As Daniel Kahnemann has pointed out to us, while the human brain is a remarkable organ, it also has its limitations. One of these is that it is built to conserve energy. What that in turn means is that we readily engage in fast, linear thinking when the challenges we face are complex. Forcing yourself to think differently is the key to untangling complex situations.
I’ve been thinking lately about the challenges facing leaders in situations that are fundamentally complex, and how these might lead us to problem-solve differently. Complex systems are characterised by multiplicity, interconnectivity and dependency. Practically, what that means is that a complicated system (like a Boeing 777, for instance) can be designed to achieve predictable outcomes. A complex one, however (like the air traffic control system) has parts that interact, making predicting an outcome impossible. How then, to tackle challenges raised by complexity?
Extreme outcomes rather than expected ones
Rationally calculating likely outcomes often leads us to think of things lying along a bell shaped curve, with the most extreme possibilities the least likely.
In complex systems, however, executives need to consider the possibility of extreme outcomes rather than expected outcomes. These can be positive (Google creates an entire category of applications and grows astronomically) or negative (Northern Rock, Bear Stearns and Merrill Lynch – once storied brand names – disappear). In either case, thinking in terms of probabilities and averages is not useful.
In complex systems the most interesting phenomena often occur at the tails of distributions, rather than in their centres. Thus, the successful entrepreneur is not the norm; nor is the successful corporate innovation; successful drug discovery or successful R&D project. This in turn means that extrapolating from past trends is a poor predictor of what will be of most interest in the future. Who knew that inventing a way to trade bales of hay in a game called Farmville would lead to one of the hottest IPO’s of the early 2010’s? Or that the now-extinct game would teach software makers about the addictive potential of social networks and the value of the data users would willingly give away.
Non-linear and path dependent
Complex systems are also non-linear and path dependent. What that means in English is that there is not proportionality between the investment in an element of the system and its effect, and that these effects accumulate over time. Some scholars call these “tiny initiating events” and they can have a disproportionate impact (this is sometimes called the ‘butterfly effect’).
For example, beginning with simple tools such as the Flip video camera (remember that one?) and moving to today’s smart phones, digital video helped establish the success of video sharing site YouTube by making it easy for people to post and share their videos. YouTube in turn has had a huge impact on society, politics and culture by becoming a forum for the posting of everything from the silly (cute kittens doing ridiculous things) to the serious (widespread disinformation and the viral circulation of videos of horrible events).
Systems change in emergent ways
The outcome of a business decision can’t be predicted in a complex environment, even if one could take into account all known variables at the time a decision is made.
Consider competition in social networking and the advertising-supported Internet we have today. Big winners include Google (worth a mind-blowing $1.53 Trillion); Facebook, or rather Meta (worth $458.69 Billion) and Twitter ($29.9 billion, depending on what kind of day Elon Musk is having).
We’ve forgotten all about many of the losers, though. Remember Myspace, for which News Corp paid $580 million in 2005, only to find itself practically giving the site away for $35 million just a few short years later. Or even worse, Bebo, the social site AOL paid $850 million for in 2008, deciding two years later that it was a ‘distraction’ and basically giving it away. Or eBay’s acquisition and subsequent divestiture of telecommunications facilitator Skype for $2.5 Billion. Or what about Orkut, a social media platform Google operated until, well, it didn’t. In each of these cases, the outcome of greatest interest lay at the extremes, small initial decisions had a disproportionate effect on what happened later, the successes are outliers and it was quite difficult to predict how the systems would evolve.
Complex system behaviour can cause entire business models to cease being viable, which can take people by surprise. Here’s a simple example: executive recruitment. Conventional head-hunting depended on the following elements being difficult or expensive to bring together: knowledge about which individuals hold which jobs within companies; knowledge about who might be amenable to an offer; and knowledge about salary levels and employment conditions in the target companies. The value contributed by the search firm was to link these elements together and present employers with viable candidates they may not have been able to identify otherwise. Today, business-oriented social networking sites such as LinkedIn and Glassdoor.com allow prospective employers and prospects to connect without the help of a broker. This shift is forcing headhunters to invent whole new ways of creating value.
And it’s catching up with today’s decision makers, who face environments in which things that used to be kept safely apart are bumping up against each other, often with unexpected results.
Making the wrong choices in a complex environment can lead to negative, often unforeseen, consequences. For example, retailers today find themselves competing not with traditional retailers but with so-called “direct to consumer” firms that develop deep relationships with their customers.
Making the right choices in a situation that is complex can also be fairly rewarding, as Ron Adner points out in his great book “Winning the Right Game.” As he describes it, Amazon adroitly navigated the emerging market for voice-enabled technologies to make its Echo technology a standard across many different kinds of devices.
Slack, discovery and learning
One implication of all this is that our conventional ways of investigating phenomena and understanding the world around us are highly unlikely to be the best way to put together an understanding of what data are relevant and what decisions should be the best. Ironically, complex systems may best be understood by qualitative and speculative practices rather than by ‘hard core’ analytics.
To get this right suggests the need for some slack in the system, time to explore around what I’ve called the “edges” of the organisation and diverse talents. This is not just a nice to have – it can be absolutely essential.
Consider the following hypothetical. You have people with qualities who are able to respond to some kind of challenge from the environment. Let’s say each employee has 5 such qualities. Have a look at this table.
What becomes clear is that many of the team members selected for this group share a common set of qualities, which is often what hiring decisions in companies are like.
Note the potential blind spot this creates, however. Only Chantelle has the qualities needed to respond to challenge number 4, and only Isaac to the one presented in condition 6. It becomes quickly evident that losing either of them depletes a team of valuable resources.