Can historical usage patterns help us prevent errors?


Most hardware and software products are designed to be used for a relatively long period of time, typically one year or more. During that time, the user will grow accustomed to how to operate the product, and the product has the opportunity to collect quite a bit of data about how its owner uses it. With this in mind, it’s surprising that so few products actually take advantage of this wealth of usage data.

Specifically, I’m interested in how products can look at historical usage patterns to help prevent errors. For example, if you always set your alarm clock for either 6 am or 10 am, shouldn’t the clock remember this and show a warning if you set it to 6 pm by mistake? Or, if you always bill a particular client at one hourly rate, shouldn’t your accounting software let you know if you suddenly try to charge them a different one?

I’m not talking about sophisticated artificial intelligence here. Rather, I’m just pointing out that a simple heuristic based on user behavior over time can be employed to prevent costly errors. Simply keep a log of how a customer uses the product, and then warn them whenever their current behavior varies significantly from their historical levels. I’ve yet to see this implemented in any piece of hardware or software, but hopefully somebody will give it a try. Until then, I’ll be lamenting how even the best-designed products learn so little from our ongoing interactions with them.