Monday, January 29, 2007

A Penny For Your Thoughts: Referral Reward Programs and Referral Likelihood

The January 2007 issue of the Journal of Marketing contains an article that will be of interest to those seeking to understand WOM and referral reward programs. Researchers Gangseog Ryu and Lawrence Feick wrote the article entitled "A Penny For Your Thoughts: Referral Reward Programs and Referral Likelihood."

Here's the abstract (an executive summary of the study can be found here):

Because referral reward programs reward existing customers and build the customer base, firms use them to encourage customers to make recommendations to others. The authors report on four experiments in which they find that rewards increase referral likelihood. More specifically, they find that rewards are particularly effective in increasing referral to weak ties and for weaker brands. It is also important who receives the reward. Overall, for weak ties and weaker brands, giving a reward to the provider of the recommendation is important. For strong ties and stronger brands, providing at least some of the reward to the receiver of the referral seems to be more effective. The authors discuss the implications of the results for the design of reward programs.
This is a very neat article so I'll go into some detail summarizing it. NOTE: This is not a comprehensive summary of the article, so please refer to the entire article for all of the details. If you're more interested in the practical/so what? implications, skip down to the section entitled "Practical Implications for Managers".

Goals of the Study (pp. 84-85)

The researchers set up four experiments where they examined the effects of the following on referral likelihood:

- the presence or absence of a reward;
- the size of the reward;
- who receives the reward (the existing customer, the new customer, or both);
- the impact of the relationship between the two parties (strong or weak tie [roughly speaking, this measures the closeness of the relationship: strong ties are well-known others while weak ties are less well-known acquaintances]);
- brand strength (strong or weak; a "strong brand" is one that has high brand awareness and well-established brand associations (see p. 86)

Theoretical Background (p. 85)

The researchers used exchange theory which, in this context, suggests that people make decisions to engage in WOM based on their perceptions of the costs and benefits associated with doing so (for example, benefits of WOM include reducing anxiety after the purchase, to be seen in a positive light by others, and expressing concern or helping another person out; costs might include the time and effort while communicating and the possibility that their conversational partner isn't happy about the recommendation they received from the other person which could damage the relationship). They add, though, that referral programs carry additional complexity when it comes the cost-benefit analysis. For example, when a referral is rewarded the conversational partner might perceive that their friend is making the referral to get the reward rather than for the intrinsic motivation of helping another friend out.

ROI of Referral Programs

For a referral program to be deemed effective, a company needs to obtain results beyond what would have otherwise occurred naturally. In the WOMMA Terminology Framework terms, the marginal gain from the "amplified" program needs to be able to cover its costs and go beyond the "organic" WOM that would have happened without the program.

Study 1: Presence/Absence of Rewards and Reward Size on Satisfied Customer's Referral Likelihood; Strength of Ties and Brands

The research method for the first study involved 275 undergraduate students at a university in Singapore completing surveys. The students were randomly assigned to different experimental conditions (whether a reward was used at all, the size of the reward, close friend versus casual acquaintance) and then asked to imagine they were wanting to buy an MP3 player and then which of two brands they would prefer (Brand A versus Brand B; brands were described as having higher/lower quality and higher/lower reputation to indicate brand strength; no actual brand names were used to avoid effects of prior brand beliefs). The size of the reward, if present at all, was 10% of the sales price (small reward condition) or 20% (large reward condition). Participants then indicated how likely they were to refer on a scale of 0% ("certain not to tell this person) to 100% ("certain to tell this person"). See the study for additional details and other variables.

Results indicated that referral likelihood was greater with strong ties than weak ties and that offering a reward significantly increased referral likelihood (however, the size of the reward didn't make a significant difference in increasing referral likelihood). When you take into account the relationship between the parties the researchers found that the presence of a reward did not affect referral likelihood with strong ties. With weak ties, though, offering the reward increased consumers' likelihood to refer. Further, the rewards were more likely to increase referral likelihood for weak brands (by 20%) but less likely for strong brands (only 10% increase).

Study 2: Trying to Understand the Mechanisms Underlying the Effects from Study 1 by Looking at the Perceived Social-Psychological Costs and Benefits of a Referral

Without describing the procedures, I'll summarize the results by stating that the researchers found support for the principles of exchange theory: consumers can and do evaluate the social and psychological costs and benefits differently when a reward is involved (p. 89). In general, any potential social benefits of a referral are perceived as lower, and the costs are perceived as higher, when a reward if offered versus when one isn't offered. But again, tie strength matters here. There are more perceived social and psychological costs and benefits between strong ties and when a reward is present. However, when engaging in WOM with weak ties, consumers are more likely to recognize the economic benefits of the reward and not worry as much about the potential social and psychological costs and benefits.

Study 3: Does It Matter To Whom the Reward Is Given? "Reward Me" (Existing Customer), "Reward You" (New Customer), and "Reward Both"

Again, cutting to the chase here, the researchers found it does make a difference who gets the reward based on the relationship of the parties to each other. With strong ties, it's not as important who receives the reward (the existing customer, the new customer, or both), though the trend indicates it's probably better to err on the side of rewarding the new customer or rewarding both people. With weak ties, however, rewarding the existing customer had the most impact on increasing referral likelihood.

Study 4: Seeking To Generalize Findings Beyond Hypothetical Scenario with MP3 Players to Actual Experiences with Mobile Phone Service

In this study, the researchers wanted to see if the same findings would apply to a service (rather than a good) and whose attributes are based on actual experience (rather than search characteristics). This study was done in South Korea and the following results of previous studies were replicated:

- Offering a reward to an existing customer increased referral likelihood between weak ties, but not strong ties;
- Referral likelihood was greater when a reward was offered for a weaker brand;
- For weak ties, there was little difference between the "Reward Both" and "Reward Me" reward schemes. For strong ties, referral likelihood was higher with "Reward Both" rather than "Reward Me."

They also found an interesting interaction effect between brand strength and the reward scheme. For a stronger brand, the "Reward Both" scheme performed a lot better than the "Reward Me" scheme. But for a weaker brand, the "Reward Me" performed just a little better than the "Reward Both" scheme. The researchers think that the underlying mechanism here is similar as for tie strength. That is, since people have a high level of commitment to a strong brand and a strong intrinsic motivation to refer, the "Reward Me" scheme leads people to think of themselves and their motives in a negative way (for example, feelings of guilt). A "Reward Both" scheme, however, may reduce the psychological costs involved with getting a reward since it's shared with the other person. Further, consumers of a strong brand tend to be less price sensitive and so the higher economic gain in the "Reward Me" scheme is probably perceived as less important.

Theoretical Contributions

From a theoretical perspective it expands social exchange theory from just the relationship between two parties, to include three parties: the existing customer, the new customer, and the brand. It also shows the utility of using exchange theory to understand WOM.

Practical Implications for Managers

Here's a quick summary of the implications the researchers identify:

Result: Offering rewards can increase referral likelihood, but there wasn't a difference in effect between smaller and larger rewards (in this study, 10% versus 20% of the purchase price).
Implication: Calibrate reward size carefully by calculating the revenue impact of the size of reward with how much additional referral likelihood is likely to occur (see the detailed results in the article for specific percentages of how much the referral likelihood might increase using different reward sizes). Then compare this calculation with the cost of other reward programs. (p. 92)

Result: Among strong ties, rewarding the existing customer (Reward Me) didn't increase the likelihood to refer. But rewarding the new customer (Reward You) or both (Reward Both) can slightly increase referral likelihood among strong ties.
Implication: Since most referral reward programs tend to involve strong ties, either explicitly (as in the case of "friends and family programs") or just end up that way (because people are more likely to interact with strong ties and more likely to know what is relevant and important to them, thus leading to higher recommendation behavior), its probably best to use the Reward You or Reward Both condition, rather than the Reward Me condition.

Result: Rewards increase referral likelihood to weak ties. Further, in weak tie recommendations, the recommender needs to receive a reward in return for the recommendation.
Implication: Since weak ties play an important role in "bridging" one social network to another social network, they are important to involve in the referral program in order to increase the spread of WOM. However, the trick is in devising programs that involve weak ties. Referrals to strong ties are likely to happen without the reward program and are likely to happen first. The researchers suggest that weak tie referrals is what really need to be encouraged and are likely to happen after the strong tie referrals.

One option suggested by the researchers is a segmentation approach based on reward size and reward scheme. First, increase the size of the reward as the number of referrals increase. This recommendation is based on the assumption that strong tie referrals are more likely to occur before weak tie referrals. The referral reward program would thus pay the least for strong tie referrals that would presumably occur naturally and pay the most for those weak ties that would be least likely to occur organically. Second, the reward scheme would shift from "Reward Both" or "Reward You" for strong ties (and presumably earlier referrals) to "Reward Me" for weak ties (and presumably later referrals). Shifting the rewards to the recommender should naturally switch the emphasis to weak tie referrals based on the social-psychological mechanisms discussed earlier.

Result: Distributing a reward to increase referral likelihood is more important for consumers who perceive a weaker brand.
Implication: The researchers suggest that weaker brands can benefit from rewards programs, even if the long-term goal may be to increase brand strength. In the shorter term, the brand may want to use a rewards program to increase trial and/or referral likelihood. Practitioners should use the Reward Me scheme for a weaker brand and the Reward Both scheme for stronger brands.

Limitations of the Study

The researchers point out a few limitations to their study:

- Future research needs to add in receiver receptivity as an important variable. Some receivers may be more receptive to a referral than others and the researchers suggest it's important to understand more about the dyadic relationship between the parties to understand this aspect of receptivity.

- Referral likelihood, rather than actual referral behavior, was studied. Since referral likelihood may not lead to actual referrals, future research needs to collect behavioral data.

- Different types of rewards should be considered beyond just monetary rewards. One option I thought of might be to grant access to exclusive information not provided to other consumers. This may reduce negative self-perceptions associated with receiving a monetary reward. Or, perhaps give the option of donating the reward to a charity or other "good cause."

- Future research should look at the effects of rewarding referrals on the recommender. For example, might rewarding the referral actually lead the recommender to feel less satisfied with the brand, less loyal, or otherwise change the attitude toward the brand? Might the referrer start to feel more like a "mercenary" and could this affect how they view the brand?

- Future research needs to look at the differences between knowledge of the program before purchase and after purchase. This study focused exclusively on post-purchase knowledge of the reward program.

- Cultural differences may play a role in the results. All of the participants were from Asia where other research suggests that Asians are more likely to see themselves as more interdependent rather than independent. This may affect how the social and psychological costs of the rewards are perceived. But the researchers ultimately suggest that because the findings about tie strength are consistent with earlier studies that involved U.S. participants, they feel that similar results would hold when using a sample of Western participants. (Of course, the distinction between Asian and Western in the study is very broad and might not take into account important differences based on national or regional variation).

- I would also add a couple other limitations that weren't discussed in the limitations section. One, rewards might have other effects besides stimulating referral behavior (such as serving as a reminder to talk about the brand; in fairness, the researchers acknowledge this on page 92). Two, exchange theory emphasizes a rational, cost-benefit analysis of consumer decision-making, and while the researchers show some support for this approach, it would be instructive to look at other frameworks that don't assume a rational actor.

OK, that's the summary with a bit of added commentary. I'll probably write more about this later on. In the meantime, feel free to contribute your thoughts in a comment.

Citation: Ryu, G. & Feick, L. (2007). A Penny For Your Thoughts: Referral Reward Programs and Referral Likelihood. Journal of Marketing, 71, 84-94.