The Need for Speed, 23 Years Later

23 years ago, the internet was quite different from the one we use today. Google didn’t exist yet, fewer than 20% of U.S. households had internet access, and those who did were using a dial-up connection.

It’s no wonder that people complained about slow speeds on every website we tested back then, because the internet and the computers used to access it were painfully slow.

What is surprising is that, despite today’s much faster network speeds and computer processors, people using the internet today are still plagued by the exact same frustration: slow websites.

The Internet Is Faster, but Websites Aren’t

You might be wondering whether people simply don’t notice how much faster today’s sites are because their expectations have increased over time. While it’s true that people’s estimates of wait times are sometimes exaggerated, in this case it’s not just a matter of distorted perceptions.

For the past 10 years, has recorded page load times for 6 million popular websites. ( is a part of the, whom you may know as the folks behind the WayBack Machine). The results are not encouraging: for webpages visited from a desktop computer, the median load time hasn’t improved. Today’s websites aren’t that much faster than they were 10 years ago.

Line chart showing change from 2010 to 2019 in median page load time and in the average internet connection speed
Median page Onload time tracked by the has remained about the same over the past 10 years, while, over the same period, the average internet speed for users in the U.S. (tracked by Akamai and has steadily increased.

As you might guess, the story on mobile is even worse — connection speeds have improved for sure, but, over the past 10 years, the mobile page load times tracked by Httparchive have actually increased.

Chart showing the change from 2011 to 2019 in the average page load time on mobile devices vs. the average mobile device internet connection speed
The average connection speed for U.S. mobile users has increased steadily in the past 10 years; meanwhile, the load times for mobile web pages over the same period have more than doubled. Connection speed data through 2017 is from Akamai (which in 2014 adjusted its methodology to record connection speeds from a larger sample of devices); data from 2018 and 2019 is from, a mobile analytics company. Page OnLoad times for mobile webpages are as recorded by

Increases in internet speed clearly haven’t solved the problem of slow websites. Of course, network speed is not the only factor that affects performance, so it’s not reasonable to expect speeds to have completely kept pace with network connectivity. But it seems like huge increases in network speed should make it at least somewhat faster to browse the web.

You may be wondering if this data is really accurate, which is a fair question, as there are many different ways to measure performance and speed, such as by sampling different selections of websites or using different milestones to identify when a page is loaded. For example, in 2018, Google reported an average time of 7 seconds for mobile pages to load content above the fold. But, since above-the-fold loading times from 10 years ago aren’t available, we can’t draw conclusions about trends in that particular metric. In any case, even this more favorable number is still 7 times slower than the recommended response time for navigating web pages.) The data is unique in that it’s been collected using the same approach for the entire decade, allowing longitudinal comparison. This data strongly suggests that the websites people encounter today aren’t that much faster than they were a decade ago.

How Slow Is Too Slow?

The basic rules of human perception of time provide a framework for understanding the effects of webpage delays: people can detect delays as short as 1/10th of a second, so anything that takes longer doesn’t feel ‘instant.’ Delays of just 1 second are enough to interrupt a person’s conscious thought process, changing the experience into one of waiting for the system to catch up, rather than feeling as though you are directly controlling the interface. This delay reduces conversion.

We’ve described how these delays of just a few seconds seriously hurt the experience of those trying to use a website. But you don’t have to take our word for it — in the past 10 years, the effects of slow-loading web pages on site abandonment and conversion has been proven time and again:

  • 2009: Google and Bing both reported that even half-second delays in load time resulted in measurably lower conversion metrics (number of searches and revenue per user).
  • 2010: A Mozilla experiment which found that reducing the page load time by 2 seconds led to 15% higher conversion.
  • 2011: Gomez reported on 150M pageviews across 150 sites and found that pages that took 6 seconds to load were 25% more likely to be abandoned than pages which loaded in 2 seconds.
  • 2016: Google found that increasing the load time of its SERPs by half a second resulted in a 20% higher bounce rate.
  • 2016: Google found 53% of mobile visits ended if a page took longer than 3 seconds to load.
  • 2017: Akamai aggregated data from 17 retailers (7 billion pageviews) and found that conversion rates were highest for pages that loaded in less than 2 seconds; longer load times correlated with 50% drops in conversion rates and increased bounce rates, especially for mobile visitors.
  • 2018: BBC found that for every extra second of page load time, 10% of users will leave.
  • 2024: just kidding, but we’re sure there will be many more compelling data points to add to this list in the future, because this problem won’t go away.

If you’re still not convinced that response time matters in your specific industry or for smaller organizations, check out this collection of case studies by, which documents how performance optimization dramatically improves traffic, sales, and usage metrics across many different types of websites and industries.


All this data still does not precisely define a magic number at which your site is ‘fast enough’ or ‘too slow.’

If you’re in a group of people being chased by a bear, you only need to be faster than the slowest person in the group. But that’s not how websites work: being faster than at least one other website, or even faster than the ‘average’ website, is not a great achievement when the average website speed is frustratingly slow.

Subsecond load times would be ideal; realistically, this target is incredibly difficult to achieve while also delivering rich and engaging content. But, if there’s one thing that’s been amply proven in all the experiments listed above, it’s that each incremental improvement in speed pays off. Instead of settling for being ‘fast enough,’ investment in performance optimization should be driven by asking, ‘how much more successful would we be if we were 1 second faster?’ Reducing page load times by even a second, will improve your users’ experience, and increase your conversion rates. The slower your website is, the more you have to gain from making it faster.


  1. “Loading Speed” Report, 2010-2019,
  2. Akamai State of the Internet Connectivity Reports, 2010-2017
  3., “Worldwide Broadband Speed League” reports. (This data is download speed rather than the connectivity speed reported by Akamai for 2010-2017, so is slightly inflated relative to the Akamai data)
  4., a mobile analytics company, “The State of Mobile Network Experience” 2019 report,
  5., a mobile analytics company, “The State of Wifi vs. Mobile Network Experience as 5G Arrives”,

Read More

Leave a Reply

Your email address will not be published.