Written by Lina Altenburg, PhD candidate at KU Leuven (Belgium)
Why is one temporary loyalty program successful and another is not? How can retailers design them to be more successful, and which external factors impact success?
source image: www.eighty20.co.za
“It is very difficult to get data on these temporary loyalty programs […] that is also the main reason why we haven’t seen many papers on this yet because the data is just not there.” Nick Bombaij
Nick Bombaij (Assistant Professor at the University of Amsterdam, the Netherlands), Sarah Gelper (Associate Professor of Marketing Analytics at University of Luxembourg, Luxembourg) and Marnik Dekimpe (Professor of Marketing at Tilburg University, the Netherlands and KU Leuven, Belgium) obtained a unique, global dataset to investigate these questions. In their recent IJRM article “Designing successful temporary loyalty programs: An exploratory study on retailer and country differences” (Volume 39, Issue 4, December 2022), the three authors investigate how design characteristics of temporary loyalty programs, as well as retailer and country characteristics, impact redemption rates and ultimately retailer profit.
What are temporary loyalty programs (TLP)?
By participating in a temporary loyalty program (TLP), consumers collect points or stamps for their spending at a retailer during a pre-defined time period (usually a few weeks/months) and redeem them for discounts on rewards such as high-quality glasses, kitchenware, cookware, or household appliances.
See below some examples of temporary loyalty programs (click to expand)
The origin of the research idea
I talked to Nick about this paper, which was the second research project of his PhD at Tilburg University (the Netherlands). As I learned during the interview, his PhD position was funded by BrandLoyalty, a company which implements temporary loyalty programs for retailers around the globe. This sponsorship provided the team with access to an unparalleled dataset and the chance to research the topic.
“They found this topic [of temporary loyalty programs] so interesting that they funded a full PhD position, which made me realize how important it is for them to know more about it.” Nick Bombaij
The research question was originated by BrandLoyalty: “Which factors contribute to a successful loyalty program?” To provide an answer, the authors estimated the effect of 4 design characteristics of the loyalty program (duration, discount, requirement and reward depth), 3 retailer characteristics (the existence of a parallel permanent loyalty program, price strategy and market share) and 4 country characteristics (retail concentration, temporary loyalty program competition, individualism and long-term orientation), as well as interactions between the first and the latter two (for more details, see the full results in the paper itself).
Explaining for academia; forecasting for the industry
To better understand temporary loyalty programs, Nick regularly worked one day a week in the office of BrandLoyalty, the data provider. He participated in the same training that practitioners received who wanted to implement a temporary loyalty program, but also provided regular updates on his research process and findings.
“At some point, I also gave them a more predictive model without the whole endogeneity correction and only based on variables that they knew beforehand, so they could forecast how successful it [the temporary loyalty program] is going to be” Nick Bombaij
The collaboration was very fruitful. Still, at some point, the interest of business and academia diverged somewhat, as BrandLoyalty was also interested in forecasting the success of temporary loyalty programs (from a logistical standpoint), whereas the academic paper focused more on explaining how the various factors (causally) impact redemption rate and retailer profit. To support BrandLoyalty, Nick even developed a tool that the company could use to predict the success of a temporary loyalty program, which they still use in their decision-making. In contrast, in the paper, the authors carefully estimated the effects of the individual factors, while controlling for the multiple levels of endogeneity.
Further advancing the paper during the review process
The review process added two major aspects to the paper: 1) another performance measure for the temporary loyalty programs and 2) a further investigation of interactions between the design characteristics and the external factors.
In the original manuscript, the authors focused on the redemption rate as the main performance measure. The reviewers challenged the authors to also collect sales data, which added valuable insights to the analysis.
“Sales was a huge challenge because while [the data provider] did have data on redemption, they did not have data on sales.” Nick Bombaij
Even though the data provider only collects redemption rates and not the retailer’s sales (due to the retailers’ reluctance to share this sensitive information) for each of the temporary loyalty programs, the authors managed to obtain sales data for a subset of the programs. With this, they could estimate the sales increase due to the temporary loyalty program on a customer (redeemer vs. non-redeemers) and a retailer level, thereby providing evidence that temporary loyalty programs are indeed beneficial for retailers, especially when the redemption rates are high.
Investigating the interaction between the design characteristics, and retailer and country characteristics was methodologically easier, however, provided challenges in another way. As Nick pointed out, the literature on the loyalty program design characteristics covered in their research is sparse, which made it difficult to develop theories for every possible interaction term.
“For the main effects we of course have some theories, but there were so many possible interactions it would just not have been feasible to find a theory for all of them” Nick Bombaij
As often happens in empirical papers, it is not clear what comes first: the theory or the data. Here, the authors chose a combined approached: They developed a theory to explain the main effect, but when studying the interactions, they applied an inductive “empirics first” approach, meaning that they explored the effects of the interactions without developing theories for them. This provided the opportunity to look into all interaction effects and to develop a basis for a more detailed understanding of temporary loyalty programs. As Nick explains, their analysis starts with classical theory testing and combines this with a theory-developing strategy to expand the knowledge in the field for the more detailed interactions between the temporary loyalty program characteristics and the external factors.
Read the paper
Interested in reading all the details about temporary loyalty programs? Read the full paper here.
Bombaij, Nick J. F., Sarah Gelper, and Marnik G. Dekimpe (2022), "Designing successful temporary loyalty programs: An exploratory study on retailer and country differences," International Journal of Research in Marketing, 39 (4), 1275-95.
Meet Nick Bombaij
Nick is an Assistant Professor at the Amsterdam Business School in the Netherlands, father of 2 kids and interestingly my academic uncle (his co-author Marnik Dekimpe is also my supervisor’s supervisor). He has participated in temporary loyalty programs himself, especially when he moved to a new home a few years ago and had to further expand his kitchen equipment with new knives, glasses and pans. When the paper on the temporary loyalty programs was published, he took his family out for dinner. Let’s hope, maybe for his next paper, he will also invite his academic family 😉
What drives you to do the research/work you do?
In academia in general, you have a lot of freedom to do what you want. Of course, with a funded PhD you are a bit more restricted to the exact topic. But at least with the methodology, you can really take your time to focus on things that you want to learn about. And that is something that I find very important.
The temporary loyalty programs are easy to explain and it has a lot of practical connection. Because it is being used by companies, it feels very useful and what you are doing is relevant for practice as well.
If you were not an academic, what would you be?
If I had to go back to the industry then it would probably be something like a data analyst. I get the most joy out of working with data and doing the analysis, so it would be something in that field. Maybe also a teacher of data analytics or something like that
This article was written by
Lina Altenburg
Ph.D. candidate at the KU Leuven (Belgium)
