For almost 10 years, I have been a commissioner of a fantasy football league. Participating in fantasy football is not only fun, it’s a model in thinking probabilistically. Each week, our host website automatically provides tons of summary statistics on each player in the league, with one click access to videos and articles providing expert commentary, and the odds associated with each player of scoring a predicted amount of points. The data and the probabilities evolve through the week, providing new information up to the moment that the actual games begin.
This provides a parallel to a situation that brand marketeers encounter each day. Every day we are thinking about how to drive growth by conditioning demand in the market for our brand(s) over the other brands that we compete with. To do this we look at lots of data on a variety of influences on consumer behaviours. We bring together data on marketing spend; quality of our communications; consumer associations, feelings, and memories; price and distribution; macro-economic… It’s a wide list, and now includes epidemiological and public policy data to account for the influence of the pandemic. But are we effectively leveraging this information to make good decisions? Continuing the sports analogy, are we re-setting the odds of our team winning or covering the spread as we approach game day? Are brand marketeers thinking and acting probabilistically in how we use data and analytics to inform our plans to drive growth?
Here are three guidelines to push us to move in this direction:
First, we need to ensure that all data, analytics, insights and recommendations are easily distributed, accessible, and ready to be acted upon in an on-demand fashion. This need for technology-enabled data strategy and engineering is not new, but it is greatly increased with the changes brought about by COVID-19 that cause business and consumer realities to change so quickly. The drivers of consumer choices summarised before must be managed and updated and visualised on a continuous basis. This data is, in effect, the inputs and outputs that allow us to engage in probabilistic thinking, but that won’t work without ease of access and interpretation. And it should be unacceptable that marketing executives often have less access and relevant information to make decisions than Fantasy Football enthusiasts have regarding which players to start on any given Sunday.
Second, we need to do something that may be very uncomfortable. We need to embrace and embed uncertainty within our decision-making processes, rather than running away from it. As the mathematician John Allen Paulos wrote, “Uncertainty is the only certainty there is, and knowing how to live with insecurity is the only security.” Too often we fall into the trap of believing that because we have gathered all the relevant data, curated it in an analytics-ready fashion, that we can then generate definitive answers of predictions from the models we run on how best to move forward. And this desire for clarity of direction is heightened in our current moment of accelerated change with many more unknown and uncontrollable influences. But the volatility associated with our inputs and outputs in a COVID-19 environment means that there is greater volatility associated with the odds of realising any predictions we make. Said differently, we should approach our analytic results with humility, and we should ask ourselves: how confident are we that we are explaining the moment in our analyses? Are we actually explaining a longer-term trend?
Last, and this is the exciting part, we should always be thinking creatively about scenario planning, simulation, and potential paths to growth that we have tried many times before through the completely new and unexpected. Back to the sports analogy, there are odds of success associated with each play or formation a coach chooses in a game – and those odds change based on the plays and formations of the opposing team, as well as the time left in the game. It’s thrilling when we see a coach choose a play that no one expected. Brand marketing needs to have a similar mindset; we can use data and analytics to shine a light into paths to growth yet to be chosen, paths that are not expected, by applying our models against simulated outcomes. There will be more uncertainty around these paths, but that is a not a reason to not consider them. And with all the options for agile experimentation, we can quickly gauge what is likely to stick or not, make changes, adapt our probabilities about potential success, just as the coach needs to do through the game.
