By Nancy Moffitt

David Bell was nervous. He’d been asked to give a talk at a retailing conference in San Francisco and, to his surprise, found upon arrival that he’d been scheduled as the keynote speaker. Before him was a room of 75 or so of the nation’s leading retailers and manufacturers, two groups known for decades of bitter conflict, with the retailers camped stiffly on one side of the room, the manufacturers on the other.

The speech he’d prepared was a somewhat abstract mathematical/economic analysis of trade promotions – deals manufacturers and retailers had traditionally used as weapons in a zero-sum game. Manufacturers had long been tied to an “off invoice” system that gave retailers periodic discounts during a promotion period – discounts that, in theory, retailers were supposed to pass on to consumers. But retailers had abused these promotions, often “forward-buying,” the practice of purchasing more than they could sell during the official promotion period and/or diverting the product to other retailers who were not privy to the trade deal, thereby pocketing the savings themselves. Not surprisingly, manufacturers despised the system, and even retailers claimed to be frustrated by funds wasted on administrative and inventory costs.

Bell, an associate professor of marketing, was about to propose an overhaul of the dysfunctional arrangement – a way to seek peace, and profits. But in a room with many more manufacturers than retailers facing him, he wondered how what he was about to propose would go over. “I was quite worried about how my talk was going to be received,” said Bell. “But there I was, so I just forged ahead with it.”

The upshot: Bell told the group that retailers are not inherently “evil” – they were simply responding to the incentives on the table. A solution could be found, Bell said, in a variation of what was then a new but unpopular type of trade promotion known as pay-for-performance – rewarding retailers based on what they sell rather than offering up-front discounts. If manufacturers were willing to sweeten their pay-for-performance deals enough, the counter-intuitive result was that they could offer a better deal to retailers, yet make themselves, and ultimately consumers, better off in the process. Give away more money to get more money, Bell told the manufacturers, wondering what the reaction would be among the embattled, cost-conscious crowd.

“But they were surprisingly candid – as were the retailers,” he said. “Both agreed that the system was problematic, but I think that after our discussion the manufacturers were less enamored with the view that retailers were underhanded and that all the retailers were doing was responding to an incentive system fueled in part by the manufacturers themselves.”

“If you give a retailer an incentive that’s based on what she’s going to buy, then she’s going to try to optimize the buying function,” Bell said. “So you have to turn the incentive around and say, ‘Ok, instead of giving you a dollar for everything you are buying from me – a discount – I’ll give you a dollar discount for every case of product that you sell, that you scan through your system.’”

Ultimately, Bell’s uncomfortable, unexpected keynote speech in 1999 was the beginning of a real shift in organizational culture. Immediately after the talk, a former Procter and Gamble executive named Bob Gibson approached Bell, explaining that the analysis the professor had just presented was the crux of a company he’d founded called Scanner Applications. Gibson, a Harvard MBA, had spent his early career in the packaged goods industry working for leading manufacturers and had grown increasingly frustrated by the enormous waste in trade promotion. He’d started Scanner Applications to promote the pay-for-performance concept, using the retailing industry’s ubiquitous scanner as a way to track sales. Gibson invited Bell to his Cincinnati headquarters to talk to his sales force about the theory underlying what they were doing. He also got Bell access to data – data that allowed Bell to test, prove and forward his theory via two high-profile research papers, a 2002 article in the MIT Sloan Management Review and another last year in Marketing Science.

Both papers went beyond purely hypothetical data, reporting evidence from a yearlong field test – an experiment that showed that “scan-backs” really work. In the field test, a national-brand beverage manufacturer implemented both off-invoice and scan-back trade deals in cooperation with retailers in four regions of the United States. Each retailer received up to four scan-back and four off-invoice deals at different times over the year. To make the two deals equally attractive to the retailers, the scan back included a deeper discount than the off invoice.

What happened? The results revealed that scan-back deals generated more sales and marketing support from retailers, greater pass-through of discounts to consumers, limited “forward buying,” the process of stockpiling discounted product for later sale, and more stable retailer demand.

Today, the sea change Bell proposed has increasingly become a reality. Roughly 65 percent of packaged-goods retailers’ promotion dollars are devoted to pay-for-performance deals, up from about one-third a decade ago. And Bell has enjoyed seeing work that began as theory take shape as practice, and strengthen an industry.

“When you compensate the retailer based on what they sell, there’s no longer any benefit in them loading up on all this inventory,” Bell said. “And all of the inventory infrastructure and shipping things all over the country – which is paid for by the retailer at the manufacturer’s expense, but from a systems point of view is a complete dead-weight loss – is eliminated. Retailers can dramatically cut inventory costs and reorient their activities around what should be their core competencies – selling and marketing,” he adds. “The news is good for consumers, too, because retailers are much more likely to pass on the full amount or even greater than the full amount of the deal on to the customer.”

A Rising Star

Bell, 38, a native of New Zealand, moved to the United States at age 25 to attend Stanford University, where he earned his PhD. He joined Wharton in 1998 after three years at UCLA, where he began his academic career in spite of early offers from Wharton and other East Coast schools because California reminded him a bit of home. An avid rugby and squash player, Bell also plays the guitar and started a band with a group of friends, including two professors from Harvard and MIT, and says he loves playing pool so much he became the PhD Association president at Stanford for the sole purpose of buying a pool table for the student lounge.

He fell into academia somewhat by accident: As an undergraduate at Auckland University, he found he particularly enjoyed his research-oriented classes. “Something about trying to answer questions of human consumer behavior really intrigued me,” he said. “I liked marketing in particular because it is very eclectic.” When he graduated at age 20, Bell admits, he opted for graduate school as a way to avoid joining the workforce. “Now that I am in this job, I feel that it really fits, but a lot of getting here was a matter of luck and circumstance,” Bell said.

A prolific scholar, Bell has published 15 research papers in leading journals between 1997 and 2003 – with nine others under review or revision – on everything from the effect of word-of-mouth on Internet sales to how store location and pricing structure affect shopping behavior.

A 2001 research paper called “Store Choice and Shopping Behavior: How Price Format Works” scrutinized shopping patterns at “every day low price” stores such as Wal-Mart versus HILO stores, such as clothier Ann Taylor, which have great disparities between the highest and lowest prices, thus relying on sales and promotion strategies to entice buyers. The study provided managers with a new shopping framework by considering both pricing and location in store performance, weighing such factors as distance to the store, familiarity with layout, product assortment and pricing strategies to create a model store managers can use to predict how and when people will shop.

A more recent study on the effect of inventory on buying habits, meanwhile, challenged traditional economic frameworks that suggest that full coffers negatively affect a person’s willingness to buy: laundry room shelves stocked with detergent, it stands to reason, mean a consumer won’t likely buy more on their next shopping trip. Bell’s research found, however, that when it comes to discretionary food products such as ice cream and soda, the opposite actually occurs. “If it’s a discretionary food product, the level of inventory seems to directly affect the rate of purchase,” Bell said. “People actually consume items like this more quickly when they have more on hand. Other more mundane product categories like paper towels, butter and detergent don’t seem to be affected.”

And in the brave new world of Internet retailing, Bell studied the effects of word-of-mouth or other “social contagion” factors on consumer willingness to try an online retailer, using data provided by Lisa Kent, the CEO of, who was introduced to Bell by a Wharton MBA student. “For retailers, ‘location, location, location’ is a familiar mantra,” Bell said. “But for the Internet retailer, the physical location of the store is meaningless, and this unique market context, with geographically dispersed customers and competitors, raises important and largely unstudied questions about the evolution of the customer franchise.”

Bell’s study found a significant “neighborhood effect,” with a 50-percent increase in the base rate of consumers trying an online retailer’s services once they talked about or otherwise observed its use locally. “What seems to be critical,” Bell said, “is the location of existing customers relative to new potential customers. And why this is really important for the manager is that if this effect is really going on, maybe there’s a way for you to feed the process. Maybe if you put a billboard up on a major highway and get people in a very urban, dense region to buy first it’s going to spread a lot more quickly.” Interestingly enough, this effect disappears for repeat purchases – once consumers have their own experience they rely on this and are more likely to disregard the actions of others.

As for the future, Bell has ongoing projects in retailing customer base evolution, customer response to marketing initiatives, and coordination issues. He is also interested in trying to explain and describe fundamental market characteristics, such as price dispersion and the diffusion of information, and is working on a textbook using his marketing strategy notes as a starting point.

“I’d like to advance the use of scientific approaches to tackle important marketing problems,” Bell said. “I’m a firm believer in the value of data, models and empirical analysis to complement managerial intuition, and I try to demonstrate this value in my own research, whether explaining a particular market phenomenon or simply quantifying the response to a marketing decision. This view of academic research has a big influence on my teaching as well.”

Bell enjoys introducing students to a new “spin” on marketing – namely that marketing problems are amenable to scientific analysis and that practitioners can benefit tremendously from this approach. This philosophy, Bell said, is typically somewhat surprising to students, as they often have a completely different view of marketing. “It’s very gratifying when they come to embrace this idea,” Bell said. “I’ve been fortunate to teach the marketing strategy elective because this course allows me a lot of flexibility with respect to content and approach. I enjoy interacting with students, challenging them and being challenged in the classroom, and I feel very fortunate to be able to interact with students of the caliber that we have at Wharton.”

Frequent contributor Nancy Moffitt is the former editor of the Wharton Alumni Magazine.