Page speed is overrated

Most page speed preachers are wrong. They're flawed by bad data.

I still see many people preaching page speed in conversations about conversion rates. Most of them are wrong. The discussion is flawed by insufficient data. Here's why.

The backing argument is always bounce rate (BR) data. For example, this article by Section states that the average BR of a page taking 2 seconds to load is 9.6%, the BR of a page taking 3 seconds to load is 13%, etc. But when you think about it, BR is a horrible way to measure the impact of page speed.

A 13% bounce rate simply means that 13% of your visitors left without seeing other pages. There are 1000s of reasons for someone to leave without seeing other pages. Irrelevant content, confusing sites, and anything else that make your site less interesting.

So, to isolate the effects of load times, "researchers" like Section need to account for all other possible bounce reasons. Of course, they don't. That's impossible. So they don't know whether a higher bounce rate is caused by higher load times or just because of all the irrelevant content that leads to higher load times. Correlation does not imply causation. All else equal, you need more content on a page to make it load slower. So maybe a slower loading page is just filled with more garbage?

Think about it. Do you really believe that going from 2 seconds to 3 seconds in load time makes 3.4 percentage points more visitors bounce? I don't. Remember: We're not talking about someone visiting 13 different products on your site and growing tired of waiting the extra second. We're talking about a single page load.

I think it's much more likely that visitor simply doesn't care about all the extra, irrelevant content that makes the page load slower.

I'm not saying you shouldn't care about page speed. You should. Optimize your images and videos. Be cautious of apps and scripts.

But don't obsess about it. It's not a magic bullet. I think you're fine if your site loads faster than 3 seconds. Spend your time improving something else.

Note: The 3 seconds are 100% arbitrary. I don't know where the "optimal" threshold is.

Finally, remember source criticism. Why do you think companies like Section talks about page speed? To sell you their products.

Godspeed,

Mathias

PS: If you still want to improve your site speed, learn how to find your speed sinners here.