[ssebro]

That's an interesting powerpoint at the bottom of the techcrunch article.

I see your argument that slower page loads leads to less engagement and higher bounce rates, and I'm somewhat skeptical - how do you know that your data isn't saying that more engaging websites are faster because they're more engaging?

If being engaging is correlated with delivering more value, they're likely more profitable and more able to invest in higher quality engineering & infrastructure than other sites. This would likely lead to decreased load times. In other words, how do you know the cause from the effect?

I think the only way to make the argument you're trying to make is to compare each website with itself when loaded slower.

[joshfraser]

Sure, correlation does not imply causation. I think most of our customers are smart enough to understand that.

The engagement charts are typically broken down for each site so that they are compared with themselves. It turns out, in both cases there is a strong correlation between performance and the business metrics you care about (bounce rate, conversion, time on site, # pages viewed, etc).