The company created a system that uses custom analytic models to predict whether a customer up for contract renewal is likely to accept or reject a price increase. It considers factors such as type of services, geography, and past behavior, as well as competition and economic conditions. Pricing managers get the analysis, then exercise their own judgment in determining whether to push for an increase.
Waste Management needed the pricing app for several reasons. One is volume. The company provides more than 100,000 service quotes a month, and relying on the experience and instinct of managers is "not a scalable model," says CIO and senior VP Puneet Bhasin. Also, the company has started a multiyear effort to increase revenue and cut costs, and it's counting on "pricing excellence" to meet its goals (see Waste Management's New Model: Garbage In, Profits Out).
Implementing a pricing optimization system can be daunting. One reason is that price can be an emotional issue. If an airline passenger who paid $500 finds that the person next to him paid $250, the $500 traveler isn't happy.
Waste Management's services, however, are different because every bid is unique, says Bhasin. Location, type of waste, weight and volume, scrap value, rules on where garbage can be taken, and competing services all go into its calculations.
Another challenge is that there isn't an easy off-the-shelf pricing optimization system that companies can buy, Bhasin says. While pricing software is available, there's a lot of hard work in learning what factors really drive market prices and in customizing the software to the industry and market.
The company first had to build its predictive models, which it did using SAS's analytics software. "It takes a commitment to science and technology not everyone has," Bhasin says. Waste Management built different models for all 25 of its markets, since the drivers differ. IT custom-coded the models into decision support software, then established a portal to give its pricing managers access to the recommendations generated.
The system will never replace a manager's judgment. But by modeling the most important market drivers, Bhasin says, Waste Management is working to be "more intelligent in how we price our products."