Improving Routines (Part 5)

You decide you want to change something in your life. You might want to find a better system to keep track of your to-dos, reorganise your pantry, or streamline the process of a particularly boring task at work. How much time should you reasonably spend on improving your routines? Find out in this article, which is the fifth part in my “Learning from Algorithm” series. All previous editions can be found here.

Not all routines deserve equal attention. Taking notes by hand is something you might do a couple of times a day, but is a rather quick task. On the contrary, you only do your taxes once a year, but it takes days to complete. How much time should you spend to improve each routine? How much for the quick recurring one, and the long one that you only do once a year? 

Luckily, the brilliant science comic page xkcd has the answer! This neat table tells us how long you can spend improving the task before you take more time improving than actually saving. This has been calculated for a variety of frequencies (on the top) and time saving through the improvement (left). For each cell we make the calculation as follows (as we make all calculations for a five year horizon):

Time that is saved * how often it is done in a year * 5

A quicker way to process your emails (50 times per day) might only shave off 5 seconds per email, but you could spend up to 5 complete days (!!!) of coming up with such a solution and be still net-positive in your time (i.e. you saved more time than you spent improving your routine). If this is not an argument to automate your inbox I don’t know what is. On the other hand, spending hours to improve a task you do once a month by a mere 30 seconds is not improvement, that’s just procrastination. 

Next time you dedicate time to improving processes do a quick calculation comparing the time you spend on the improvement and the time you save through it. The results might surprise you!