When Twitter suffered a brown-out last weekend, the "twitterati" had a collective conniption. I suppose the good news for co-founders Evan Williams and Biz Stone is that the bad news kicked up such a storm.
Lots of people are so addicted to Twitter that the intermittent problems wreaked havoc with their daily routine.
Lead architect Blaine Cook and VP of engineering and operations Lee Mighdoll are now gone. And it's up to management to come up with a plan. But this isn't the first time a popular online communication service found itself a target of criticism. In August 1996, America Online got in even bigger trouble after going dark for 19 hours.
How big a deal was it? Consider this: AOL's outage was the lead news item on the evening news programs for ABC, NBC, and CBS. If you thought the grumbling about Twitter was bad, remember that AOL back then had more than 5 million subscribers and they were not a happy lot.
History rarely repeats itself exactly, but the Twitter-AOL parallels were close enough for me to seek Barry Schuler's perspective. Schuler was a high-ranking executive at the company and became chairman and CEO of the AOL unit after the company merged with Time Warner in January 2001. These days he is a managing director of Draper Fisher Jurvetson's Growth Fund. Here are excerpts from the interview:
It was very traumatic, a real shock. We had surpassed Compuserve and Prodigy, which had been the top online services of that time. We used to take the service down in the middle of the night (to) upgrade our software. The concept of 24/7 data centers hadn't yet happened...When we went down, it was triggered because of a third-party back-end bug and a confluence of new software that took (the) system down. Obviously we were very upset.
Then the calls started to come in. Lots of people were telling us, "Hey, I'm out of business today because I'm dependent on your e-mail." It was an epiphany for us. We knew we were growing and obviously thought there was a big future for the business, but we didn't realize how integral AOL was becoming to peoples' lives.
In a way, it was like the old Coke-new Coke episode. Remember, they brought out new formula to beat Pepsi. And lo and behold, they found out how much their customers really loved old Coke. So here was an opportunity for them to create more loyalty. So our takeaway was that it was not good enough to be up most of the time. The services we were providing needed to reach dial tone quality. We needed to figure out how to be up 24/7 and have redundancy and built-in procedures and controls on the backend that wouldn't allow major outages to happen.
The difference between way AOL grew and some services like Twitter or Meebo is that we integrated a huge audience. We had millions of people online simultaneously and that took a huge investment in computing horsepower to make it all work. When it comes to Web technologies, no one has replicated that sort of scale when AOL was at its prime. But now they're facing that kind of growing pain issues that we had to power through.
My message to Twitter is no different. We're getting habituated on your service and we're dependent upon it. You can't take the point of view that it's free and downtime is to be expected--not if Twiter wants to keep the loyalty of consumers. You cannot underestimate the importance of being dial tone grade.
We may accept low quality service cell phones. But when we pick up the phone, we expect a dial tone. When you send e-mail, you expect it will be delivered. So this clearly is something Twitter has to take seriously. If they do go ahead and understand that they need to have a bullet proof back end--and that sort of thing is expensive--they could wind up with the same good news success story. If not, people are going to go to someone else.