The optimization trap

2 minute read

There was a great development team that was integrating Twitter for the first time in their web application. The team was very knowledgeable in usability and able to create simple but very effective interfaces, with a good skill in creating fast and focused flows. They took exactly the same approach with Twitter integration, reducing at the bare minimum the number of clicks and the hassle for the user.
Once released, they found that most of the users were very, very unhappy about the new feature integrating with Twitter. At first they didn’t understand, because they also tested it with a few friends and all were very happy of that. But it took the time of clicking on a feedback and reading the first message to understand why: “I never told the application to tweet!”.

From another perspective: could you ‘optimize’ a game? What about a big button that says: “I win” and you just click it? Well, that’s almost how Progress Quest works, but no, a game works exactly because it creates an experience through challenges (I’m simplifying, of course).

The optimization trap appears when you take one good metric and you try to apply it to everything. The problem is that we are human beings and as such we are complex and we have very different needs and motivations helping us to achieve things. So, it’s very important always to double check the metrics and principles you are using to focus your design.

In the above scenario, optimizing for execution time was absolutely great and the right thing to do, but only to a certain extent: it should have taken into account that people had a very specific need of control that is more important than execution time.

It’s very easy to fall into this trap because it’s not related to doing something bad, but in exceeding doing something good. In my experience it surfaces more the more you are skilled in doing some specific task. That’s why feedback and team work is so important, because different people will show you the different perspective and needs.

So, how to avoid it? First of all, get hands-on feedback from strangers, that’s the best thing you could do. It’s important to be hands-on, because if you just show something to a friend, he will just say “nice”, because the feature is nice. It’s the subtlety of the action that explodes in a social context when you actually think of the impact of your choice – or lack of – that will tell you if a design choice works or not.

Do you have any story about the optimization trap to share? Or do you have any technique to avoid it?