What To Mind: Picking a Metric

Saturday, September 13, 2014
By dreeves

A little anthropomorphic ruler

We use the word “goal” a lot but, ironically, we agree with Scott Adams (of Dilbert fame) who argues that goals are for losers. [1] He points out that the most amazing people he knows tend not to just have goals that they achieve and then are done with, but systems for constantly improving. This is the biggest difference between Beeminder and StickK. StickK is about reaching a goal and the key question for setting up a goal in StickK is defining the success criteria. With Beeminder the key question for setting up a goal is defining a metric. [2]

If you want to be a writer you could beemind time spent writing, or number of pages, or number of words, or stories finished, or articles published. How do you pick which one to actually beemind?

  1. Prefer things you can automate

    E.g., time spent as measured by RescueTime or words added/deleted on Draft. The Beeminder frontpage lists all our autodata integrations. [3]

  2. Prefer things that make inherently interesting graphs

    We call it the QS First principle, and in fact I view Beeminder as foremost a Quantified Self tool. For example, beeminding miles run is more interesting than beeminding number of runs.

  3. Prefer actions over outcomes

    I.e., things you have direct control over. If you are setting a goal to “get tenure”, beemind the number of papers submitted for review rather than number of papers accepted for publication. We also recommend beeminding calories and workouts over beeminding weight (or at least beemind the latter very conservatively).

P.S. We made these points more concisely on Twitter, thanks to @sushimustwrite.


[1] Scott Adams’s criticism doesn’t really apply to S.M.A.R.T.(E.R.) goals.

[2] In fact, most Beeminder goals don’t have success criteria. They’re things like “work out at least 3 times a week, forever” or “spend an hour a day writing, forever” or “keep my weight below 100 kilograms, forever”.

And you should upvote autodata sources you think we should add next!

Tags: , , , , , , , , , ,