Notes on Designing for Antifragility

In this session, we looked at how to design for antifragility
Here are the principles put forward by Luca Dellanna and I’ll add some of my own comments (if applicable).

  • Bottom up (or maybe … outside in)

Objectives trickle down, innovation trickles up.
This is pretty standard for any strategy deployment since whoever is defining the intention of the strategy cannot also define how is to be achieved on the ground. Additionally, any learning requires a feedback loop. To me, ‘top down’ and ‘bottom up’ still sound very much like a hierarchical organisation. I would personally prefer to use a term something along the lines of ‘innovation trickling from the outside into the centre‘, and objectives trickling from the centre outwards, see diagram below.

  • Skin in the Game (SITG)

Essentially, if good things happen to the group, good things happen to you. Conversely, if bad things happen to the group, bad things happen to you. 
SITG acts as a filter and ensures that there people are taking ownership for their actions (avoids moral hazard). You should take risks but any outcome is on your head, not someone else’s. Incident investigations, retrospectives and kaizen events are all mechanisms that bring SITG. I think a component of this is also ‘single wringable neck’ – so a single person takes ownership of an action or a decision. Bureaucracies are often designed so that accountability evaporates through decision by comittee.

  • Fractal

Instead of a monolithic system, the system has constitute parts that make up the whole (like diagram above). There should also be mechanisms in place to ensure cross pollination of ideas happen.

  • Not brittle

  • Circuit breaker

A circuit breaker stops the risk spreading to other parts. The key thing about a circuit breaker is that it is automatic, and that there is learning that takes place as a result. The feedback for the circuit breaker needs to occur upon both leading and lagging data. Leading = information that shows a trend towards a particular outcome, e.g. people clicking on a help button. Lagging = the outcome has happened, e.g. clients are fed up and cancel their subscriptions.

  • Make leading indicators stressors / damaging

Unless the leading information results in a stress to the system (a harm) from which the system can learn, improve and adapt, then nothing will happen we will be at the hands of the lagging results – much more damaging, and the result is increased fragility. How to make leading indicators stressors, brings us back to SITG.

Part II – how to enable these principles?

  • Enshrining

Enshrine the feedback loops so that they are tight, specific, with both leading and lagging indicators. Cultures are made of rituals and of values who are considered worth expressing even when inconvenient -> principles

Interesting point, ‘Only institutionalised culture lasts’.

  • Teaching about principles

This is about teaching why the principles exist, even if they are inconvienient and may, at face value, appear to be slowing things down. As the Japanese say, sometimes you need to ‘slow down to speed up’.

  • How to demonstrate the values

It’s more important that a team lead repeats the need for a, e.g. security check once per day for a month, then once per month for a year. For a habit to form it requires regularity and can be formed after a few weeks (but not sporadically over a lengthy period of time).

In summary

First principles create the conditions for antifragility (to solve the problems)

Second principles ensure the first stay over time (to ensure longevity)

I’m going to ponder about this a bit more. I’m wondering if there’s anything else.

What comes to mind is this is mainly about the organisation and the execution, but there could be more to be said at the strategic level (intention, direction, and decision making). But so far, it was an extremely interesting insight.