What It’s Like To…Achieve 100% Uptime

Adolescence is awkward, even for a company. AbeBooks was in the middle of a growth spurt—and frankly, struggling to adjust—when I joined the e-tailer in June 2003.

For AbeBooks, which sells new, used and rare books online, the growing pains came in the form of downtime. At one point, we had outages every two weeks. They usually lasted 45 to 90 minutes. We lost about $50,000 each time, and marketing would point out that its efforts were canceled by the blunders.

So I got all the managers together and told them I refused to believe that downtime was inevitable. My point was simple: When you’re 13, you can act a certain way. But you grow and mature, and you can’t act that way anymore.

First, I changed the operations team’s compensation structure. Instead of raises, we put in bonuses based on uptime. Their attitude changed immediately.

Then, we shifted the focus of the development group to quality and brought in some of the new code-scanning tools. That started to pay dividends too.

My goal for the May through July ’04 quarter was 99.998 percent uptime, but we were so busy that quarter that I didn’t get a chance to look at the numbers. Finally, in the first week of July, someone asked me how much downtime we’d had so far. I thought about it and said, None that I’ve seen.

I checked the logs and sure enough: zero downtime. I mentioned it to the operations guys and they shushed me. Like a no-hitter in baseball, if you talk about it, you jinx it.

Monday morning after the quarter ended, the ops team came into my office and asked if I had seen the uptime numbers. I had: three months, zero downtime. They told me I owed them and their wives dinner, which I had promised. They were beaming. These are people who normally hole up in their offices. Suddenly, they were visible and proud. They had T-shirts made that read: “We never sleep.”

Executive management was happy too. They felt as if now they could make strategic decisions using IT instead of fretting about it. Marketing was ecstatic. They’re planning television advertising now without worrying that the systems will collapse from a rush of people logging on.

The board? Well, it wasn’t quite champagne flutes clinking. But now they’re backing me on a lot of decisions whereas they wouldn’t have before.

Everything in IT is about trust. Now we have more of it.

And I’m sure there’ll be other 100 percent uptime quarters down the road.

Of course, you never forget your first.

-As told to Scott Berinato