Not a Leadership Blog #2
So, I had an interesting conversation with a friend yesterday. Well, not really a friend, more of an acquaintance bordering on estrangement during last couple of years. We did what grown-ups do. ‘Hello, long time no see’ as if that’s a bad thing. ‘Yeah, seems like ages, innit?’ not that I thought of him more than once during lockdown. But still, we caught up, took off our respective masks and struck some conversations over coffee. Floral decorations on the café terrace lit up by a sudden burst of sunshine as two self-important clouds decided to move away from each other.
At one point my pseudo-friend said, “I simply can’t understand why everyone’s so agog about Agile”. Clearly he trusts me to open up thus but I quickly said, “Hush, don’t let them hear you” “Who?” “Well, the waterfall guys, of course” I said. He is in the same domain as I am but still doesn’t get it. It is the same people who were proponents of conventional waterfall type planning are mostly the ones driving Agile in Financial sector. Why?
Recently I have been reading Kissinger’s classic ‘On China’. Absolute marvel from an absolute strategic genius. There he has explained at length how China has been able to retain a strong projection of its national identity for thousands of years despite being tormented by civil wars, dynasty changes, invasion and atrocities, and despite being militarily weaker than the invaders in most cases. The technique most often used by Mandarin and Han officials was of psychological nature. Convince your intellectual superiority over the ‘barbarians’. Humour them by wrapping your intellectual ideas in the garb of the foreign thoughts and thus appropriate their culture. Simple, right?
So basically rename everything you’ve been doing for ages so that it sounds like the new approach and you sound like its greatest proponent so that no one will bother you to change. It’s still taking years to deliver products? Who cares, we’re having sprints! We still can’t manage ongoing changes without escalating to the top? Heavens, laddie, we’re busy in scrums and spikes! We still need annual budgets instead of an error budget? Begone, we reshaping backlog!
While reading Kissinger’s book another revelation hit me. Core tenet behind Agile is that of a decentralised incremental innovation. But that in itself is a cultural by-product of liberal democratic thought. Can we have the shape of the final product emerge as an aggregation of many independent agent’s actions over time? Once we start to see the political parallel enabled by the core philosophy, next question seems obvious, “Does it always work?”
My Not-so-much-friend-Friend suggested vigorously, “Yes, of course, because liberal democracy has worked well, you see?” But problem is that it hasn’t. In many instances the freedom of choice, if not guided by an unifying purpose, can and has led to chaotic, and often violent, mayhem. It works where the system is closer to a steady state, but imposing it on systems that aren’t ready has bloody consequences.
I told him my anecdote on Random Forests. It provides a curious insight into counterintuitive failings on liberal methods. Think of an aggregation of decision trees whereby the final model makes a decision using a voting across the trees. Now think of introducing some new piece of information as a feature. You know that this is a fresh piece of information unrelated to any that you’ve already used in your model. This should help to improve the model, right? Think again! The new feature can end up introducing a number of new and weak classifiers and when you take a vote the weak classifiers can override the performance by their sheer number, making the model itself weak.
Incremental innovation works when your underlying agents of change are-
- Empowered decision makers (not a weak classifier) and
- United by a single vision and purpose
If you’re leading innovation anywhere, you always have two paths — liberal democratic one or the collectivist central plan. I’d recommend to consider both of the approaches in the light of what you’ve got as resources. For an organization that’s already running with a strong vision, liberal democracy may work for the better. But for more disjointed teams, it might be useful to develop the coherence first by running a central roadmap that leads to the vision.