The Disadvantage of Doing Good

When it comes to engagement and retention, gambling and gaming apps have a certain kind of advantage over do-gooder products like DuoLingo or Coursera. A gambling app can optimize exclusively for habitual use (if the use is bad for the person, we call it addiction). It can spend every moment of the time it has with the user in that way.

But health and education apps — apps that want to provide some benefit — have to use some of those moments for other things: for whatever drives benefit and efficacy. And the way a designer uses a moment of user attention when he’s using it for efficacy will sometimes be different from the way he would have used it if he were optimizing for habit.

So who wins at engagement?

Imagine a skinner box. Say you make a contest between two scientists. Each scientist has a bunch of skinner boxes, and a bunch of rats. Their goal is to get the most lever presses out of a rat in a 2 hour window. One scientist only has to optimize for lever presses in the two hour window. But the other scientist has to also optimize for some other goal, like how well the rats solve mazes 2 weeks later. Who’s going to win the “engagement” contest?