Journal Must-reads from Jean-Pierre Dubé, University of Chicago

March 22, 2018

Jean-Pierre Dubé is the Sigmund E. Edelstone Professor of Marketing at the University of Chicago Booth School of Business. He is also director of the Kilts Center for Marketing at the Booth School and a Research Associate at the National Bureau of Economic Research.

His recommendations include publications from leading core-discipline journals in economics and science as well as marketing, with authors from economics, medicine, and law. “This multidisciplinary approach is important for innovating marketing practice and methods,” he notes. “That is, besides ‘porting’ (or poaching) ideas from other areas, we should facilitate dialogue between these areas to ensure that new approaches are held to appropriate standards of rigor.”

Recommended reading
1
A Structural Model of Sales-Force Compensation Dynamics: Estimation and Field Implementation by Sanjog Misra and Harikesh S. Nair, Quantitative Marketing and Economics

Easily one of the best quantitative marketing papers in the past decade. The study was conducted in collaboration with a large optics manufacturer using a complex, incentives-based salesforce compensation scheme with salary, commission, a quarterly bonus after a quota, and a cap on the bonus.

“Economic theory tells us such compensation could lead to unintended consequences if agents ‘game’ the scheme. A clever and detailed analysis of the data reveals striking evidence of gaming. On average, agents sell more in the third month of each quarter than in months one or two. Agents who are far from quota ‘give up’ and delay sales to the next quarterly cycle. Agents who are close, accelerate their sales. Agents who are near the cap stop selling and delay sales to the next quarter.

“By estimating a state-of-the-art economic model to the data to predict sales agent ‘effort’ the authors derive an ‘optimized’ incentive scheme, consisting only of salary and commission. They implement the recommended scheme and find: (1) gaming completely disappears and (2) total sales increase by nearly 30%. A triumph for economic theory, quantitative marketing and the returns to collaboration between firms and researchers.

 

2
Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field Experiment by Thomas Blake, Chris Nosko, and Steve Tadelis, Econometrica

“eBay used to spend nearly 30 million per year on paid search advertising for keyword searches for their own brand. A consulting firm documented a strong correlation between these paid search impressions and sales on eBay.com, concluding an ROI over 600%. This seemed way too good to be true. The authors of this paper (who were working at eBay at the time) got into a fight with the marketing team over the plausibility of these findings. They exploited a natural experiment – eBay shut down its paid search advertising on Yahoo! and MSN due to a dispute – and using a before-after design, they found that the increase in organic leads was almost identical to the loss in paid search clicks. Worried about confounding trends in click behavior, they cleverly used Google to de-trend (a difference-in-differences design).

“The authors find that over 99% of paid search leads are cannibalizing what would have been a naturally-occurring organic search lead. This is because eBay appears at the top of organic search results. Making matters worse, organic links are free. The authors conduct a follow-up, national randomized field experiment and find the ROI on paid search for the brand keyword to be -75%.

This study along with a string of subsequent digital advertising studies raise some strong doubts about traditional methods used to measure advertising effects in practice by firms and their media consultants. The strategic way firms place their advertising would likely lead to confounds similar to the ones that plagued eBay’s original measurement approaches. More recent work also documents precision concerns. For instance, Lewis and Rao (2015) conduct power calculations for display ad experiments across an array of different industries (retail, financial services, etc.). They find that even if a firm was willing to run a randomized controlled experiment to measure advertising (and avoid the problems discussed above), they would require sample sizes on the order of 10 million subjects just to detect an effect if it is present.”

 

3
The Value of Purchase History Data in Target Marketing by Peter Rossi, Greg Allenby and Rob McCulloch, Marketing Science

“Perhaps one of the all-time classic quantitative marketing papers in Marketing Science. This paper was arguably a decade ahead of its time. The authors use the purchase history of individual customers in supermarkets and then state-of-the-art Bayesian statistical methods to estimate customer-level demand for CPGs. They then use the individual estimates to design a profit-maximizing, personalized pricing strategy that targets individualized prices. The authors find that even with only a short purchase history for customers (e.g., two or three observed transactions), they can achieve considerably higher profit improvements relative to traditional segmentation schemes that target prices based on demographics.

“Amazingly, leading analytics suppliers are still not using such sophisticated methods. In spite of the use of machine learning and other new tools, common approaches to targeting are typically based on ad hoc rules and would not necessarily optimize profits. Many people associate personalized pricing with the recent Big Data revolution. But as this paper aptly demonstrates, supermarket chains could have implemented such pricing back in the 1990s using loyalty card data.

“In practice, pricing is often treated as an afterthought and most firms do not apply microeconomic reasoning to pricing decisions (to their detriment). In a recent collaboration with Ziprecruiter.com, for instance, Sanjog Misra and I found that price optimization increased their monthly revenues per customer by over 60% (a personalization scheme had an over 80% increase). These new pricing structures were tested and validated out of sample.”

 

4
Effect of Temporal Spacing Between Advertising Exposures: Evidence from an Online Field Experiment by Navdeep Sahni, Quantitative Marketing and Economics

“This paper makes an important contribution to the decades-old ‘effective frequency planning’ approach to advertising scheduling. To the best of my knowledge, this is the first paper to test how individual consumers process repeated exposures to advertising over time – using real field data.

“The design of the experiment is a big part of the scientific contribution. Besides showing a large, positive treatment effect of advertising exposure and carry-over, the author measures the role of temporal spacing between exposures. Spreading ad exposures apart in time increased their effectiveness. The author then builds a quantitative psychological model of demand and memory to measure how temporal spacing affects purchases. His results reject our classic models of demand with carry-over. In a series of strategy simulations with the estimates, he shows how the advertising schedules optimized against the memory model produce much higher ad effects and ROI.

“This paper finally gives us the tools to design effective frequency planning in practice, not to mention testing the underlying components of a scheduling practice that has largely been built on intuition without hardcore evidence. In addition, the paper really digs into the psychology of the consumer to teach us why ad response has the unusual ‘shape’ past research has detected. There are few papers in the quantitative marketing literature that genuinely blend economics and psychology into the primitive assumptions used to derive the model.”

 

5
Advertisements Impact the Physiological Efficacy of a Branded Drug by Emir Kamenica, Robert Naclerio, and Anup Malani, Proceedings of the National Academy of Science

“The authors conduct randomized clinical trials to test whether the treatment effect of direct-to-consumer advertising has a causal effect on a subject’s physiological reaction to a drug. In particular, a branded antihistamine was found to be more effective when subjects were exposed to that brand’s advertising as opposed to a competitor brand’s advertising. This finding is remarkable. Subjects literally responded physiologically faster to a drug when they saw its advertising.

“This reaction was measured using a standard medical ‘wheel test’ that indicates the diameter and height of the area of the skin that reacts to the histamine treatment. The wheel test indicated that advertising-treated subjects literally responded minutes faster. This complementarity offers one explanation for the inconsistencies in laboratory studies of consumer preferences in conditions where brands are removed (blind taste tests) versus conditions with brand labels are present (e.g., Allison and Uhl 1964).

“These findings also support the view of advertising as a consumable (albeit intangible) product that is complementary to the physical, branded product. In this case, the effectiveness of a branded drug is enhanced by the consumption of the brand’s advertising. This idea has its origins in the classic paper by Becker and Murphy (1993). The more common view of advertising is as a product characteristic that shifts demand.”

References

Allison, Ralph I., and Kenneth P. Uhl (1964), “Influence of Beer Brand Identification on Taste Perception,” Journal of Marketing Research
Becker, Gary S., and Kevin M. Murphy (1993), “A Simple Theory of Advertising as a Good or Bad,” The Quarterly Journal of Economics 108 (4)
Lewis, Randall A., and Justin M. Rao (2015), “The Unfavorable Economics of Measuring the Returns to Advertising,” The Quarterly Journal of Economics 4(1)

By using MSI.org you agree to our use of cookies as identifiers and for other features of the site as described in our Privacy Policy.