Research Focus: Effectiveness of Grocery Retail Loyalty Programs

retail customer

Are loyalty programs effective? With the amount of money invested in customer loyalty programs, this is a question that matters greatly to a lot of stakeholders. We may now finally have some scientifically sound answers, at least as it relates to grocery retail loyalty programs. In the largest scale academic study of 358 retail brands across 27 different countries, Professor Bombaij from Tilburg University and his coauthor found some good and some bad. Their research findings are reported in a paper to be published in International Journal of Research in Marketing. In this research focus feature post, I would like to discuss the key findings from the research and what they may mean to your loyalty program management.

Overview of the Research

The two researchers were interested in finding out if loyalty programs really work, what program designs are effective, and if program impact depends on the retail type and on country and cultural differences. They started with the top 15 grocery retail brands in 27 countries (17 in Western Europe and 10 in Eastern Europe). For each brand, they tried to gather information on its sales, loyalty program design (if there is one), and other business characteristics. Combined with country level information, these data allowed the researchers to systematically analyze the impact of different factors on each brand’s sales per square meter. Importantly, they used rigorous statistical methods to make sure that any differences found are real differences due to the loyalty program, not because of some other things mixed in there.

Do Grocery Retail Loyalty Programs Work?

68% of the brands in the study offered a loyalty program. Out of these, 115 or 47% had a positive impact on sales. 127 did not show any impact, and 3 showed surprising negative effect on sales. When the researchers extrapolated their findings to eight other countries not in their sample, they found that, assuming a moderately competitive setting, supermarkets and hypermarkets had a half-and-half chance of succeeding/failing with their loyalty programs. These insights are valuable because previous research has usually studied a single program in isolation. This large-scale analysis offered a big picture view of loyalty program impact. An important takeaway is that loyalty program performance varies widely. To say such programs simply work or not work would be oversimplifying matters.

What Loyalty Program Design Factors Help?

So what types of loyalty programs have a higher chance of being successful (as measured by sales)? The researchers looked at four design factors:

  • Reward timing: does the program offer immediate rewards (e.g., cash discounts at the time of purchase) or delayed rewards through point accumulation?
  • Reward type: are the rewards offered directly related to the business (e.g., a free ham for a grocery store) or not directly related to the business (e.g., a movie ticket for a grocery store)?
  • Reward structure: does the program offer a constant reward structure where consumers are rewarded the same way no matter their spending (e.g., buy 10 coffees and get the 11th one for free)? Or does the program offer a progressive reward structure favoring high-spending, high-tier consumers?
  • Single vs. multi-vendor program: is the program run by a single main retail brand, or is it a coalition program that includes multiple brands?

Somewhat surprisingly, neither reward timing nor reward type had any impact on loyalty program success in this sample. But the last two factors did. Specifically, a simple constant-reward structure was found to be more effective than a progressive-structure program. In addition, single-brand loyalty programs worked better than multi-vendor loyalty programs.

It is important to note a couple of caveats with these findings: (1) These are the top grocery retail brands in each country. Single-brand loyalty programs make more sense in that setting. For small retailers that do not have the ability to manage a standalone program, a multi-vendor program may still make sense. (2) Also because of these brands being well-known top brands, the brand-reinforcement effect from offering direct (brand-relevant) rewards found in previous research may not be noticeable here. (3) Grocery purchases are typically considered transactional and utilitarian. This will also affect how people respond to simple vs. complex programs and how they favor immediate satisfaction vs. delayed gratification.

Do Business Types Matter?

Across different retail formats, the researchers found that loyalty programs were equally effective for supermarkets (similar to Kroger) versus hypermarkets (similar to Wal-Mart). But such programs were less effective for discounters with lower price and more limited product assortment and service (similar to Aldi). How big the brand is or whether the brand follows a high-low pricing strategy or an everyday-low-price strategy did not make a difference on loyalty program success.Across different retail formats, the researchers found that loyalty programs were equally effective for supermarkets (similar to Kroger) versus hypermarkets (similar to Wal-Mart). But such programs were less effective for discounters with lower price and more limited product assortment and service (similar to Aldi). How big the brand is or whether the brand follows a high-low pricing strategy or an everyday-low-price strategy did not make a difference on loyalty program success.

Additional analyses within the same research also showed that business type does tend to drive the design of the program. For example, bigger retailers tended to favor delayed rewards. In comparison, discounters favored immediate rewards, indirect rewards, a constant-reward structure, and a single-brand program. Basically discounters seem to offer simpler programs with less moving components, an understandable choice.

One thing this research was not able to answer is whether a particular type of design would be more effective for a specific type of business. We know from the above that different types of businesses tend to choose different types of program design. Therefore, it is difficult to tease apart whether a particular design worked well for a particular type of business because the business type favored that design for some pre-existing reasons or because those program designs are particularly suitable for that business type. A much larger sample would be needed to answer that question, where each design variation needs to exist in at least a few businesses within each type.

Do Markets, Countries, and Cultures Matter?

Very little research has been done on whether loyalty programs may work differently under different cultures. This research fills that big gap. It found that loyalty programs tend to be more effective in individualistic cultures. As individualistic cultures value the uniqueness of each individual and believe in success through individual efforts, it’s not surprising that the individual reward based loyalty programs worked better in such cultures. How much a culture is long-term oriented mattered too. Since loyalty programs typically involve long-term commitments, they worked better in more long-term oriented cultures.

At a market level, how competitive a country’s retail market is did not make loyalty programs more or less successful. But if a lot of competing grocery stores all offered loyalty programs, the effectiveness of the programs tended to suffer. These results echoed my own analyses of airline loyalty programs a while back. However, there I also found that category flexibility (i.e., how much the industry can attract demand from other related categories) can make program saturation less of a problem.

Putting It All Together

One important lesson from this research is that loyalty program effectiveness is highly contingent on program design and operational settings. Often there is no single rule that will work across the board, although there are some emerging patterns. The researchers’ simulation analysis showed what would happen if the programs modified their design characteristics to the more optimal structure. That is, if the programs all switched to a constant reward structure offered by a single dominant brand. In that case, 160 or 65% of the programs would have become effective instead of the original 115 or 47%. Do keep in mind though that these findings were based on the grocery retail context. The same rules may not work for another industry, say, cosmetics, apparel, or travel.

Good things are better shared!