If our brains were simple enough for us to understand them, we’d be so simple that we couldn’t.
(Ian Stewart: The Collapse of Chaos: Discovering Simplicity in a Complex World)
I am talking about computer modelling, specifically Agent-Based Modelling (ABM). I believe this is a skill that most can master and that it can be a powerful tool in understanding complexity. In a world where the need for digital skills seems to be increasing exponentially, we are starting to realise that, in fact, we cannot keep up with everything. In fact many digital skills will be handed over to bots in the not-to-distant future. This will leave opportunities for humans in managing technology and relationships; something AI is a long way from. The message is to keep calm and be selective about the computer skills you really need. I am arguing that ABM could be on that list.
But what is ABM, and why is it important. What does it have to do with leadership? ABM is a surprisingly accessible method for simulating the social world as well as other uses. The agent is a simulated entity which can act autonomously; given a set of characteristics or available actions. For example, the agent could be vendor selling items to other agents, or a business in a specified business environment. What makes ABM really instructive is that once agents are programmed and the environment specified, they will generate life-like complexity out of a few simple rules. It can be incredibly surprising and educational how a simple programme can develop in unexpected ways using apparently quite predictable scenarios.
ABM has been growing rapidly over the last 20 years, mainly in the scientific and academic fields. However, easy-to-use platforms such as Netlogo are making ABM accessible to all. Netlogo is a simple to learn programming language using a graphic interface which makes developing simulations easy. A good start is to take a look at the Netlogo website library (see here) of different models, play with the parameters and run the simulations. Although the models are largely scientific, it is easy to see how models could be developed for business. Uri Wilensky, developer of Netlogo, argues that ABM literacy should be strived for universally, including professionals (Wilensky and Rand 2015). Apart from anything else, many of the models are fun. For example, there is the famous wolf-sheep predation model. By adjusting the growth rate of the grass, or the number of wolves the delicate ecological balance shifts one way or the other, resulting in a world of grass, or starving wolves. It is easy to visualise how delicate these complex systems are when you see the direct results of tinkering with the parameters. Ecological models have clear read across to organisational studies and business strategy; it would be relatively straight forward to write a model for, say, survival rates of digital start-ups.
Adjusting the parameters only slightly leads to a world of green, white or black.
The clever bit about ABM is the astonishing computational power of the models underneath the apparently simple interfaces and visual presentation. The ability to enact simple rules at various levels within the model causes unexpected phenomena to emerge, just as they do in real life. Emergence is a key factor in any complex model and is the thing that catches leaders out the most. When simple factors combine and interact at multiple levels the emergence of new complex and unforeseen features occurs. So running simulations of the effects of business decisions makes sense. Not because the ABM will copy exactly the business environment but because it can highlight possibilities leaders may not have envisaged otherwise.
Leadership itself is an emergent property. Real leadership takes time, and it is based on relationships rather than individuals. This is why a successful leader can bomb when placed in a completely new team. Seeing leadership in this way is easy with ABM. As a part of my leadership research recently, I created an agent-based model simulating how information passed through social channels based on the prestige of actors in the group. This was a real-world effect I noticed during my data collection. Running the simulation on R, a statistical programming language, I could see straight away how leadership existed in complexes of prestigious individuals, rather than on specific individuals. By changing the conditions of the model I was able to demonstrate that non-hierarchical models can exhibit higher levels of leadership, a key message for the digital world.
A simulation of a real world social network based on military teams. Larger nodes are prestigious individuals. There is no clear single leader but a collaborative hub of prestigious individuals of differing ranks. Of course there is a single hierarchical leader but the model demonstrates how important collaboration is even in a strict hierarchy.
R takes several months, even years to get to grips with but with platforms like Netlogo, it is possible to build basic models in hours. It would take a few days to learn to write complete models, but well worth the investment when you think how this powerful tool could be used in strategy development or for teaching leaders about strategic decision making.