I love the Quantified Self movement, am a product manager for Twitter Analytics, love OKRs, and tend to frame everything in terms of metrics and data.
At the same time, I’m a huge skeptic of data-driven culture, fitness trackers, A/B testing, and the statistics you hear about on tech blogs and in the mainstream news.
The reason for this seeming inconsistency is that I feel like there’s a difference between data and meaning. And yet, our default assumption as an increasingly data-driven culture is to assume all data has meaning, and is worth optimizing in the “good” direction.
If you’re a writer who makes a living off of book sales and page views, it seems obvious that you should optimize for those things.
If you’re a CEO who raises money based on monthly active users, it seems obvious to optimize for that.
If you’re a student who wants to be smart and successful, it seems obvious to optimize for good grades.
And yet, optimizing for those things above everything else will lead to click bait headlines, black hat growth hacking, and a deadening of curiosity and self-driven exploration.
It’s sort of subtle, but also crucially important, to prioritize leaving your most important metrics intentionally unoptimized.
When writing, leave page views intentionally unoptimized and instead optimize for audience connection.
When building product, leave MAU intentionally unoptimized and instead optimize for blindingly obvious and unique value to the user.
When going through school, leave grades intentionally unoptimized and instead optimize for mastery of a topic of interest.
Leaving something intentionally unoptimized, however, doesn’t mean you can’t measure it as a proxy for your less tangible real goal.
The difference is that the number you measure isn’t your primary goal, but merely a data point to reference.
The twist is that by leaving some proxy metrics unoptimized, while focusing on real value, they will often grow faster than if you tried to optimize them directly.
That’s my BART commute musing of the morning.
Don’t let data drive you. Use it as a fallible indicator of progress towards the ineffable quality of a meaningful impact.
— written on BART