This research presents an empirical study of the impact of a retail frequency reward program on store sales. We examine both the “points pressure,” or short-term impact, and the “rewarded behavior,” or long-term impact. The points-pressure impact is due to forward-looking customers increasing their purchase levels in order to earn the reward. The rewarded-behavior impact is evidenced as purchases above baseline levels after an individual has received a reward and could result from either behavioral learning reinforcement or positive affect resulting from the reward. We investigate a turkey reward program that awarded free turkeys to shoppers who accumulated the required sales levels during an 8-week period. We find both a points-pressure and rewarded-behavior impact. These effects are statistically significant and managerially relevant in that the program is apparently profitable. The points-pressure impact is especially strong among customers who do not place value on frequent shopper programs that in general deliver immediate price discounts. The key implications are that frequency reward programs of the form, “buy x, then receive xx” can be profitable, are segmentation strategies, and can complement a store's overall frequent shopper program.
Frequency reward programs have become significant marketing activities for many companies. Airlines reward travelers with free flights after they have accumulated a required level of travel miles. Hotels reward their customers with free rooms after they have stayed at the hotel a certain number of nights. Bookstores award free merchandise to customers who have purchased a requisite number of books. The common thread is that frequency reward programs provide a tangible benefit to customers for repeatedly purchasing the company's product(s).
The benefits and costs of reward programs are debated in the literature. On the plus side, reward programs are seen as powerful mechanisms for increasing sales or brand loyalty (Kopalle & Neslin 2003). On the negative side, reward programs can be costly, complex, and precipitate significant competitive response (Dowling & Uncles 1997; Kopalle & Neslin 2003).
Fundamental to sorting through these trade-offs is the question, “What is the sales impact of frequency reward programs?” While there is a rich theoretical literature on the economics of these programs (driven by the sales impact) (Beggs & Klemperer 1992; Kim, Shi, & Srinivasan 2001; Kopalle & Neslin 2003), there is less empirical study of the sales impact itself (Bell & Lal 2002; Bolton, Kannan, & Bramlett 2000; Drèze & Hoch 1998; Lewis 2004).
What makes the sales impact question particularly intriguing is that reward programs can increase sales through two mechanisms: “points pressure” and “rewarded behavior.” The points-pressure mechanism is the short-term impact whereby customers increase their purchase rate in an effort to earn a reward. The rewarded-behavior mechanism is the long-term impact whereby customers increase their purchase rate after they have received the reward. If reward programs have a points-pressure impact but no reward-behavior impact, they function essentially as multiperiod, but still episodic, promotions (Kopalle & Neslin 2003). If, however, there also is a rewarded-behavior effect, frequency reward programs can be a strategic tool for building the brand. It is, therefore, important to understand the relative contributions of these mechanisms to the overall sales impact of a particular reward program.
The purpose of this paper is to measure the short-term (points pressure) and long-term (rewarded behavior) effects of a frequency reward program. The setting is a “Turkey Reward Program” implemented by a supermarket chain. We investigate two executions of the program over a two-year period. Indeed, we find evidence for both points-pressure and rewarded-behavior impacts, and evidence that these impacts are related to customer characteristics.
The rest of the paper proceeds as follows. First, we elaborate on our framework and discuss previous research. Second, we describe our method, including the research setting, the design of the turkey reward program, and data collection. Third, we describe our analysis and results. We conclude with a discussion of implications for both researchers and practitioners.
We have examined the impact of a turkey reward program offered in two successive years for a supermarket using actual field data. The results for both years show a current period points-pressure impact and a post-program rewarded-behavior effect. Storewide weekly sales increased on average by 6.1 percent in Year 1 and 6.4 percent in Year 2 during the 8-week program period, as shoppers increased their purchase levels in order to receive a reward. This points-pressure effect was particularly strong among shoppers who ordinarily do not consider frequent shopper programs to be an important benefit.
We also found that the shoppers who were rewarded with a free turkey experienced a rewarded-behavior effect. That is, the redeemers’ purchase rate did not immediately return to preprogram levels after they redeemed the free turkey. The magnitude of the effect is an increase in store sales of 1.8 percent per week over a 7-week period.
It is interesting that the points-pressure effect is larger in aggregate than the rewarded-behavior effect. This is consistent with work in the sales promotion field that finds repeat purchase effects to be smaller than current period effects (Gedenk & Neslin 1999; Keane 1997). However, the rewarded-behavior effect is isolated to a minority of customers who redeem a reward, and can be quite substantial among these customers.
One question is the extent to which these findings generalize to other venues. While a free-turkey reward program seems specialized, it is not dissimilar in form than the “Buy x-times, get a free xx” structure of many reward programs common in the supermarket industry, the retail industry, and in other consumer products industries. However, further research is needed to examine how our results differ according to the values of “x” and “xx.”
