Lessons in Embracing Messiness
Managers don’t solve simple, isolated problems; they manage messes.
-Russell L. Ackoff
By education and hobbies, I am an electronics engineer. Most problems I face in this domain can be solved through the application of mathematical principles and known solutions. Complex problems can be decomposed into multiple simpler problems and solved in the same way, so called reductive thinking.
Don’t you wish that many of our leadership and technology issues were like this? Instead they are often messy, continually changing due to multiple interdependencies and contradictory or incomplete information. This is why reductive, simple thinking that sets out to determine the solution cannot solve complex issues such as climate change or financial and social inequalities. This is also seen in the reality of most companies’ strategies which themselves are emergent, developing iteratively as a company learns more about its environment, and having to adapt to changes in the environment that happen because of the company’s own actions.
So why do we apply reductive thinking religiously to many complex business problems? I believe it’s because there is something seductively simple and assertive about coming up with the answer. I’ve seen this manifested many times. For instance, the idea of a multinational company doing something globally once can be simplistically justified as better than doing it multiple times. This thinking fails to deeply consider complex aspects such as agility, country differences, political capital required, or vendor capabilities, or even the changes the proposed strategy itself will have on the environment it intends to address. Similarly, arguments are made for being completely abstracted from any particular cloud at great cost to reduce perceived risks, a notion dispelled in Thomas Blood’s blog post.
The relationship between technologies and people is becoming increasingly interwoven, exasperating this issue and leading to significant programme failures. So how can you deal with it? I’d like to share four mechanisms that I have seen succeed.
Differentiate Between Complicated Issues and Messy Ones
Time invested in really understanding a problem or opportunity is rarely wasted and yet in our always-on world, we often only graze the surface of a problem. In our rush to find solutions, we often mistake messy issues with complicated issues, with the latter able to be solved through decomposition into multiple simpler problems, the former not. I have two favourite tools to help here:
- Systems mapping can help define the multiple stakeholders and variables involved in a problem and their interrelationships. Done with the help of a professional facilitator and a diverse set of thinkers from your organisation, this exercise unfailingly identifies new assumptions and dependencies.
- De Bono’s “Six Thinking Hats,” which forces each participant to take different perspectives on the same challenge.
Both techniques help to ground teams with the same knowledge and appreciation of a problem, allowing them to design appropriate interventions to improve the situation being analysed.
Curiosity Didn’t Kill the Cat
Be inquisitive. The allure of throwaway phrases — digital transformation, data-enabled, AI-enabled, and agile, to name a few — often hides our lack of understanding of their true meaning, and the consequential lack of useful action that results. The best leaders I have met ask incredibly insightful questions on topics where others may want to hide their own lack of understanding. The ability to ask good questions and to learn should be a core consideration when selecting leaders as solely looking at past experience is one of the poorer indicators of future success.
To be inquisitive requires a humbleness that Mark Schwartz recently wrote about. The Amazon leadership principle “Leaders are right a lot” illustrates this. The principle does not encourage leadership omnipotence; rather, it exhorts that leaders should “seek diverse perspectives and work to disconfirm their beliefs.” How cool is it to encourage leaders to learn and be open to being wrong so that ultimately their judgment calls are more often right?
Define Hypotheses, Not Requirements
When facing messy problems, state your beliefs about the solution as hypotheses to be tested, rather than requirements to be coded. With a requirement, we assert that we have knowledge of the solution, and when it doesn’t work as planned, we inevitably debate about whether the requirement was correctly specified or coded. With a hypothesis, we openly declare that we believe we know what is required, but that it needs to be proven, or not, by experiments approached with a humble learning mindset. Interestingly, this is at the heart of many companies’ struggles with becoming data-enabled. People are often wired to look for proof that their position is correct rather than searching for a truth that might contradict their beliefs. Particularly for those interested in machine learning, it is essential to start from a (null) hypothesis that you’re comfortable disproving rather than the often-used approach of trying to prove, at all costs, why a new idea is right.
Consider Future Options over Just a Now-Oriented Solution
Our ability to predict the future is notoriously bad. In a world that often equates leadership with decisiveness, we are often pushed to come up with the answer quickly. As we’ve seen repeated throughout history, this belief closes off options that might prove more suitable once more is learned about the problem, and prevents people from searching for another solution through confirmation bias.
To steal from the financial and natural resources industries as well as the military, preferred methods of solving or satisficing complex problems include scenario planning and choosing options that will enable multiple avenues to be pursued in the future rather than just one. This is particularly important in emerging areas such machine learning where technology continues to evolve rapidly, and where the application of the technology will have profound known and unknown implications on society, businesses, and individuals. Although this platform approach does not suit every situation, when it is used, it helps support future possibilities that have yet to be considered.
All of this is enabled by the cloud and modern organisational philosophies such as the agile organisation. The combination of the two enables organisations to experiment and learn, to adapt rapidly to changes in the environment, and to tap into multiple technology options cost-effectively, deferring decisions until they are truly needed and informed. It is an uncomfortable method for us leaders taught to seek answers, but a practical one that keeps our minds — and systems — open and agile.
Unlike what you might tell your children, embrace your messiness!
Superforecasting: The Art and Science of Prediction, Philip Tetlock