Be a bit worse at things to get better

It has often struck me that, no matter how good an organisation is, how well managed their operations are, and how well-regarded their capability is, a decent auditor will always find something to fix. Similarly, an investigation almost never concludes that everything is fine as it is (I have done that in the past for investigations and management are very uncomfortable with accepting it as a conclusion).

Why is this? And if it’s finding issues to improve, what’s the problem?

Traditional audits and investigations look at specific components of our management systems and approaches. But those individual components form part of a bigger system and how they behave individually is not the same as their contribution to the whole. When you consider the system, the interactions between components are often more important than the components themselves.

To optimise a system, you have to have some parts that are sub-optimal to accommodate their broader role. Or, more simply, a team of stars does not make a star team.

Sporting analogies are grossly over-used in business, but it’s appropriate here. In athletics, you generally have as simple a goal as possible. If you are a sprinter, you can largely optimise for speed and not worry much about endurance, or agility. If you’re a marathon runner, you can optimise for endurance and not power. This is why sprinters are muscular, distance runners are slight, high jumpers are tall and shot putters are wide. But the best decathlete is sub-optimal in every one of the ten disciplines that they compete in. Because their goal is more complex – it’s a balance of point scoring across every event. If they optimise for power to get 100 more points in the discus, they may put on too much muscle and lose 200 points in the 1500m run.

This is even more complex and tricky within a team sport – each individual rugby player has to have good strength, good endurance, good speed and good skill. But different team positions require different combinations of those that then need to link with team-mates in a way that optimises the team performance, not that of the individuals. A team of stars does not make a star team.

Our workplaces are typically more multi-event, team based complex systems than they are simple and direct.

So, to optimise the system performance, some components will inevitably be sub-optimal. And in reverse – optimising individual components often reduces the optimality (if that’s a word? Optimalness? Optimus Prime? Yes – optimality. Googled it.) of the system.

This is, of course, context-specific and different systems have varying degrees of sub-optimal components that can be accommodated. You don’t want brakes in a car to be too far off optimal, for example. But a person in a team can be technically quite weak, if they’re very good at collaborating with other functions.

When the performances of the parts of a system, considered separately, are improved, the performance of the whole may not be (and usually is not) improved

Russell Ackoff

So, we spend time auditing, investigating and optimising at component level, only to find that our system performance is never getting any better. In fact, we may be making it worse by improving individual things.

What to do instead? There is still a place for audits and investigations, so long as they are mindful of the context of the system. I am often asked to carry out audits and performance reviews and I don’t go around telling my clients it’s a bad idea (not good for repeat business). What I do is to supplement the detailed review with a higher-level, step-back consideration of what the system is trying to achieve. I come up with a series of themes and collate my detailed findings into them before I start setting out any recommendations. This way, I can think about the connections between them. It also means not worrying too much about the little things that so many audits pick up on. It really doesn’t impact on the system performance if a document review is a little out of date. But it will deteriorate your system performance if you recommend diverting resources to check every document for currency.

You also tend to end up with fewer, more meaningful actions and I’ve never yet met anybody that complains about that. Give it a go.