Dunbar Number for data insights
Like how Robin Dunbar said humans can have about 150 social relationships, there is a similar number for the number of top level business metrics executives can track comfortably
Many of you might have heard of the “Dunbar number”. British anthropologist Robin Dunbar ran a regression (log-log, I think) between animals’ brain sizes and the size of their social groups. There was a rather good correlation. Extrapolating this, he proposed that humans can “comfortably maintain 150 social relationships”.
In other words, “Dunbar Number is 150” or I should say, “Dunbar Number for people is 150”.
This number has been used by several anti-social media activists to argue about how social media is bad for you, and there is no way that you can really maintain connections with all those connections you have. Then again, that is besides the point of this post.
On Monday, I was in Berkeley, talking about Babbage Insight with an acquaintance. He was talking about how, when he was CEO of Uber, Travis Kalanick would apprently himself drive the company’s weekly business reviews. “He had all the numbers in his head, and would know when something was off”, my correspondent said.
This tallied with how a bunch of my unsuccessful sales calls have ended - with the person on the other side saying something to the extent of “the number of metrics I need to track is so small that I don’t need any tech or analytics help for it. As long as I can see the numbers, I can quickly deduce what the stories behind them are, and what we need to do to correct it”.
Dunbar number for data insights is the number of metrics that one can “track in their heads without any analytical support
So this leads me to the new concept - “Dunbar number for data insights”. This is the number of metrics that one can “track in their heads without any analytical support”. I mean you will need reporting support - to report to you what these numbers are, so that you can make your conclusions. However, as long as the number of (high level) metrics you are tracking is well within this, the moment you see a number you know for sure if it’s within range or too high or too low, and whether you need to be worried about it, and if you need to be worried about it, what you should do about it.
I interrupted our conversation on Monday to talk about Dunbar number and how we can apply it to data insights. We then started discussing what the likely number is - the number of “high level” (without including the finer details) metrics that a business leader can keep in their heads, and have a clear view of what is happening, what the trend is, and what needs to be done about it.
My instinctive answer was 10. The person I was meeting said “maybe 5 or 6”. What do you think it is? Of course this varies from business executive to business executive. I guess the ones “with larger brains” can track more metrics in their heads?
PS: I’m in the Bay Area for another six days (until Tuesday evening). I’ll be largely based around Palo Alto during the day. If you want to meet, hit me up.
Firstly, you have just reinvented the Miller number. 7+/-2 is the number of things that human being can hold in their working memory, according to a famous paper by psychologist George Miller. That's a more useful starting point for analysis than the Dunbar number. But this too is a wrong starting point. The Miller number is the number of independent things we can hold in our working memory. If you're a CEO, many of the metrics will be in long term memory. In addition, you can track many more numbers when they are systematically connected.