>A while ago, someone claiming to have worked at Amazon said that downtime doesn't really affect things as much as you'd think. He said most people simply just come back later.

Brick and mortar stores found that out ages ago. They were closed for large parts of the day/night and the customers just came back the next day.

If people didn't leave Tumblr and Twitter, with their constant massive outages (at some point in their life), when why would the leave Amazon, a huge established player, for a few hours outage?

Funny enough, I've gone to a brick and mortar store to find it was closed, then went back home and bought the item on Amazon. I wanted it right then, but since I would have to wait until the next day anyway I just ordered it online.

Contrast that with going to Amazon and finding their site is down or performing poorly; I have never gone to the store to buy something instead. If I was already going to be ordering it online, I was already resigned to waiting a day or two for it to arrive.

I doubt many would leave Amazon permanently because of a service outage, but it probably does cost them something in impulse purchases, if the impulsive desire for the item passes during the downtime.

No, you would think that, but actually if you read the comment referenced above, what they actually found is that outages seem to have little effect on revenues. That's why it's so surprising, really. The implication seems to be that their customers don't actually spend a whole lot on impulse buys.

Tumblr and Twitter have orders of magnitude fewer users than Amazon.

Seeing that Twitter has 500 million users, an order of magnitude more would be in the range of 5 billion, and a second order of magnitude would be around 50 billion.

So I seriously doubt Amazon has "orders of magnitude more users" that Twitter. For Tumblr, maybe, but I doubt that too (maybe just one order of magnitude for it).

Besides, all that's orthogonal to my point. Except if you mean that the reason that Twitter and Tumblr have less users than Amazon is that they left those services due to outages.

Twitter has much better uptime than it used to, too.

When you get in the hundreds of millions of users, you've run out of early adopters and end up with users who are a little more demanding about uptime.

I imagine the difference between latency and downtime is that latency tends to occur every time you visit, while by definition, downtime is more rare. In other words, latency provides a bad experience, while downtime provides no experience.

The Oatmeal actually has a great comparison of how people react to latency vs how they react to downtime: http://theoatmeal.com/comics/no_internet

the latency vs cost curve will not be linear. After some latency, increase in latency wont affect cost much.

Also, latency (a constant lack of resources) and down time (an extraordinary lack of resources) are two very different things. I wouldn't be surprised if some down time had little impact on sales, whereas latency has a lot.

true! latency is also location dependent. so in latency network and health of other external 'resources' also matter. Whereas downtime is only a server issue. So they are two very different things.

This is true by my experience. I went online last night to buy a replacement garden hose. Was surprised Amazon was down. After confirming it wasn't just me (thanks isitdownforeveryoneorjustme.com), I gave up and ordered it this morning.