A/B testing your life

Posted by on May 9, 2010 in BlogPost | 2 Comments
I have been writing a bit about how to use the thinking of offline innovation to make better digital products. (flow over feature & Innovation Architecture). However, lately I have been pondering a bit about how to take online best pratices and apply them to my real offline life. Got the idea of doing A/B testing on areas of my life that I wanted to improve but was unsure what was the most optimial method to achieve it.  First I needed to make sure I had measurable data. The good news is that quite a few people  (e.g. fellow personal stats junkie Sam Lessin) are increasingly blogging about and enjoying datamining all sorts of aspects of our lives – and there are now loads of tools that allow us to do so. Here are a few I use:
  • my weight (My Withings)
  • my sleep pattens (Sleep Cycle) (waiting for my Wakemate that I orded 6 month ago)
  • my running speed (Garmin 310)
  • my heart rates (normal blod pressure machine)
  • my household energy consumptio  (Altertme)
So lately I have started to do A/B testing on all these things. Basiccly, try out different thesis on a specific subject and then do test of each of them for a few days. Applying the same methodology that we use for optimizing userflows and optimize towards what works the best. Here are some of the a/b tests I am running:
  • Effect on weight depending on intake of alcohol, different type of foods and time of eating dinner
  • Breathing technics influence on blood pressure
  • Most optimal running technics
  • Optimize home appliances to reduce energy consumption and so on.

So besides creating a bunch of ultra personalized behavior generated content and the pleasure of playing with data visualization – I am finding easier and more optimal ways to achieve life goals.

  • http://bluxte.net/ Sylvain Wallez

    Interesting idea, Henrik! But we do so many different things in a single day that can influcence the parameters you want to measure, that the results of the test may be skewed by some parameter you're not aware of, or haven't thought of. The human body and brain are still way more complex than any system out there!

  • http://www.neverreadpassively.com/ jasonb

    Sounds more like hypothesis testing than split testing. Use some statistical tools to test significance and you've got your self statistical hypothesis testing.