When the Data and the Anecdotes Disagree, the Anecdotes are Usually Right

Mostly metrics is proudly sponsored by Mercury.

I know I’m the metrics guy and all, but here’s something I believe: When the data and the anecdotes disagree, the anecdotes are usually right

Park Conrad, founder of Rippling, and Jeff Bezos, founder of Amazon, each shared stories about prioritizing anecdotes, and chasing down ground truths at the event level, before making decisions off of aggregate data.

Logan Bartlett was talking to Conrad about Founder Mode on his podcast and the concept of chasing down anecdotes to ground truths came up:

❝

“The best leaders tend to solve problems by going really deep and going into the weeds. If something is wrong in customer support one approach to solving this is to look at your team and say do I have the right leader? Great, they are they working on it. They’ll figure it out.

I’ve found that doesn’t work. If the problem gets to your level wherever you sit in the organization, you’ve got to go all the way to ground.

So the way to solve that is to look at the last 50 support interactions. Or you look at the last 20 Gong calls. You go and inspect the anecdotes rather than the data. You can get to an answer much more quickly by looking at a stream of anec-data.

And once you understand the problem it’s usually pretty clear what you have to do about it. You can thrash for a long time if you are looking too top down at it and trying to operate through your direct reports and management team.

It’s equally true for directors, vps, c-level execs, as it is for founders.”

-Park Conrad, Logan Bartlett Show

Conrad explains that data is great, but solving a problem needs to start at the ground level before you start to lean on data. I mean, how do you know if it’s even being measured correctly?

Plus, if he were to just wait for updates from his direct reports, context would get lost in translation, and the feedback loop to solving the problem would be extended.