Nothing But Net: Bayesian Network Modeling and the NBA Finals

Posted by: Charles Swann
  • June 14, 2016
  • 1
Photo Credit: Cavsnation.com

Photo Credit: Cavsnation.com

If you have seen any news coverage about this year’s NBA final between the Golden State Warriors and the Cleveland Cavaliers you may think it is a simple battle of titans: the Warriors’ Steph Curry versus the Cavaliers’ Lebron James. The sports commentators frequently say things like “It’s all on Lebron James’ shoulders tonight. If he has a break-out performance, then the Cavaliers will win this game.” However, for those of us who have been watching the finals, the games are much more than the Steph & Lebron show. In fact, several games have been decided by amazing performances from “support” players who work together to outscore their team’s titan.

So what can marketers and researchers learn from this? And what does this have to do with drivers analysis?

Just as sports commentators tend to look at players’ individual impact (and focus on the players with the biggest individual impacts), traditional drivers analysis looks at the individual impact of attributes and encourages us marketers to focus solely on the strongest drivers.

The simplistic one-player-at-a-time approach, while it may be right at times, is more often misleading because it doesn’t look at the interactions between the players. When Steph Curry isn’t playing his best, the players around him can respond. What’s more, when those players work together they form a team that is greater than the sum of their individual contributions. The traditional approach to drivers would fail to explain how the Warriors won the first few games because it doesn’t value the interactions of the other players. It also leads marketers to focus solely on the strongest drivers, even when they aren’t the true key to market success.

To help marketers develop a more sophisticated strategy we have been using a technique called Bayesian Network Modeling. While traditional approaches look at players’ individual impacts, Bayesian Network Modeling looks at those individual impacts and all of the interactions between the players. This means we can answer questions like “What combination of players are most effective to use when Cleveland is playing small ball and Steph Curry is on the bench?” Or, more importantly for marketers, “should I focus on the biggest driver (and compete head-to-head with everyone else) or is there a collection of brand attributes that, on their own are not the most important, but when taken together have a huge impact on customers?”

It’s unlikely that sports commentators will stop focusing on Steph Curry and Lebron James—after all these two players can (and have) single-handedly carried their teams to a win. However, I do believe that marketers will become much better at creating winning strategies when we look beyond the attributes with the biggest individual impact and understand how all of our consumers’ attitudes and behaviors interact to drive their decisions.

Categories: Marketing, Modeling
1 Comment
  • Bethany
    June 14, 2016
    So true, Charles! Just look at last night's game 5, where Draymond was out--totally changed the dynamics of the players.

Add Your Comment