Pom-poms, a Bulldozer and Donuts—and the Right Optimization Project Sponsor

Wednesday, February 21, 2018

The project sponsor recognizes the potential value of optimization and often initiates the contact with the optimization consultants to further define the right opportunity. While titles vary, this is generally a person near the top of the reporting chain of the people who will be using the software day in and out. The sponsor is usually a level or two above the SMEs: in most cases a vice president, director, or general manager.

Once the area of opportunity has been selected and the project team is up and running, the sponsor is much less involved on a daily basis than the SMEs are. The sponsor’s primary role is to make sure that the project receives the required attention and resources, as well as to make strategic calls at key decision points. Sometimes an effective sponsor needs to change his or her role as the project evolves. As one sponsor told me only half in jest, “I have been known to show up with pom-poms to cheer the team on, with a bulldozer to clear the way, and with free donuts to get the right people to show up for meetings.”

When Marriott’s Sharon Hormby and her team first put forward a proposal to develop a Group Price Optimizer (GPO), a wide circle of people needed convincing. The majority of Marriott’s properties are actually owned by independent property-investment groups who contract with Marriott International to manage their properties. To implement the optimization system, Marriott would need to gain permission from these owners, a task made more complicated by the fact that Marriott would need to charge them for the development and deployment of the system. Fortunately, Hormby and her team had strong project sponsor to support them: Bruce Hoffmeister, senior vice president of global sales and marketing information technology [now Global CIO] for Marriott International.

Hoffmeister came out of finance, making him exceptionally well prepared to speak about money issues, the value of the project, and key project milestones: in short, to help make the business case. In Hormby’s view, he was an ideal sponsor. “Bruce let us develop the problem definition as we wanted to. He is very logical, so if you gave him a good argument he would go with you. More important, he was extremely useful in helping convince the owners to let us roll out the project. Bruce actually came with us on field trips to different hotels to speak with the owners about the implementation and its value. He was invaluable.”

At ArgentAir, the optimization team had a very strong sponsor: the executive vice president of scheduling and logistics. What made this individual so unique was his combination of business and technical skills. Although head of operations when the optimization project was launched, he had previously held the position of CIO. He had actually designed and implemented ArgentAir’s computerized operations system, which our optimization software needed to interface with. If all that weren’t enough, he had an Operations Research background. With such strong and knowledgeable support, the project encountered minimal headwinds.

It may be that there is nothing more valuable than a high-level executive sponsor to fly cover for an optimization project, but this is not the level at which such projects are typically conceptualized. Intel’s Karl Kempf reports that in the more than 20 years he has been working in OR and at Intel, he has never had a senior executive come to him and request or propose that optimization be used as a way to make better decisions. At first blush, this may seem surprising, given the widespread diffusion of technological expertise in a company such as Intel. But on further reflection, the fact that optimization is usually driven from below the top tier should not be surprising. Recall John Kotter’s classic delineation of management—which is about dealing with complexity—and leadership—which is about coping with change. As he comments, “Companies manage complexity first by planning and budgeting—setting targets or goals for the future (typically for the next month or year), establishing detailed steps for achieving those targets, and then allocating resources to accomplish those plans.” Leading involves dealing with change through setting future vision and formulating strategies to achieve it.

Executives’ trusted advisors, who are on the front lines and struggle with complex operational decisions on a daily basis, are generally the first to look for decision-making assistance. At Intel, the first successful project that Kempf ever undertook started in conversations not with a senior vice president, but with a factory shift manager named Bruce who had a degree from MIT and had been contemplating alternative ways of making decisions about his shift. As Kempf reports:

Working together, Bruce and I designed a series of alternative approaches to making decisions during Bruce’s shift. At the time, Bruce’s factory was running four shifts around the clock. Within a few months, the performance on Bruce’s shift sparkled, with throughput 10 percent higher than the other shifts. Shortly thereafter, at a shift-manager’s meeting, the plant manager asked, “What the heck is Bruce doing that you other guys aren’t?” The resulting pressure helped get the other shift managers on board and in time, performance for the entire factory shot up, bringing pressure from above for the other factories to figure out what they should be doing to keep up.

This type of “bottom-up” development is common in successful optimization efforts. Not only are senior executives unlikely to launch an optimization project; at times they can actually represent a formidable roadblock, given their deep confidence in their own decision-making abilities. At Amazon, Greg Linden was leading a software development team and had the idea of presenting recommendations at checkout time based on what was in a person’s cart. He began to play with the idea and to test what kind of impact it would have on purchase patterns. Two suggested hypotheses were that (1) it would stimulate additional purchases or (2) it would reduce conversion rates because it distracted buyers and prevent them from successfully completing the checkout process.

An Amazon senior vice president of marketing felt strongly, based on his personal intuition and many years of marketing experience, that the second case was sure to be the outcome. He ordered the tests halted. Fortunately for Amazon, quants are not always the best listeners. Linden rushed a quick prototype test into production and found overwhelming support for the conclusion that making recommendations at checkout significantly increased cross-selling and the size of the “average purchased basket.” The rest is history. Today, almost every successful retail website makes suggestions for additional purchases at checkout.

Greg Linden was also fortunate in that Amazon had a strong culture for “letting the data speak.” Once the evidence was in, the senior vice president backed down. It’s hard to argue with success. The moral of the story: seek the support of senior executives—perhaps even recruit them as sponsors—but don’t sit back and wait for them to originate your optimization project.


This post is excerpted from The Optimization Edge: Reinventing Decision Making to Maximize All Your Company's Assets (McGraw Hill) by Steve Sashihara, www.optimizationedge.com.