This week's Commonplace piece is about 's ability to diagnose a product team's issues within 5-20 minutes of talking to them, and what we learnt when we did a tacit skill extraction session together:
Conversation
When John talks to a product team, the call progresses through four stages. The bulk of the magic happens in stage 2 (John intuitively knows what to ask and where to dig), with some leftover awesomeness happening in stage 3 (John gives them an ordered list of experiments to try).
1
2
A lot of what he does is picking up on tiny cues.
For instance (and also the biggest thing I learnt): good product teams are comfortable with uncertainty, and able to articulate what they know and don't know and are working to find out.
Bad product teams talk about the future.
1
5
Another cool thing: John is able to tell when someone is new to product. The cue — they talk about frameworks! Not always, but "good product leaders put all these things to practice (from the frameworks), but they don't talk about the frameworks. It's just something they do."
1
3
There's also this weird bit about product teams, which is sort of the reverse of the Anna Karenina principle: "bad product teams all kinda look the same, good orgs can succeed for very different reasons."
1
2
Another visceral thing I got from this is just a sense of the sheer VARIETIES of good product orgs there are out there.
John's expertise is really cool, and the number of years he's taken to build it is considerably larger than the hours we took to extract it ;-)
1
3
If you want more of this, read the rest of my tacit knowledge series (which includes the technique we used, in Part 5):
1
2
And check out cta.institute — which was set up by the researchers who invented Applied Cognitive Task Analysis. 🙃
1
3
Finally, follow me if you'd like more posts about tacit knowledge and skill extraction — I've got two interviews lined up right now. (Psst: newsletter link — commoncog.com/blog/subscribe)
Thanks for reading!
ACTA's really basic and yet super fascinating! Can't wait to try a second time.

