Everything is ok, until suddenly it isn’t

OK, so I’m kind of a geek, accepted, right?

Here’s a mechanism from complexity science for understanding how sudden catastrophic failures happen in systems. It explains both the machine and human elements quite neatly. This science was originally intended for things like physics, population studies and financial markets, but has been successfully applied in many businesses to avoid productivity loss through breakdowns. The science/math behind it is called “attractor landscapes”. I first came across this in physics when studying how systems ‘trend’ to a particular state (equilibrium in thermodynamics).

Told ya I’m a geek.

Ernest Hemingway was once asked how he went bankrupt. His reply was “gradually, then suddenly”. It’s a great quote. Often things seem ‘stuck’, until suddenly they change.

Let’s look at a simple example to see how attractors work:

• My car leaks a little oil……
• I put oil in my car regularly to lubricate the engine — too much oil and I might do damage to the engine, too little oil and the engine seizes
• As long as I stay within a given range of oil levels (normal operating range) all is well and the car runs fine

We can show this on the graph below:

As long as I stay in the ‘normal operating range’ I have no problems. A little more, or a little less makes no difference. The system behaves like a ball rolling down a series of hills and valleys. The top of the hills are the ‘tipping points’. If we remain within the valley of the normal range, then all is well. The ball always wants to roll to the bottom of the valley — it’s equilibrium point, which keeps the system stable and functioning correctly.

But, you can see that there is a second attractor or state of equilibrium. This represents another state that the system moves to if we pass the tipping point at the top of the hill.

So long as I add oil to the engine before I reach the tipping point, the system will stabilize and return to equilibrium at the bottom of the valley, even outside the normal operating range. But if I don’t add oil before the tipping point, the system quickly and unexpectedly changes (a WTF! moment) and moves to the next stable state — a seized engine (catastrophic failure). Everything is OK, until suddenly it isn’t. Once I cross the ‘tipping point’, there is no going back, the system will always move to the next stable state, which in this case is a knackered car!

If you’ve ever experienced a WTF! moment, then you’ve just passed a tipping point and something unexpected has occurred.

There are a couple of things to take away from this:

1. Just because a system has been characterised by stability in the past, does not predict that the system is safe from failure in the future. Maybe we just haven’t reached the tipping point yet.
2. These failures can be human (behavioural) or mechanical. There are plenty of times machines have failed with no apparent warning. There are plenty of examples of people getting away with risky behaviour, until the day they didn’t.
3. For systems (processes) where the cost of passing a tipping point can be catastrophic, we need to have a pretty good understanding of the system and it’s tipping point(s), and create wide margins of error.

The actions to consider:

• finding the processes in your business that are vulnerable to tipping points,
• understanding the conditions that give rise to volatility for these processes and –
• defining robust limits and checks for them.
• Visualising early trends is the key.