Note: This is a follow up to Thinking probabilistically about engineering.
As a lead of leads, one of my mantras is that everybody has two bad weeks a year. It may be because one of your reports is having some serious career doubt, or because something unexpected happened in their personal life, or for no reason whatsoever. The point is that it’s outside of your control. Thus, for a lead of leads with 25 indirect reports, on average, every week one team member will be going through a difficult situation the genesis of which their manager likely have no control over. It’s a probabilistic inevitability.
Accepting this mantra, a manager’s focus should be two-fold: (i) do whatever is necessary to lower the probability of bad events, but understand and accept that it will never be zero; and (ii) always have a contingency plan for the most damaging unavoidable events.
Interestingly, one of the things that makes experienced managers stand out is knowing what those unavoidable events can be. New managers often don’t even know all the kinds of failures their team can experience, which is why they often focus on the wrong things (like doing individual contributor work like writing code) instead of establishing process and back-up plans.
Getting back to the “2 bad weeks a year” mantra, managers should always be ready for anybody on their team to have a very serious set of misgivings about your company, their career, their work life balance, their growth, their compensation, their teammates, the strategy, the product, etc. The deterministic approach — the illusion that you can even stop this from happening — will only lead you to feel like you aren’t good at your job when these situations necessarily occur.
And the real key to applying the probabilistic approach here is to understand those factors you can somewhat control (e.g., belief in the company strategy) and those you can’t (e.g., events in people’s personal lives that inevitably end up leaking into their work life). For the former, you need to understand what parts of the environment you can control to lessen their probability. But in preparation for the latter, you should have a clear process and plan for recovery when they inevitably do happen. You must know what you’ll do if any important person from your team leaves (and your bus factor for anything critical should never dip below two).
In a probabilistic world, you don’t judge yourself based on bad events happening, but based on your ability to lower their probability of happening. You don’t think in terms of individual cases, but in terms of scale, of process, of culture, of org building. And you aren’t concerned with what you do before the fact, but with how well prepared you and your teams are at handling the inevitable ensuing difficulties — in as systematic a way as possible — all without forgetting the very human and very personal side of these events to your team members.