Should I challenge my own futures training?
I was brought up in the school of strategic thinking built upon the fundamental assertion that the future is too complex and uncertain to be able to predict. I was trained in scenario thinking by the founders of Global Business Network (GBN), many of whom had been instrumental in successfully propagating the use of scenarios at Royal Dutch Shell. The ground-breaking Shell Scenarios Team, was itself established by the mercurial Frenchman, Pierre Wack, back in the 1970’s.
I once made the mistake of saying to a client, in front of Napier Collyns (one of the aforementioned Shell executives and a co-founder of GBN), that while a powerful tool, there was nothing magical about scenario planning. Napier visibly balked at this suggestion. He of course had be heavily influenced by Wack, a man who studied Sufi mysticism and would disappear for weeks on end to India to meditate and contemplate the future. Pierre believed strongly in the need to combine spiritually heightened awareness with an intense analysis of real world facts and the vast range of possible futures there presented. For Pierre, anticipating the future involved “being in the right state of focus to put your finger unerringly on the key facts or insights that unlock or open understanding”.
Note, futures (plural) not future (singular). For Shell (GBN and all the scenario methodologies that followed) developed multiple futures, or scenarios, to help better understand what might be (what if?) before contemplating their respective implications (so what?). I have been struck recently by how may articles and claims have been made that; yes, humans were once bad at making predictions, but now we are getting much better. Data scientists will point to the breakthroughs offered by the collation and crunching of big-data. Although for every success story I have read, I have seen just as many failures. As Tim Harford of the FT (incidentally also a former Shell scenario planner) pointed out back in March, “‘Big data’ has arrived, but big insights have not.”
Has the Era of the “Superforecaster” Arrived?
Interestingly, Tim has just penned another article on “superforecasting“, published in the FT Weekend Magazine, this weekend. Harford reviews the work of Canadian-born psychologist, Philip Tetlock and his establishment of the Good Judgement Project, an attempt to improve the success of geopolitical predictions for the benefit of the US intelligence community. Considerable success is claimed, in extracting insights and predictions from a 20,000 strong network of forecasters largely on the back of:
1. Basic training for those forecasters in probabilistic reasoning
2. Leveraging the “wisdom of crowds” phenomenon, i.e. teams of good forecasters produce better results that good forecasters working alone
3. Having an actively open mind, increases your chance of making a successful prediction.
Wikistrat has also received a strong press in recent days, based on their own success in using the wisdom of crowds to outperform traditional geopolitical forecasting. Although, in their case they use a small, carefully hand-picked network.
I suspect that Pierre Wack (sadly no longer with us) and Napier Collyns would have no issue with the last two of Tetlock’s success factors., In fact quite the reverse, scenario planning actively promotes the need for diversity of input, and openness of mind. The clue to the original success of GBN was in the name – Network. Collyns and his co-founders built an extraordinary network of so called Remarkable People, based on Wack’s practice, to provide the very challenge and diversity of input that Tetlock and Wikistrat seek.
Probability vs. Plausibility
The thorny issue remains however, should we take a single future view (prediction) or carefully selected, divergent, multiple perspectives (scenarios) to better understand was is to come. This perhaps touches on the quasi religious debate about the “2 Ps” – probability vs. plausibility. Forecasters focus on the former, scenarists typically focus on the latter. My own view is that it has to be “horses for courses”. The development of predictions and scenarios , while related, should be used for very different purposes in decision-making and setting strategy /policy. Scenarios, while data-driven, are essentially narratives, stories if you will, about the future. They help inform the intuition of decision-makers to understand the causes and context of what might happen and explore the critical implications for their organisations, before determining on a course of action. Data, whether historical, or predictive, should be part of the mix and it is the balanced integration with future narratives that enable the creation of robust decision-making in light of future uncertainty.
Comments