A little thread about resiliency and efficiency. This might become a blog post; right now I’m thinking aloud.

Imagine an airline, back when plane travel was safe.

Running an airline is expensive, so management tells everyone to tighten their belts in the name of efficiency.

The bean counter looks at the planes. They have two engines, but you only need one for most flights. Building and maintaining engines is expensive.

“What if we built planes that only had a single engine”, they ask the mechanic.

Would you fly on an airline with this attitude?

The bean counter looks at the cockpit. Each flight has two pilots, but either one can fly it all. Hiring and training pilots is expensive.

“What if we ran flights that only had a single pilot”, they ask the flight instructor.

Would you fly on an airline with this attitude?

The bean counter looks at the takeoff checklists. Each plane is checked before it flies, but usually the checks don’t find anything.

“What if we skipped doing some of the checks”, they ask the ground crew.

Would you fly on an airline with this attitude?

The bean counter has a point: all of these things would make the airline more cost-efficient!

They’d also make air travel safe, and less resilient to failures in a flight. A previously-manageable incident would become catastrophic.

Taking efficiency to the extreme is dangerous.

The airline could hedge against dual-engine failure by fitting each plane with four engines.

Less efficient, more resilient – but eventually they’d run out of money, and then they don’t get to fly any planes at all.

Taking resiliency to the extreme is dangerous as well.

For most systems, the sweet spot is somewhere in the middle – you want a balance of efficiency and resiliency.

They’re inherently opposed – more resilience means less efficiency, and vice versa.

If you only ever prioritise one and not the other, bad things will happen.

And of course, people always want to prioritise “efficiency”, which is usually a weaselly way of saying they want to cut costs.

The same people then act surprised when there’s no slack in the system, and it falls over at the first sign of a crisis.

You can’t make a career of crying “efficiency! austerity! cut costs!” at key infrastructure, then act surprised that it can’t cope in a major crisis.

Cutting spare capacity might look like an easy way to save money, but it’s a gamble and it can come back to bite you later.

If you’re missing the subtweet:

The reason the NHS and other healthcare services are struggling right now is because of years of calls to cut costs and improve efficiency.

That leaves very little slack in the system to cope with a sudden surge like, say, a global pandemic.

This follows a previous thread about resilience/efficiency: https://twitter.com/alexwlchan/status/1300000406349713411

I have lots of thoughts about this that I’m still organising, so I’ll leave it here for now.

tl;dr: efficiency and resiliency are connected, and you can’t change one without affecting the other.