The obvious first step to dealing with critical issues (once you’ve defined them) is to spot them.
“But Evan,” you might say. “It’s pretty damn obvious when our site goes down.”
True, often it’s going to be very clear. You get 20 tweets and 50 tickets saying your site is down and you get to work.
But there will be plenty of times when critical issues that are a lot harder to spot…which can lead to very angry customers and/or lasting damage. A great example: if customer data is disappearing gradually, you might not put together that it’s a site-wide phenomenon right away. You might think a customer or two fat-fingered something and accidentally deleted their data. But once you realize a lot of folks are experiencing the issue, a ton of customer data may have been lost forever. Goodbye customers, hello TechCrunch exposé.
Here’s a few tips for catching those less-obvious critical issues…
Regularly communicate with your team
Especially as your supportteam grows, no one person is likely to have awareness of all the support tickets you’re getting. this
See something that would be worrisome if it were a trend? Share with your team in a more casual setting. We use HipChat and will frequently share concerns in the Support room:
Most of the time the answer is “no, I haven’t seen anything like that.” But occasionally you discover that everyone’s had one or two tickets about the issue. Uh oh – this might be a widespread critical issue!
Note issues in your bug tracker
If you use a bug-tracking system, make sure that all support agents search for a bug before submitting a new one. If the bug already exists, leave a comment. It’ll become clear what’s trending pretty quick as the comments stack up. (Thanks to UserConf alum Mathew Patterson for that tip!)
Give your users descriptive custom fields
Support ticket submissions are pretty generic (“I have an issue”) unless you prompt your users to provide more info. But if you include descriptive custom fields, many customers will fill them out and give you a good idea of what’s going on. “Installation>Error message” is much clearer than just “It won’t work. (Props to AVG’s Jon Meyer for that nugget!)
A good helpdesk will provide you with reporting about trending custom fields, tags, etc. (We’re biased, but we recommend our own UserVoice helpdesk.) A spike in a single field makes it clear that something went wrong.
Optional: give users an emergency lever
As I wrote about a few years ago, Rapidbuyr actually lets customers mark something as “urgent”, which will send a text message directly to their Head of Customer Support. Sounds scary, but they claim they didn’t see much abuse, and it immediately gave them a head’s up when something was going very wrong.
Hopefully you’ll find these useful. But there are always more tactics to implement. Which did I miss?