Hacker News new | ask | show | jobs
by akamaka 6222 days ago
Edit: whoops. this is supposed to be a response to the first comment.

I would guess that the 1% figure is the average loss for each 100ms of latency. So if it took 5 seconds longer to load each page, 50% less people would ultimately make a purchase.

Which is still a very high number, and makes me wonder if most of Amazon's customers are impulse shoppers, or whether these numbers were just fabricated to help sell AWS.

Either way, it's not a very meaningful without knowing the context and overall probability distribution of lost sales vs. latency.