dave snowden: managing in the present by designing parallel safe-to-fail experiments

dave snowden’s interview on the humancurrent was one of the few podcast episodes this year i felt like need a repeat listen. i could probably listen several more times and still be getting new things from it (the other was the last episode of season 1 of how to survive the end of the world).

even just his takeaway point(s) contained more subpoints than i can fully process. two things that i did pull down were (a) managing in the present is all we can actually do and (b) managing in the present happens best by running parallel safe-to-fail experiments.

managing in the present is how to handle complexity

although i don’t remember a ton of what he said about this in this episode, i’m a bit familiar (mostly via presencing) about managing in the present / learning to the future. in a sense, “managing in the present” is set up as a foil to two old paradigms of management: managing from the past and trying to manage the future. managing from the past is what happens when people look at old data, situations, practices, experiences, etc and make plans for the future from that information. the problem here is that the future is not guaranteed to be like the past.

and trying to manage the future is when people make predictions and things and then try to move from today into a future desired state via plans. the problem with that framework is that it’s not possible to know today if you’ll even want to be where you’ve decided you needed to go. by the time you arrive, your data will all be outdated. the world will be different and you might realize (too late) that you actually should have been trying to go somewhere else.

so in opposition to both of these paradigms, managing in the present emerges. but how does one manage in the present?

creating parallel safe-to-fail experiments

there’s actually much too much about this to write in this little snippet, but i wanted to talk out at least a smidge of what i’m thinking right now. basically, this idea of running parallel safe-to-fail experiments is opposed to two things. first, it’s opposed to creating plans that are failsafe. the problem with failsafe plans is that they don’t maximize learning because there is no risk. when you don’t risk anything, you don’t stand to gain the most valuable information you can in a moment/experiment. one way i think about this is that old saying “you learn a lot more from failure than from success.” that’s not to say that you don’t learn when you make failsafe plans. it’s just that leaving yourself the opportunity to fail, you can either (a) learn a huge amount from that failrue or (b) realize that you actually have figured something out and can take a bigger leap (i.e. creating the possibility to fail your way into the future). sidenote: this means, as g says, you have to learn to design experiments with risk that is in correct proportion to acceptable loss, but that’s a whole nother post…

second, it’s opposed to managing future states. by design multiple simultaneous experiments, you prevent your current self from placing normative judgment on the future. because the you of today can’t actually know what the you of tomorrow will want, it’s best to not decide today what will feel right tomorrow. this stance gives you the most adaptive position. it allows you to move forward today and yet still have the greatest number of options in the future. in a sense, this way of behavior is a resilience booster.

also, by having lots of experiments running, you are learning into the complexity that is. in the language of the cynefin framework, this is the difference between operating in complex vs complicated realms.

phew! that was much more than i thought i had on all of that. and that’s just from snowden’s takeaways. the episode is too dense to truly comprehend today. maybe i’ll listen to it several times next year and see what sticks next time.

anyway, that’s all for now. happy holidays!

words / writing / post-processing
685w / 21min / 8min