2 ms·
Not exactly. It's more that making the page faster is fairly predictably going to get you $X in additional revenue. Adding a new feature is going to get you a
by nostrademons 11y ago
Not exactly. It's more that making the page faster is fairly predictably going to get you $X in additional revenue. Adding a new feature is going to get you anywhere from $0 to $Y in revenue, but you can't know what $Y is until you've launched it and given users some time to learn about it. So the only way to avoid getting stuck in a local maxima - your current feature set, as optimized as possible - is to periodically try to shake up the page, add some new features, and measure their effects. After a couple years or so, the features for which $Y < their cost in latency & developer maintenance are killed, and a new round of optimization & code cleanup starts based on the current feature set.