Faster sites mean users do more stuff

I remember a study, can't find it at the moment since I'm on the phone, that showed that slower sites have much higher engagement than fast loading sites. Fast loading sites also have a larger bounce rates than slower loading sites. I have no idea why that is but the data was conclusive. So maybe fast loading sites are great for transactions (shopping etc) where the speed is important but maybe not so when you're loading someone's profile. My best guess is that by waiting, people are investing time into a page and they want to check it before they hit that back button. (ping me if you want the source and I'll dig it out).

I don't have the exact numbers on hand, but the game I worked on saw a similar 10-20% increase in daily pageviews as an immediate result of making the site faster. I would optimize a few core pages or add more capacity and, like magic, traffic would jump the next day.

Interesting. And it comes after I realized how slow Salesforce was (for my dev account at least).

What would be the things to work on to increase the performance? The article mentions mostly the size of the files. I also understand that the lower the number of different files is, the better. What else should be taken into account on the front-end? For example, if I were to use jQuery only for a couple of things, would it be better to have a bigger but simpler JS file? (as in: I rewrite what I need myself but can't match the file size)

The quote about the 30 results sounds tenuous at best.

Did the user search 20% less often because of an imperceptible page load time, or because they became overwhelmed with the number of search results and decided to quit?

User experience drives more activity than performance. Perf is important, but that particular example sounds totally bogus to me.