John Handy Bosma (Boz) proposed a personal productivity random moment study. His goals are:
- Find out how he’s spending his time in terms of the proportion between important and unimportant task
- Show the connection between what he’s working on and the business priorities
- Improve his productivity
- … and do all of that with at most 5 minutes of tracking a day.
The interesting thing about randomness is that it might have a different effect on behaviour. If you can’t anticipate when you’re going to get polled and you’re honest about your responses when you do, would that help you focus on more important things so that you don’t catch yourself goofing off during the polling time?
What are good questions to ask during the sampling moment? Boz has:
- What are you working on?
- Who are you with?
- How important is this?
- How is this related to the business objectives?
These questions also helped Boz stay focused – immediate benefit.
Questions/ideas related to tracking:
Is the effect of uncertainty worth the added effort required to build a custom tracking solution (or buy one), or will fixed time intervals be acceptable? If fixed time intervals are okay, then off-the-shelf apps can be stitched together for this functionality.
Is there value in full randomness (ex: five reminders randomly set for one day, even if those reminders all come in the morning) or is it more about moment-to-moment randomness (ex: a reminder set randomly in each 2-hour period)?
In which circumstances would interrupt-driven methods like this be better than time tracking or time-and-motion-type studies? Boz shared that he never quite got the hang of time tracking, so it might be about enabling a different set of people to explore this class of experiments.
Does measuring time (either through sampling or through time-tracking) offer significant benefits over, say, tracking quantity of tasks completed in different categories (like Andy Schirmer does) when it comes to measuring alignment with priorities?
I might give it a try. I like my time-based analysis, though, so I may increase the granularity of my time-tracking (track at the task level whenever possible). I can then simulate work-sampling based on that data. I might also try fixed-interval sampling using KeepTrack on the Android, although I tend to skip interruptions.