Coping with complexity

Have you ever seen a product launch fall flat, or a high-scoring ad campaign tank and wondered why – only to be told that “it’s complicated”? At the time that might not have seemed like a satisfactory explanation, but it is true.

We live in a world of interlocking systems, some of which are ‘simple’ in the sense that we can model them and make accurate predictions based on those models, while others are ‘complex’ for which the above just isn’t possible.

“Nobody really agrees on what makes a complex system ‘complex’ but it’s generally accepted that complexity arises out of many interdependent components interacting in non-linear ways,” writes Duncan Watts in his book, Everything is Obvious Once You Know the Answer.

Of course, simple systems can be highly sophisticated – the movement of the planets, the handling of a car or the aerodynamics of an aircraft – but what’s important is that we are able to accurately model the system. If we can model the variance in a system, then we can predict the activity of that system.

Complex systems, by contrast, are inherently unpredictable as we cannot account for a high proportion of the variance that influences outcomes. In these circumstances the classic ‘butterfly effect’ can occur, whereby small changes in one part of the system are amplified to create large disruptive effects elsewhere.

Network of choice
Much social and economic activity is complex. Within market research we recognise that much of what we are dealing with is ‘multi-faceted and non-linear phenomenon’¹. Complex systems, therefore, help to explain why so many product launches fail and why some marketing campaigns are less successful than others.

But increasingly there are ways of mapping these complex systems. Network science is a huge field that has grown rapidly in the past decade. During this time numerous types of network have been identified and we are starting to see the emergence of three core ‘species’² that are useful in explaining social and economic behaviours and attitudes. These are:

• Scale-free networks – where most people have a small number of connections but a few have a large number of connections. This can be represented graphically as a ‘hub and spokes’. The people in the hubs may or may not be particularly influential, but the large number of connections that they have is important in driving change
• Small-world network³ – which can be described as overlapping groups of friends of friends. These are groups that are typically closely connected, but then a few people within those group have longer-range connections to other groups. In order to drive change it is important to identify these ‘long-range connectors’
• Random networks – groups where there are no consistent patterns in the way in which members are connected. They are a purely random assortment of connections. The common cold, for example, is transmitted through this type of network

Common to all networks is that they are ‘robust’ – that is, for much of the time significant change does not happen. Think of attempts by brands and their marketing agencies to generate viral campaigns. Most of the time they simply do not succeed. However, there are occasionally very big changes that occur where significant numbers of consumers within the network ‘copy’ each other. It is at these times that we describe the network as ‘fragile’, and at these ‘fragile’ moments the success of a proposition is less to do with the fine tuning of the features, pricing etc. and more to do with the way in which copying has spread across the network.

The nimble-footed facilitator
Identifying the types of networks we are dealing with can help us to understand complex systems and, as such, better manage activity within those systems. Networks provide marketers with a completely new way to manage change – but it’s important to remember that we still have limited control over complex systems and there remains limited predictability.

Systematic approaches are therefore needed: using research to check progress and then amending marketing strategy as appropriate. Careful allocation of resources which are used intelligently over time is the nature of the challenge in a complex system – the antithesis of launching large-scale, one-hit campaigns that merely hope for the best.

It is likely that we will start to see the end of the more definitive ‘predict and control’ approach to marketing, moving to a more nuanced approach where we map out the nature of the network environment, assess the nature of the market sentiment that is resonating within these environments and then carefully experiment to assess which of a range of different propositions will resonate.

If we accept the premise that we live in a world full of complex systems then the way we look at each stage of the marketing cycle changes. Optimising to meet the needs of specific market segments is still important but managing the underlying network becomes critical. Market research needs to become even more of a nimble-footed facilitator, constantly checking the structure and status of a network and acting alongside marketing teams in a highly iterative and fluid way to work out how best to achieve success.

Colin Strong is managing director of GfK’s technology division in the UK. He’s previously written about smart data and social influence.

Reference:

1. David G. Bakken ( 2005 ) The Bayesian revolution in marketing research. Esomar Innovate Conference, paris 2005
2. Paul Omerod, ( 2012 ) Positive Linking: How Networks Can Revolutionise the World. Faber and Faber
3. Watts, D.J.; Strogatz, S.H. ( 1998 ). “Collective dynamics of ‘small-world’ networks”. Nature 393 ( 6684 ): 440-442