Antifragile Metrics

For some time I’ve been thinking about metrics for antifragility.

I first came across Troy Magennis’s framework: the ‘Six Dimensions of Team Performance” (perhaps worth reading first) and remember thinking at the time that his framework was fairly close to describing an antifragile system already.

I thought I’d have a go at modifying it and supplementing it, mainly as an opportunity to think aloud and hopefully get a conversation started.

Provisionally: The eight dimensions of antifragility!

How much / amount – assuming that an organisation is there to deliver some value, how much value is being delivered.

How fast / responsiveness – this could be the cycle time between request and delivery or perhaps between stress (harm) and response.

How predictable / consistent – the responsiveness needs to balanced by an ability to deliver or respond according to a steady cadence – can an organisation respond steadily or is it start & stop? I’ve typically used velocity here.

Do it well – outcome delivered vs outcome planned. I’ve refined this one by using Stephen Bugay’s terms here regarding the ‘three gaps’. Tracking bugs and technical debt.

Value – outcome delivered vs outcome expected. Again either for the value delivered or the response to a stressor. Rework and satisfaction are popular metrics.

Resilience – can the system continue like this or is it burning out? There’s no point an organisation speeding up with the team burns out. Metrics I’ve used have been team health.

Adaptability – ability to tactically take advantage of opportunities or a surge in demand. A metric could be related to built in redundancy or perhaps a portfolio of opportunities.

Risk – A propensity take risks and even the leverage to take risks.

The beauty of Magennis’s framework is that the metrics need to seen as a whole – usually visualised as a live dashboard. If, for example, an organisation targets to improve a particular dimension, it may have a negative impact on others! So all need to be viewed. 

I added adaptability and risk to the equation and it probably still needs some thought.