In some of my initial prototypes and sketches for the applications, I envisioned the system as something more game-like. The system would need to somehow determine a score that would apply to a user’s distraction level

One method of scoring was based on the idea of a speedometer. The faster you worked, the higher your speed would go. (Although, the opposite metaphor could also have been applied.)

In one sketch, I thought it would be interesting to give out badges to users who had achieved certain scores during their day of work.

After some more thought, I scrapped the concept of badges primarily because it made the system feel too much like the game. Likewise, I dropped the idea of keeping a score based on an entire day’s work. If a user had a distracted morning, for instance, it would be discouraging to have a low score for the entire day. Likewise, a high score in the morning might send the wrong signal that a user could relax for the rest of the day. Instead, I decided that the score would represent only the last hour’s worth of activities .

Eventually, I simplified the concept even further. The score would be based on the amount of time that a user had spent on a productive activities. Distracting activities would subtract from the score. The side effect of this decision would be that distraction would have twice the impact that productive activities would have. In early prototypes, this didn’t seem balanced.

Later on, I balanced it a bit further so that the system would give users credit for productive time, and would not remove credit for distracted time. This made it somewhat simpler to describe: The score measures a user’s productive time.

However, this still felt too complex. The score ranged from 0 to 100, with 0 being completely distracted and 100 being fully productive. The score number still felt opaque. Instead, I decided to use the raw numerical value. The score would represent the number of minutes in the past hour that had been used for productive time. The score would work on a well understood scale of 0 to 60, and it would be easier to understand how to improve the score: spend more time on productive activities.

The score system cascades through the rest of the system. I had struggled with the decision of what to do after a user clears an intervention and therefore decides they want to spend more time being distracted. By using time, the decision was made easier. An intervention would appear when the score fell below a certain point. By clearing an intervention, a user would buy more time before another intervention would appear. And the complexity of the intervention would change with the level of the score. High scores get easy interventions (so it’s easy to distract yourself), and low scores would get very difficult interventions, so the user would focus on work.

The score also cascaded through the social system of the game. When a user asks a teammate to “unlock” an intervention, it was unclear what would happen. But after deciding to use time as the scoring system, that decision was simplified too: An unlocked intervention would have a longer period of time before another intervention would appear.