Why measure pageviews per month?
I don’t get why companies talk about how many pageviews they get per month. Maybe it is a handy metric when thinking about revenue, but when thinking about performance, it’s not a convenient way to measure.
For example, “we get 50M page views per month.” Let’s translate that into something useful for an engineer: 20 page views per second. How about 2.5B pageviews per month? Roughly 1000 per second.
Much easier to think about what this means. And not so impressive.
Further Reading:
Your calculations usually doesn’t make sense either though because chances are pretty good your site doesn’t have a constant flat level of traffic. It’s precisely because of that kind of quick calculation becomes attractive that pageviews per month is useless.
Most people who aren’t naive talk about performance, especially in provisioning, talk about peak qps or pages/s and when provisioning in the cloud where you can more easily scale down also talk about variability rate so that they can predictably scale down to save money.
If anyone is talking about their number of page views per month to a technical crowd they’re doing it wrong.
Guillaume Theoret
1 Oct 10 at 10:22 am
I thought about peak -vs- average too, but got distracted and forgot to write something about that. Thanks.
Xaprb
1 Oct 10 at 10:59 am