August 02

Leadership hack 027 – as a leader you need to change when the problem changes

We all know there are different types of problems.  Noble prize winners in physics solve different problems to firefighters rescuing people from a burning building.  David Snowdon attempted to categorise problems into four types of simple, complicated, complex and chaotic.

Leaders need to adjust their approach based on the problem they face.  The first thing to do is to understand the problem.  This is more difficult than it sounds.  There are three common traps:

  • Relying on your experience and intuition – ‘I have seen this problem before’
  • I do not understand it, the problem must be chaotic
  • Assuming that the problem will always be the same

These traps result in miss-categorisation, and then a leader applies the sub-optimal approach.  Here are some ways to prevent these traps:

  • If you have time, stop and analyse and if possible experiment
  • If you don’t have time, try to limit yourself to decisions that could be reversed
  • Find someone who has a lot of experience in the specific problem and learn from them
  • Keep analysing the problem, make sure you can identify when it changes


Here is more information on each problem type, how to identify them, and how to respond.


Simple – the domain of ‘best practice’

Simple environments are deterministic – there is a clear, simple and obvious link between cause and effect.  For example, a ball rolling off a table on earth will drop to the floor.   Under these conditions, leaders can learn very quickly by using “sense–categorize–respond”.  Sense what is going on (ball fell off the table), categorise what happened (gravity) and then respond (turn the ball into a square so that it cannot roll).

Leaders need to be careful and not oversimplify problems just so they can use this approach.   Leaders also need to ensure that they do not become complacent, they can do this by having a clear channel of communication between the bottom and the top.

Complicated – the domain of ‘experts’

A Rolls Royce Trent 1000 aircraft engine is incredibly complex.  Despite it being made of thousands of parts, skilled engineers can take it apart and put it back together the same way a hundred times.   When facing complicated problems, leaders should “sense–analyze–respond”.  If a Trent engine suffered a failure, you need to sense the mechanism of failure (e.g., birdstrike), analyse the internal (components) and external (weather) factors so that you can respond (get better data on bird flight paths).

Leaders need to create an environment that allows experts to solve these problems.  Good access to data, an environment where experts can disagree safely and time are essential to solving complicated problems.

Complex – the domain of ‘experimentation’

The stock market is complex.  One day a rise in oil price will coincide with a stock market rise, on another day the market could be flat or go down (see here for an analysis).  Complex systems have so many dependencies and variables that it is very difficult to determine cause and effect and can usually only been determined in retrospect.  In a complex environment, leaders need to “probe–sense–respond”.  A stock trader may run simulations of possible strategies before they execute them for real, or a leader of a software company may launch multiple MVPs into to see what customers value.

Leaders have to be very careful about taking a ‘reductionist approach’, that is taking apart all of the components parts to see how they work.  For example, you will not understand the traffic by analysing how a car is put together.   Leaders need to allow people and teams to conduct experiments (with clear metrics and goals).  People and teams need to be able to take risks and for well planned and executed experiments to fail.

Chaotic – the domain of ‘action’

A firefighter rescuing people from a burning building operates in a chaotic environment.  The fire could do anything, and there could be damage and dangers you cannot see.  Chaotic environments are too confusing and urgent to analyse in detail or run experiments.   Leaders must use “act–sense–respond”.  Enter the burning building to gain more information, then sense if it is safe enough to proceed, then respond by either exciting to safety or going further in.


I have found the table below (copied from an HBR article here) a great synthesis of Snowdon’s work.

Image result for simple complex complicated chaos


You can find more information at David Snowdon’s webpage, his video, or on Wikipedia, the title image by Edwin Stoop.