simply visit the next website in the search results.
Not “48% form a negative impression.” Not “48% are mildly annoyed.” Nearly half go to your competitor, who is one tap away, and who paid nothing to acquire that visit.
2025: Forty-Seven Percent Draw the Line at Seven Seconds
One year later, the number moved again.
The Pixel surveyed 1,000 UK-based online consumers in 2025. Among the findings: over 80% said they were likely or very likely to leave a website that takes too long to load, and 47% cited seven seconds as the maximum acceptable load time.
Ten seconds in 1993. Eight seconds in 2024. Seven seconds for nearly half of consumers in 2025. The direction is unambiguous.
A necessary honesty note about this evidence. A third survey complicates the picture and you should see it. Liquid Web, in a 2025 report surveying over 1,000 consumers, found that 43% get frustrated if a website doesn’t load within ten seconds, a far more permissive threshold than either figure above.
These three numbers are not contradictory. They are answers to three different questions:
| Survey | The question being asked | Answer |
|---|---|---|
| Forbes Advisor / OnePoll (2024, n=2,000 US) | What load time do you consider acceptable? | ~8 seconds |
| The Pixel (2025, n=1,000 UK) | What is the maximum you will tolerate? | 7 seconds (47% of respondents) |
| Liquid Web (2025, n=1,000+) | When do you become frustrated? | Over 10 seconds (43% of respondents) |
“Acceptable,” “maximum tolerable,” and “actively frustrating” are three different mental thresholds, and they should sit at three different numbers. Read together, they triangulate to a current consumer tolerance window of roughly seven to ten seconds, tightening toward the bottom of that range.
All three are consumer surveys rather than controlled experiments, and two of the three were commissioned by companies that sell performance-related services. That is worth knowing. It is also worth knowing that they were conducted independently, in different countries, with different question framings, and still point the same direction.
2026: The Professional Standard Is 2.5 Seconds
Now here is the gap that should get your attention.
While consumer tolerance sits somewhere between seven and ten seconds, the technical standard your site is actually graded against is far stricter.
Google classifies Largest Contentful Paint: LCP, the time until the biggest visible element on your page finishes rendering, which is the closest technical proxy for “the page loaded”, on a three-tier scale. “Good” is 2.5 seconds or under. Between 2.5 and 4.0 seconds is “needs improvement.” Over 4.0 seconds is “poor.” These thresholds are applied to real-user field data collected from actual Chrome sessions, measured at the 75th percentile of your visitors, and they are the values used in the HTTP Archive’s Web Almanac 2024 analysis of 16.9 million websites.
Separately, and as a matter of practitioner convention rather than a single citable Google directive, performance specialists have long targeted two to three seconds for mobile page loads. That convention and Google’s 2.5-second LCP threshold happen to converge on roughly the same place.
So: your customers will tolerate seven seconds. The standard you are measured against is 2.5.
That 4.5-second spread is not a contradiction. It is an opportunity window, and it is the single most useful fact in this article. If your site loads in six seconds, most visitors will not consciously abandon it, you are inside the tolerance window. You will simply lose to any competitor sitting at two seconds, on every visit where the two of you appear side by side, without ever seeing the loss in your analytics.
Why the Compression Keeps Happening
Three forces are pushing the number down, and none of them are slowing.
Habituation. Your customers spend their day inside applications that respond instantly. Every one of those interactions recalibrates their baseline for what a competent digital experience feels like. Your website is judged against that baseline, not against other small business websites.
Brand judgment has entered the equation. This is the qualitative shift. Slowness used to cost you patience; now it costs you credibility. In Liquid Web’s 2025 survey of over 1,000 consumers, more than 50% said a website’s speed reflects the quality of the brand. In The Pixel’s 2025 survey of 1,000 UK consumers, 42% said a slow website makes a brand feel definitely or possibly untrustworthy. Your load time is no longer a technical characteristic. It is a statement about whether you are a serious business.
And you do not get the patience that big brands get. The same Liquid Web survey found that 73% of consumers would wait longer for a site to load if it belonged to a brand they know and love. That buffer is real, it is measurable, and it does not extend to you. An unfamiliar local business gets judged on its worst second.
What This Looks Like on a Tuesday Night in Tampa
Consider, hypothetically, three restaurants competing for the same dinner reservation.
A couple searches for somewhere to eat and opens three results in sequence. Restaurant A’s site renders in 2.1 seconds; they are reading the menu before they have thought about waiting. Restaurant B’s site takes seven seconds; it is inside the tolerance window, so they wait, but their patience is now spent, and when the reservation form takes another beat to respond to the first tap, they back out. Restaurant C takes eleven seconds; they never see it.
Restaurant A did not have better food, better reviews, or a better location. It answered first. All three paid the same amount to appear in that search result, which is to say, they paid for the visibility and then two of them declined to collect on it.
Restaurant B is the instructive case, because Restaurant B’s owner will never know this happened. There is no line in any analytics dashboard labeled “lost because the form felt sluggish.”
The Number Has Changed, and It Will Change Again
Ten seconds in 1993. Eight in 2024. Seven for nearly half of consumers in 2025. Measured against a professional standard of 2.5 seconds. The direction has been consistent for three decades and there is no reason to expect it to reverse.
Which means “is my website fast?” is the wrong question. The right one is: fast compared to what, measured when, and against which competitors?
A free performance assessment from iServU answers that concretely. It pulls your actual real-user field data, what your visitors genuinely experienced, not a simulated test, maps it against Google’s current thresholds, and identifies where you sit relative to the businesses competing for the same searches.
Gabriel Lopez-Seco runs it directly. No account manager, no handoff to a junior analyst, no report generated by someone who has never spoken to you.
Schedule a free assessment at iServU.
Sources
- Nielsen, J. (1993, FOUNDATIONAL). Response Times: The 3 Important Limits. Nielsen Norman Group, synthesis of Miller (1968) and Card, Moran & Newell (1991). https://www.nngroup.com/articles/response-times-3-important-limits/
- Forbes Advisor / OnePoll (2024). Consumer survey, n=2,000 U.S. adults, conducted by OnePoll on behalf of Forbes Advisor. Reported by Search Engine Land. https://searchengineland.com/people-wait-website-load-2024-stat-445223
- The Pixel (2025). Consumer survey, n=1,000 UK-based online consumers. https://www.thepixel.com/insight-news/why-website-speed-matters
- Liquid Web (2025). Lost Value in Site Speed report, survey of 1,000+ consumers. https://www.prnewswire.com/news-releases/conversions-skyrocket-27-with-a-single-second-site-speed-boost-reveals-new-liquid-web-report-302361364.html
- HTTP Archive / Web Almanac (2024). Performance Chapter: Core Web Vitals thresholds and CrUX field data, 16.9M websites. https://almanac.httparchive.org/en/2024/performance
