Leveraging Customer Loyalty Data to Personalize the Service Experience in the In-Person Customer/Employee Context

Andrea Fineman
Carnegie Mellon University

ABSTRACT

Most people value the highly personalized service they receive from their barber, handyman, or financial advisor. These providers get to know individual customers, and they tailor their offerings to better match specific needs. This leads to loyalty, as customers return for the personal service. Interestingly, many national and international service companies (such as retail stores and hotels) buy their customers’ loyalty by running loyalty programs that offer points and rewards. Loyalty programs collect information on individual customers; however, this information never gets passed on to a customer service representative, allowing them to personalize the service they provide to an individual customer.

This paper investigates if loyalty program data can be collected and fed back to a customer-facing employee, allowing them to personalize their performance to an individual customer in a way that both provides value to customers and makes the employee feel they are better at performing their job. To arrive at my service design recommendations for customer/employee interactions in in-person settings, I conducted interviews as well as a large online survey to test my design concepts.

INTRODUCTION

Most people value the highly personalized service they receive from their barber, handyman, or financial adviser. These service providers get to know individual customers, and they tailor their offerings to better match specific needs. This leads to loyalty, as customers return for the personalized service.

Many large companies, such as retail stores, airlines, and hotels, buy their customers’ loyalty by running loyalty programs that offer prizes and rewards. Loyalty programs collect information on individual customers, but interestingly, this information never gets passed on to a service agent, which would allow the agent to personalize the service they provide to an individual customer. Companies currently use the data to improve marketing, but the companies could use the data to improve the quality of the service as well.

On the digital side, however, companies like Netflix and Amazon customize the online experience based on the user’s prior interactions with the service. These personalization features, such as recommendations, may be a free add-on to the core service, but the fact that the leading digital service companies do feature personalization so prominently is a sign that customers find it valuable or interesting.

In this paper I explore this phenomenon, specifically investigating how in-person, human-to-human interactions can be improved using customer data. There is an opportunity to leverage customer data in service delivery, allowing service agents to personalize their service delivery in a way that both provides value to customers and makes the service agent feel they are better at performing their job. The service experience can be enhanced in utilitarian ways, such as providing useful information to the customer, or in emotional ways, by making the customer feel cared for as an individual.

This paper explores the nature of the customer/employee interaction from a psychological and social perspective, in order to understand more fully the nature of the interactions. This understanding is necessary before implementing such a feature in any customer-employee interaction context. A technological solution for how to provide customer data to an employee comes second, and is outside the scope of this paper.

The insights drawn from this study of customer interaction with commercial services can be applied to other types of service as well. I did not study healthcare interactions, non-profit services, or public-sector and government services, but I believe the insights that are drawn from this research project may be extended to those contexts. 1/

METHODOLOGY

My work unfolded over three distinct phases. First, I conducted expert interviews and interviewed a target customer population (frequent business travelers) in order to gain insights on how people perceive value in specific service encounters connected to loyalty programs. Second, I analyzed the data I had collected and used it to construct service-design scenarios. I created a customer journey map using different frameworks for understanding customers’ reactions to services, then sketched several design concepts and wrote and drew storyboards to depict new service design scenarios. Finally, I used an online survey to evaluate the scenarios while also probing the boundaries of what customers and employees feel comfortable with, when it comes to exchanging privacy and engaging in new social interactions in return for the utility of enhanced service. I concluded with design recommendations about how to leverage customer loyalty data in service delivery.

Generative research phase: User interviews

For my first round of user research, I chose to study the air travel context—for a few reasons. I knew my research would depend heavily on user interviews and evaluations by everyday customers and employees. Choosing one specific industry to gather data on would make it easy to compare data from multiple sources. I thought that the airport context might make users more receptive to thinking about scenarios surrounding the use of personal data. Since the airport is a highly instrumented environment prone to surveillance to begin with, where customers provide personal data as a matter of law before entering, I thought that it would make the user testing much more seamless, allowing me to focus on the service innovations instead of convincing the user to take the innovations seriously. Furthermore, because airlines already have such sophisticated loyalty programs, I thought that the airline industry was a realistic context for experimentation.

After investigating the state of the art in terms of personalized service, I talked with frequent airline customers and employees (gate agents and flight attendants) about the travel experience. I wanted to learn about the pain points of air travel customers, and I chose frequent travelers (specifically, those in tiers of different airlines’ frequent flyer programs) to study because they would have a nuanced viewpoint and plenty of experiences to draw from. I was able to do hour-long scripted interviews with 14 business travelers. I asked exploratory questions but also used the Critical Incident Technique (Bitner, Boom, and Tetrault, 1990) 1/ to delve into the customers’ feelings regarding negative experiences and service recovery during air travel. I also asked about the customers’ feelings of loyalty toward the airlines. While I knew I couldn’t assess each interviewee’s actual behavioral and emotional attitudes regarding loyalty to the company, I still wanted to ask each interviewee to try to articulate the level of loyalty he or she felt.

Synthesis of interview data and design of service scenarios

After conducting the interviews, I used synthesis methods common to the interaction design discipline, including customer journey mapping using various frameworks and sketching.

I created a customer journey map depicting the current state of my interviewees’ experiences as air travelers (Figure 1). I combined every interviewee’s description of his or her experience, from arriving at the airport to leaving the destination airport. Then, I overlaid the journey map with a couple of well-known frameworks from interaction design research. The first is called “Rose, Bud, Thorn” and consists of marking everywhere on the journey that is a positive touchpoint (a “rose”), every place that has potential to become a positive touchpoint (a “bud”), and every negative touchpoint (the “thorns”). This helped me see which areas of the customer journey cause the most negativity, and therefore provide an opportunity to neutralize those problems, and which areas are positive, and should be emphasized. I moved the positive and negative attributes to the lower row of the journey map. Then, I overlaid a second framework known as “Doing, Thinking, Feeling.” This is an exercise to help designers get in touch with the inner lives of users or customers. Because I had extensive interview transcripts, I was able to intuit the thoughts and feelings of the passengers as they navigate the various touchpoints. This helped me consider areas of the journey that may not be outright “roses” or “thorns,” but which contribute to the overall experience passengers have during the air travel process.

Findings

Using the synthesis methods mentioned above, I was able to cluster my insights from the interviews into four main categories: customers’ relationship to information, service recovery, most positive and most negative experiences, and social aspects.

Customers’ relationship to information
Some of my most relevant conclusions from the interviews I conducted with business travelers had to do with those participants’ most positive and most negative experiences with air travel. What was very clearly shown in the data is the fact that many travelers, ranging from those who love to fly to those who only fly when driving isn’t feasible, have had their worst experiences caused by a lack of information. Particularly, customers’ worst experiences were caused by times they felt that information was being withheld from them. Rolling delays–when a flight is delayed incrementally by just a few minutes over and over–were a chief source of this kind of information disparity. Especially those interviewees with high status felt that being given insufficient or misleading information was insulting. Said one interviewee,

“I got to [my layover airport] and [the airline] had overbooked the airplane, and they said they had to kick off some people because of weather but I knew that actually it was overbooked. So I didn’t believe their claim.”
Another said, regarding a narrowly missed connection,

“The flight in Houston was still on the ground, but they’d JUST closed the door. We ran all the way there, but they wouldn’t let us on. But then it sat there for 45 minutes and they still wouldn’t let us on. … They told us on the outgoing flight they would never ever open the door [once they’d closed it] to let on additional people. On the way home, the same thing happened though and they opened it! Some guy … in that case they opened it and let us on it. Bottom line we just thought the whole experience was really sketchy.”

These cases of insufficient or “wrong” information are frustrating to the customer in obvious ways, and undermine customer trust in the airline business considerably. Based on my reading of the literature, I shouldn’t have been surprised that information and lack thereof is a major component in customer satisfaction and has been cited in the literature (Bitner, et al).

In order to understand this factor in the context of air travel and the airport experience, I spoke with a service design consultant who has considerable expertise on the airline business about these interview results, and he had some interesting things to say about the nature of transparency. Customers think they want transparency–more information–but with the existence of status tiers, would it actually be desirable (for any party) to have full transparency between airline, employee, and customer? Sometimes, the airline wants each customer’s status to be made clear: some high-status customers like the feeling of being served first, in front of the other passengers on the plane or at the gate, and the airline may benefit from this performance as well. At other times, making clear the existence of different passengers’ status levels could have a negative effect on the average customer experience.

Another case of data imbalance is the sense that the gate agent has a certain amount of discretion he or she is allowed to exercise. Customers don’t know how much power individual gate agents have when it comes to rebooking or reimbursing an inconvenienced customer. When a gate agent does “help” a customer, the customer may feel like the agent did it because he or she wanted to help the customer, and when the agent denies the customer’s request, then vice versa. Opening up these policies and facts to all customers is probably not in the airline’s best interest. Striking the right balance between giving customers plenty of information and not overloading them or causing awkward situations is something I explored in designing and testing my service concepts later on.

Customers’ relationship to service recovery

When I asked the interviewees about their most positive experiences during air travel, participants relayed responses about times that an airline or airport employee helped save them from a self-inflicted mistake. A couple of people said they had gone to the airport on the wrong day, not realizing they’d purchased a ticket for the next day. When the ticketing desk agent let them go ahead and get on a flight that day (one day earlier than what they’d purchase), this constituted a “best” experience. Another participant said he had accidentally packed something important in his checked luggage on a trip overseas, and only realized his mistake at the gate. An airline employee worked hard to find a baggage handler who could get the item before the passenger had to board; this was a “best” experience in his mind as well.

This is an interesting corollary to the service recovery paradox. 2/

While it is well known that customers can sometimes report greater satisfaction with a service that went wrong, then was corrected satisfactorily, than with a service that went off without a hitch, I found that customers (at least in the airline context) had an extremely high rate of satisfaction when an employee helped them get out of a jam of their own creation. In other words, I found that not only were customers satisfied with a service when employees did something to fix an error; customers were even more satisfied when an employee fixed an error that the customer had undeniably caused. 3/ Because I hadn’t read about this phenomenon in the literature, I asked around among the customer experience experts I knew. None had read studies specifically about service recovery when the customer causes the mistake. However, one expert I spoke to conjectured that the emotional component of making a “mistake” in a highly regimented environment like the airport has something to do with it:

“Much of the service recovery literature focuses on organizational mistakes (process or outcome errors). But I can understand how addressing a customer error would be even more memorable because it has an emotional component. With a customer error there’s no one to blame but yourself. You know you’ve screwed up; you have no right to complain and no way to solve the problem yourself. Then the employee magically fixes everything. That’s like absolution! Emotional relief and gratitude for unexpected help.”
He cautioned me, though:

“[A]re they actually more strongly positive than recoveries from employee-caused incidents? That’s less clear. That type of comparison would only be possible to evaluate when both types of incidents happen to the same person on the same trip. It seems like that would be rare.”

The notion of helping customers avoid making mistakes is something that I explored in my design concepts as well.

Customers’ positive and negative experiences during the customer journey
I found that there were few to no negative experiences reported in the time between clearing TSA and arriving at the gate. This is something marketers call “the golden hour”—a period of time when the airline passenger has free time to spend in shops or at food service outlets, and is enjoying the relaxation that comes with clearing security and being on time. 4/ I also found that customers who used this period of time (or their layover time) to engage in travel rituals (such as getting a massage or visiting a particular restaurant) tended to be customers who enjoyed air travel more overall and recovered from service breakdowns better. I wondered if encouraging formation of rituals could result in better customer satisfaction.

Customers’ attitudes toward socializing during the air travel journey
Many customers mentioned social interactions as an important positive aspect of traveling. Frequent business travelers reported enjoying their social interactions with favorite gate agents and flight attendants (whom they got to know quite well because they frequently traveled on the same airline) as well as with friends and acquaintances in the airport. I wondered if airlines or other organizations involved in the air travel journey could support social bonding and increasing the number of social interactions that customers have during trips.

Design phase

During the design phase, I moved back and forth between synthesis methods and more generative design methods as needed by sketching while using the customer journey map. I also wrote text to accompany visual storyboards depicting service scenarios. Sketching is widely considered (in the interaction design and HCI communities) to be a research method, while also producing design output. (Figure 2)

A few challenges arose while sketching because of some specific aspects of the air travel context. Some experimental types of customer/employee interactions may not be applicable in the airport context. For instance, air travel has an extreme tendency toward self-service and efficiency, which causes customers to avoid any human interaction with the airline company, and excludes many opportunities to share interesting information with customers. The relative similarity of passenger needs and can also make the space somewhat less rich for experimentation. Therefore, I designed for different service and retail industries beyond air travel going forward in my project. (This also had the benefit of making it easier to test with a variety of survey takers, because the scenarios wouldn’t rely on the survey taker’s experience with air travel.)

I translated the insights from my interviews and synthesis exercises into five categories of service interaction that have a high opportunity for experimentation with regard to customer-employee interaction:

Category 1: Employee gives customer information about something new or valuable related to the store (i.e., new information relevant to customer’s interests, new products, information that previously was not relevant to customer, but now is)

Category 2: Employee helps customer make a better choice. This includes avoiding customer mistakes.

Category 3: Acknowledgement and recovery from service breakdowns

Category 4: Recognizing the customer’s loyalty. This includes quantified self-y stuff, thanking the customer, showing familiarity/recognition.

Category 5: Service orientation data collection & provision. (New data enters the system via the human channel.)

I designed a total of 20 scenarios with storyboards and text describing the scenario from the customer’s and employee’s point of view. I designed from both perspectives for two reasons: first, in order to understand the service from both points of view and make sure that I was creating something beneficial and humane for both populations, and second in order to test both versions of each scenario in my online survey. A list of all 20 scenarios is in the Appendix.

Once I had designed my scenarios and revised them under the guidance of my faculty advisers, I evaluated them further by gathering data from a variety of customer and employee populations. I needed to understand customers’ and employees’ attitudes toward the use of personal data in in-person interactions. What characteristics might mitigate users’ concerns about privacy violation? Which industries or settings are users most comfortable with using their personal data to customize the service experience? I also wanted to test the desirability of the designs I’d created. For instance, are designs relating to one of the five categories above more desirable than other types of service interactions? Finally, I wanted to see how employees react to being provided with customer data.

Evaluative research: On-line testing

In order to answer the questions that arose after creating my design scenarios, I created an on-line survey. I predicted that my design innovations would have a small positive effect on customers, and in order to provide evidence of the improvement, I would need a large sample size. I also hoped to reach a broad range of respondents from different walks of life this way. Using an online survey does have its drawbacks. For instance, I was unable to ask respondents follow-up questions to the free-text responses they provided. Nonetheless, the survey was very effective at collecting both qualitative and quantitative data on my designs.

Survey mechanics

The survey, constructed in Qualtrics, provided 10 scenarios to each respondent. At the beginning of the survey, respondents were asked if they had worked as a customer-facing employee. If the respondent said yes, he or she received five customer-point-of-view scenarios followed by five employee-point-of-view scenarios. If the respondents said no, then all ten customer-point-of-view scenarios were displayed. The scenarios were presented in random order so that respondents who quit the survey early or got survey fatigue wouldn’t confound the data on the scenarios at the end of the list.

Each scenario had the same questions. Customer-point-of-view scenarios had these questions:

Please think of a time when you recently took a flight. How does the above scenario compare to your own experience?
Less desirable   No preference   More desirable
0 1 2 3 4 5 6

Would this kind of customer service make you want to fly on this airline instead of a similar one?
o Yes
o Maybe
o No
Why?

Employee-point-of-view scenarios had these questions:

How well do the tools that this salesperson uses help to make them better at their job?
Not at all   About the same   A lot better
0 1 2 3 4 5 6

How? Why?

How would you customize your delivery if you were serving customers in this context?

These questions allowed me to collect quantitative data as well as qualitative free-text responses.

Survey results

The survey received 204 usable responses. Of the 204, 90 respondents indicated they were or had been customer-facing employees.

The numerical ratings of the scenarios served as a starting point for assessing the designs’ success. As no scenario received an average score below 3 (scenarios were ranked on a seven-point scale from 0 to 6), I then analyzed the scenarios’ ratings to see if any were polarizing (i.e., receiving many scores at the low end of the range and the highshow illustrate the survey results.
Highest average ratings

In analyzing the survey data, the natural impulse is to compare the scenarios’ average ratings. (Figure 3)

Interestingly, from the employee perspective, two scenarios didn’t get any scores below 3: the hotel information scenario and the anxious traveler airport scenario.

Polarization

In addition to looking at the scenarios’ average ratings, I also analyzed the scenarios according to an additional criterion: I wanted to know if the highest-rated scenarios show some or all people responding with strong positivity, with the rest of the respondents giving a moderate rating (i.e., the scenario isn’t engendering a small but fierce population of objectors). To determine the answer to this question, I analyzed each scenario for polarization, using the quantitative rankings as well as the free-text responses that many survey respondents entered.

Some scenarios had a lot of agreement (most people gave the same score), whereas other scenarios split the population. For instance, the Movie theater scenario (customer version) was not very polarizing. Out of 138 respondents, 39% of the respondents gave it a 3. The free-text responses indicate that the interaction is creepy to some, and that it’s not useful because the employee doesn’t offer a free popcorn or ticket. Similarly, the Paint primer paint scenario (customer version) received overwhelmingly high scores.

A couple of scenarios were more polarizing. For instance, on the Cereal sale customer-version scenario, respondents were split between saying it’s useful and creepy. Out of 154 respondents, 44 said it was useful, and 44 said it was creepy. (Ten people said both.) Respondents for the employee version also mentioned the creepiness factor in this scenario. (Figure 5)

REFLECTIONS

Lessons learned

The data analysis can be distilled into several lessons learned for service designers:

  • While the theory that the service experience can be improved by providing information to customers was resoundingly confirmed, my survey results showed that customers and employees both would like to have access to more information. The amount of information distributed to individual customers can be determined based on a customer’s prior interactions with the service or company. Employees may find it useful to have personal information about customers, even if customers find the interactions enabled by that information to be mediocre. Giving employees data to help them feel more competent at serving customers may be beneficial to employee job satisfaction.
  • Customers are bored by quantified-self-type data, for instance in interactions that provide them with about their past usage (e.g., “This is your 50th visit to Cinema Delux!”). These kinds of interactions should be avoided, as they provide no value to the customer nor to the employee.
  • Using personal data in a way that appeals to an individual customer’s service orientation or that provides utilitarian value won’t be perceived as a privacy violation. On the flip side, data given out of context, even if highly useful to the customer, emphasizes the store’s access to personal customer data (e.g., alerting a customer of a favorite grocery item sale while the customer is shopping in a non-grocery department). Many customers will find this out-of-context data provision intrusive.
  • Finally, my finding regarding the service recovery paradox from the frequent business traveler interview phase—that customers are grateful for interactions that help them avoid a self-inflicted mistake—was confirmed by my online survey. Naturally, it is important to find a way to notify customers of the impending mistake in a sensitive and non-condescending way.

Opportunities for further study

There are many opportunities for further study. Two open questions stemming from my own results have to do with service orientation and service recovery. My research revealed that customers who have a utilitarian service orientation may be more likely to grow impatient or irritated with the use of personal data when it isn’t providing them any useful service. More research on customer segments that may be more prone to offense at the use of personal data in less-specific contexts may be needed in order to help companies implement this kind of service innovation in the most sensitive way.

The finding regarding service recovery stemming from a self-inflicted mistake is another interesting topic for which I could find no specific past research. More research is needed to understand this phenomenon fully.

About the author
Andrea Finemanis a 2015 graduate of Carnegie Mellon’s master’s program in interaction design. An interaction designer interested in digital experiences as well as physical spaces, she is especially excited about designing for services and mobile design. Before joining the master’s program in interaction design, she studied the history of architecture at Brandeis University, worked in advertising briefly, and spent three years at a boutique customer experience consulting firm in Boston. more

Appendix

References

1. Bitner MJ, Booms BH, Tetrault M. The service encounter: Diagnosing favorable and unfavorable incidents. Journal of Marketing 54:71-84. JAN 1990.

Endnotes

1. Many non-commercial (i.e. business-to-business) services these days are adopting interaction patterns from the consumer world in order to increase engagement as well as “customer” satisfaction. http://www.economistgroup.com/leanback/consumers/siegel-gale-b2b-consumerization-study/
2. http://en.wikipedia.org/wiki/Service_recovery_paradox
3. For example, a customer who packed their passport in the checked luggage, and required the help of baggage handlers to retrieve it so as not to miss a weekend abroad.
4.  http://www.forbes.com/sites/sap/2014/08/08/airport-retail-the-golden-hour-and-4-missed-opportunities/

Figures

1. Customer journey map
2. A preliminary sketch
3. Average rating for each scenario
4. Highest rated scenarios, customer and employee versions
5. Distribution of scores given to three customer-point-of-view scenarios to demonstrate polarization among survey respondents