You don't want it to happen. You fool yourself into thinking that it won't. But then it does: Your app or site goes down.
You've done your best to prevent this, but sometimes there are forces beyond your control, like the power outage in Northern California on Memorial Day evening that brought down part of Amazon Web Services' Elastic Compute Cloud, and along with it many sites and many apps, including ours.
This sucked for obvious reasons but also because we had recently gained a lot of new customers who would be a) wondering what the heck was going, and b) ready to drop us like a crashy hot coal.
Besides swiftly addressing the emails that were pouring in, Twitter and Facebook were the obvious places to post updates for our customers. But what would we say, exactly, and how would we say it?
See what others were saying. We needed to address the fact that yes, we were down (it's us, not you) and that we were working hard to recover, but I wondered if we should say why. Would people know, or care, what "AWS issues" were, or what a power outage in Northern California had to do with the app in their hands? Would we seem like we were blaming someone else for our troubles?
Knowing that others had also been affected by the outage, I searched Twitter for terms like "AWS outage" and "AWS down" and found some great, straightforward examples ("We're currently down due to AWS issues") that told the truth in a non-accusatory tone. Even better was inlcuding a hashtag (#AWS) so that our followers could see that while it was us and not them, it wasn't only us, and they could learn more about the outage if they wanted.
Update regularly, even if nothing has changed. In his book, The Ten Faces of Innovation, Tom Kelley warns us about the Doorbell Effect: that seemingly interminable lag between when you push the doorbell and someone answers the door. The longer you hear only silence, the more likely you'll turn around and leave. But if you hear someone call, "Coming!" even once, the more likely you'll stay.
We knew we had to keep reassuring our customers that we were still aware of the problem and that we were still working on it. Tweeting or posting a new message about every hour or less seemed to be the right amount.
Be a little vague. A few hours after our app went down, it seemed to be back up. However, because it was still unstable, we didn't want to say, "The app is back!" only to have it go back down again.
We settled on a message that communicated our progress but remained a little vague: "We're recovering from #AWS issues from earlier this evening. Hope to be 100% soon!"
Thank your customers. Thank them for their patience, for sticking around, for putting up with your issues. They could easily dump you and go somewhere else.
Also, be patient! You know you're working hard to fix stuff, but they don't necessarily. Nor do they know they're the 20th person asking, "Hey it's not working, what's going on?" Apologize, thank them, briefly explain the issue, and give an estimate as to when the problem will be resolved, even if it's just "hopefully soon."
Have some fun -- your customers and your co-workers will thank you. While you don't want to downplay or brush off the problem -- the last thing you want to do is invalidate your customers' frustration -- you might as well have some fun.
And what's more fun than a hamster eating a tiny burrito? Only a hamster eating a tiny pizza. Inspired by creative 404 error pages, I decided to include links to such videos in our outage updates ("While we work hard to bring you back our app, enjoy these hamster butts"), figuring it was an opportunity not only to practice good customer service but to be entertaining and showcase our voice and personality.
But having fun with our outage messages had an unexpected benefit: My co-workers liked it too. Watching rodents nibbling adorably on miniature food helped lighten the mood during a serious situation. Sure, things really suck now, but look at how cute this hamster is! The videos made us laugh, and laughter and humor, researchers say, have a ton of benefits, including reducing stress, improving problem-solving abilities, increasing hope and optimism, and reinforcing group cohesiveness, all of which we especially needed during that time.
We could have handled the AWS outage in a more straightforward way. We could have posted a simple message and left it at that. But we decided to inject a little humor, and as a result, not only (hopefully) made our customers laugh but ourselves too, which only stands to benefit our customers in the end (so to speak).