> "Gmail is twice as reliable ... and higher reliability translates to higher employee productivity."

Actually, I bet the email outages are some of the most productive time in a company :)

Productivity is winding its way higher and higher up the meaningless words list, soon to join "natural", "selected" and "actually".

In fact, I may propose a pseudo-Victorian movement for Natural Productivity. Motto yet to be determined, actually, natch.

Uptime is great. What about Trust though?

Can you trust a another company like Google or Amazon with your personal/private data or your company or client's sensitive data?

Can you trust whatever random IT guy you hire to manage your exchange server?

I've thought about the trust issue many times... I think there might be a market for a "new" secure/encryption technology/business model focused on securing the cloud.

They may have millions of users, but with Google's searching, you can never be lost in the crowd, er cloud again.

"Gmail, and we thought it best to simply share our reliability metrics, which we measure as average uptime per user based on server-side error rates."

If you are just using error rates from the requests that make it to the application then it seems this measure of reliability doesn't account for a lot of infrastructure at Google (ie, load balancers, switches, etc). Service interruptions and downtime with infrastructure like this should definitely be accounted for by a measure of reliability for Gmail since the end-user needs all of it to work to use the app.

One solution could be factoring in some sort of availability monitor at the border of Google's networks that polls/uses the Gmail service like an end-user.