Whenever I hear the term “incentives,” I instantly think of that iconic I Love Lucy chocolate factory scene. Lucy, a new factory employee, is responsible for wrapping chocolates as they come down a conveyor belt. The manager sternly instructs Lucy that if a single chocolate makes it down the belt unwrapped, she’ll be fired. From the manager’s perspective, the incentive seems logical: wrap chocolates quickly so that none will leave the belt. From Lucy’s perspective, it’s very different. As the belt speeds up to an impossible rate, Lucy meets her manager’s requirement by guzzling the onslaught of chocolates instead of wrapping them. When the manager comes back to inspect her work, unaware that Lucy has eaten and hidden the majority of the product, she praises Lucy with the punch line of “You’re doing splendidly!” and then, to someone off-screen “Speed it up!”
Advanced analytics is all about process change – changing how people do their current job to make it more efficient, reduce costs, and engage the right people. I’m sure we could brainstorm a dozen ways to approach the chocolate factory fiasco and improve that process. However, no matter the potential value of an analytic solution, unless you can get key players to buy into it and allow for appropriate changes to happen, the solution will not meet its potential value. Simply put, if we can’t get the manager engaged in finding a solution the problem, Lucy’s employment situation looks grim.
Several years back, I was working with a state department of taxation to develop predictive models to determine which individuals who owed money to the state were “collectable.” We developed models that identified people that would pay anyway, people that would pay when contacted, and people who would never pay. To me, the strategy became a very clear:
- Don’t call/contact those that will pay anyway.
- Determine the best contact strategy for each one that will pay, or pay sooner.
- Sell off the bad debt of the people with a very high propensity to never pay.
As good as it sounded, I quickly learned that there would be no process change. The department was required to treat all taxpayers with the same contact strategy. The exact same process that was in place will continue to be in place. To this day, I don’t know why we did this project.
I found a similar situation at a call center that I once tried to sell into. Our value proposition, “turn your call center into a profit center,” fell on deaf ears. While we were trying to convince the call center manager that this could make him more valuable to the organization, he knew that he and his team got paid on number of calls and call length. For him, a longer phone conversation meant less money in his pocket.
Going back to the I Love Lucy example, you’ll see a similar environment as the call center. The manager could change the incentive structure so that Lucy is paid per every chocolate wrapped, but this wouldn’t alleviate the problem of the too-fast conveyor belt. The only way to fix the inefficiency would be to slow it down to a reasonable rate. If the manager is unwilling to do this, then the majority of the chocolates would end up eaten or hidden instead of reaching customer’s hands.
Find out how companies like GE are changing business processes to get the most out of predictive analytics. And read about other considerations when planning for analytics in the Myths and Realities of Successful Analytics series.