Case Study: Incentivizing Referrals

The name of the real company in this case has been change to “The Company”.


“The Company”’s main problem is running an affiliate program that wasn’t effective. Despite giving out promotions and offering a promotion in which users can refer their friends, it wasn’t increasing usage further beyond the initial trial. With this information we were also given a set of barriers which included no budget for external merchandising, advertising, or extra materials to aid in our research.


The first step toward creating a successful analysis of the situation was gathering and analyzing all the current data available. This included restaurant usage statistics, customer acquisition, promo code usage, and even email statistics. In order to form a hypothesis and create actionable results to test our theories, we needed to create an entire picture of the data. We ran regression analyses on the data, but found no correlations that were significant. As a next step, we performed trend analyses on data such as total bookings, profit per booking, cover rate, and retention rate since January 2019.


After analyzing the data available we were able to model the following scenario: it will take up to three years to gain profitability with the current referral program in place (5 euros per successful referral). In other words, the current cost of acquisition is too expensive to be profitable short term. Of the 5,384 customers “The Company” currently had, we found the returning ratio to be 16.94%. Even though some users are being referred, the majority are not sharing the program and thereby not creating the intended network effect.

Typically a referral program should be more effective than paid advertising. Studies show that “Referral marketing generates twice the sales compared to paid advertising.(6) and that “every referring customer makes an average of 2.68 invites.(7)” Assuming this number is true, the current program was not as effective as they could be. Referrals were below the 2.68 invites at less than 1 per user on average. That led us to believe that either the wrong marketing methods were used or there may be an overall issue with the affiliate program itself.

To further determine what was actually going on behind the scenes we needed to 1. Create actionable results that we could measure 2.Understand more the theory in why people can refer and what issues in the main platform may be causing the disconnect and creating a barrier to referring.


Using the initial data we formed several actions to take. We identified two challenges to tackle based on behavioral surveys that “The Company” had sent in the past from their CRM software, which targeted current users and asked about their usage behavior.

Hypothesis 1: Build credibility through product features and media coverage

Initial surveys provided to us indicated that user questioned whether the restaurants were the so called “left-over” restaurants, those that were of a lesser quality. By analyzing the survey responses, we identified the following customer characteristics: 1. Live in medium and big cities 2. Speak German and English 3. Use the app rather than the website 4. Dine with one other person 5. Usually book in the afternoon 6. Eat on the weekend. This information was critical in determining where to target: Berlin food blogs, news agencies, and universities.

Hypothesis 2: The right ambassador and incentives will drive referrals

We identified several groups that seemed best to target based on ease of obtainment and most likely fit as “cheerleaders” of the program. Our target? Customers that have used “The Company” within the last month and a half, brought in the highest revenue for “The Company” and booked more than 15 times.

Creating the right referral program was key. According to a Texas Tech survey, “while 83% of customers are willing to refer after a positive experience, only 29% actually do.(5)” This means creating the right incentive while removing any barriers that make referring seem difficult. In designing an effective referral program we also took into account Forrester Research analyst Tina Moffett’s findings that, “Marketers must understand what makes customers loyal, what their expectations are, what motivates them, and makes them feel valued.(9)” At the same time it is important to create not only an “enticing referral reward, but most importantly, crafting compelling messaging that instills an expectation of imminent gains in social capital.(10)”

Keeping this in mind, we created a referral program which was easy to use by laying out a multi- tiered system where customers can more easily gain a sense of benefit from the program. This meant giving rewards that ambassadors would strive to achieve and unique offerings that utilized existing resources that the company had in order to reduce cost. Our program was designed to use resources such as meeting the founders, and utilizing partnerships with restaurants on the platform in order to pass some on some of the cost of the program.

To further test our hypothesis and make sure our referral program hit on the appropriate points, we sent a survey which resulted in some interesting data points. For instance, 59% of users had signed up for “The Company” but did not know about the referral program in place nor how to use discount codes. Additionally, of the survey respondents who identified themselves as students, 54% weren’t sure they would use it within the next three weeks with some concerns, such as the restaurants did not seem enticing enough.


We found that our first hypothesis was outside our scope. It did not fit with the budget we were given as most potential ambassadors had not used the app before and wanted financial incentives outside of the reward program we developed. This mean’t that either our program didn’t fit the needs of our target group of ambassadors or that we weren’t appropriately targeting the proper ambassadors.

After sending the referral invite email to our second target group, 91 users we identified as valuable “The Company” customers, our goal of a 5% response rate was successful. From the invitation test email, one user made 3 referrals which according to our devised incentive program made them an ambassador. The program needs to be monitored and tweaked in the future to ensure maximum efficiency and retention. Each user is expect to generate an additional 15 users yearly​ not taking into account fourth and fifth tier ambassadors. These ambassadors will create a 5.19 time of boosting based on referrals​. In order to be considered effective and profitable the program needs to be monitored and ambassadors encouraged, to hit a target of 45 new ambassadors in three years in order to be profitable.​

Thirdly, website changes need to be made in order to ensure that referrals have their maximum reach. As outlined in our PowerPoint report, barriers exist that are preventing referrals from being made as easily as possible. Following the recommended changes will improve usage at least 30%.

Critical Assessment of Project

Our main concerns were analytical. Could we back it up with enough data? We had less than a year of data so it was hard to create an accurate forecast. We were also not given data such as when the team started and stopped advertising which would also influence results. While we initially asked for website analytics we were not given that information until weeks into the project, and at that point we did not have enough time to incorporate it into our analyses. Being able to combine restaurant usage, email marketing, and web analytics would have given us an extra level of insight into the project.

Without that additional data, I am also concerned there could be bias in the data set used to confirm the hypothesis, especially data found after we formed our final conclusion.

Problem Areas

What we did wrong at the start was not focusing enough on getting answers to the questions that were most important. We should have posed questions first and secondly looked for data needed to find the answers to those questions.

Creating hypothesis from the start may have framed our thought process toward a certain path, and prevented us from thinking of other ways to approach the referral program. It set us in a certain frame that may have led to confirmation bias, which “leads the individual to stop gathering information(11)” once the formed hypothesis is proven. Once we were able to prove or disprove our theories, we may have stopped looking for further possibilities. Had we better ensured a way to prevent this bias, we would have had a large set of ideas to work with rather than the small subset we discovered.

We also should have had more direct communication with the marketing team to compare the ideas they have in motion with our own ideas. This would have prevented overlap and would have avoided our group from presenting ideas that the marketing department already had in the pipeline. On the other side though, the fact that our team and the marketing department came up with some similar conclusions separately, maybe have as well confirmed our idea. At the same time this would have also given us additional insight and ideas into the project. As we were working at the client’s office every day, we would have gained further insight by attending some company meetings or asking for any new insights that had been uncovered in order to prevent repeated actions.

Working with the team on their ideas would have given us more insight into the problem as seen by the marketing department and given us additional ideas to explore. By working essentially as two separate teams, cross-communication was lost that would have given both the ESMT team and the marketing team additional ideas to consider.

Improvable Actions

From the start of the project, being more diligent about getting the information we needed was key. At the same time, we had additional setbacks that wasted more time than should have. For example, one instance that did arise was when we wanted to create behavioral surveys. This took more effort than originally anticipated as the company could not move as quickly as we thought to approve the surveys to be sent out.

While originally our survey concept and layout were approved, it required never ending changes. While one manager would approve our design, another would raise concerns, ultimately requiring that the survey get passed along to several managers. A better way to approach this could have been to either have them design the survey themselves using information we provided (to allow us to focus on more productive measures), or just schedule a meeting with all parties that required sign off on the project.

Generalizability of the Project Findings

Our project can be applied to other companies trying to create a referral platform, to determine whether it is worthwhile or to understand if the referral program will create enough customers to counteract the initial cost. We used a few theories that while applied to this program can also be kept in mind when developing any application. In this theory a nudge is “a relatively subtle policy shift that encourages people to make decisions that are in their broad self-interest.(2)”

The main points that others should learn from the project are some of the following:

Social Nudging: Take for instance the example from “Streak CRM which does this by adding the label ‘most popular’ to their corporate plan on their pricing page. (3)“ We suggested that “The Company” use social nudging to help validate the quality of their restaurants. By listing how many people are visiting each restaurant and which are the most popular it gives social clues to the restaurants. Other companies can use this same theory to consider when developing their apps. Indicating what is the most popular, what people are using, gives the indication of the popularity and clues toward the action they should complete.

Another nudge theory that is also important to consider when developing a site is the use of framing. Let’s take this example from airport security: “Would you like to undergo a full-body pat-down by an agent or a quick and easy body scan in our machine?(4)” What would someone more likely choose? The option framed as easy or the option that was framed as potentially more uncomfortable? Websites can also frame choice options as something that is hard, or something that is easy. Take for instance the company’s current referral program, the way it had been presented initially seemed too hard to understand, left the reader confused as to the exact proposal and framed it as a difficult proposition.

After we did a behavioral survey, we discovered that many customers and potential customers saw the app as not having enough selection, however, in the short term there were easy ways to mitigate this concern by changing the “framing”. What we suggested is to make sure the type of restaurants listings are spread out enough as to make the site seem as if it has more diversity than it actually does. While it is important to actually address the issue, there are ways as well to give the appearance of diversity even if it may not be the case.

We also suggested utilizing existing assets. This can mean strengthening relationships with current users to further the likelihood of customer referral or strengthening relationships with businesses to create more cross-promotion opportunities such as the restaurant putting “The Company” signs on their tables . Looking internally is a good way to see what can be offered instead of looking for new processes that add additional costs. By getting creative with resource usage, companies can add value.


There are several factors in how this project at the same time cannot be generalized. For instance, each company may have different needs in designing an incentive program. While the theory applies, this is not a one size fits all situation.

For instance, what drives and motivates users can be completely different depending on the situation. While a three-tiered motivational system would work fine for “The Company”, the same system could be too difficult for customers of a small coffee shop to claim the same level of value. These customer (and the shop) could likely benefit from a simple program, such as buying five coffees and getting a sixth for free.

Additionally, referral programs may not always be necessary. In fact, a good product should inspire those around it to share it with others regardless of incentives. Products don’t often simply go viral because of an incentive. According to Inc magazine it important that a product “evoke emotion and provide real value (12)” in order to really go viral.

Key Learnings

Understanding what customers want is important when providing the right incentive program. Blindly adding incentives means that customers may not get their needs met which in turn disincentivizes them to participate.

Even if a program is well designed, too many barriers to its actual utilization will hinder any referral program. Analyzing current behavior and in the case of a website, doing a full analysis of how it currently runs and ways to improve it, can make sure that the current inefficiencies don’t hinder performance. In the case of “The Company”, the current system itself needed to be improved. If new ambassadors need to take extra steps to motivate their friends and referrals, it will overall make the program ineffective.

Lastly, it is important to follow up on behavioral surveys after running them. This lets you gauge whether actions taken had any impact on previous results. “The Company” didn’t repeat any surveys which meant it was unable to gauge the effectiveness of their effects. At the same time, designing the right behavioral surveys are also key toward figuring out what needs improvement.



When it comes to team work, most of the time our group was efficient. We met deadlines and made sure to communicate often with “The Company” in order to keep them informed on our current status of research as well as to gain further information from their side. This helped build good rapport with the “The Company” team and avoided any animosity on both sides.

One difficulty of working on a team of your peers is that we didn’t always have clear leadership. As I initially self-sourced this project and put together the team, I was often the de-facto leader. However, as a group effort most of the decisions were democratic in that everyone generally agreed before a certain action took place.

While this was a good way to create conversation around an idea before acting upon it, it wasn’t always the most efficient way to make decisions. We quickly noticed that if we attempted to agree upon everything, it often took more time than it should to get anything done, reducing efficiency. So how did we conquer this and improve our speed of operation?

One of the main ways to improve our communication was through following agile methodology. We used Asana which initially helped us organize tasks. However, despite our best intentions we didn’t use the organizational tool as much as we should have, but we were still able to work in an agile structure.

One of the main principles of agile work is “is face-to-face conversation as well as working together daily throughout the project.(1)” We made sure to work together whenever possible. While it would have been easy to work from home, we set a daily time to be in the client’s office and worked from there as if it was a normal job. We also discussed at the beginning of each day what each team member would work on so we made sure no one was sitting idle. We also made sure to check in as to each team members progress throughout the day to make sure we were still on task.

This face to face communication also worked well being within the client’s office. When we needed information from the marketing department or another sector of the business, we could work with them directly. It was also easier to keep up with the client if information wouldn’t get sent or if we had any questions. Despite our great work flow there were still at times both internal team conflicts as well as conflicts within the organization.

A counter-productive situation occurred during the final preparation of our presentation. Certain group members got angry and yelled at another member creating a tense and uncomfortable situation. Two members raised their concerns personally with me outside the team study room. I advised speaking directly to the member raising their voice and as well talking to the group member to which the anger was directed.

My initial thoughts were, was there a miscommunication? Of course stress was effecting the situation which made everyone on edge. However, by talking to the person who lashed out, we were able to understand what happened. It appears that stress, the amount of work and the responsibility was creating tension and this person lashed out. They took it upon themselves to apologize and listened to our feedback. This open dialogue made for a better work environment, and rather than blame them directly, allowed for real constructive feedback for which we recommended maybe taking a moment outside to think it over before lashing out.

Personal Reflection

How did I manage communication with the team? Our main initial channel was through Whatsapp chat. This allowed any message to be shared among the group and create a dialogue outside of our workspace. I made sure to keep the team structured through a Microsoft Teams shared workspace which helped keep everything organized. I also alerted the team whenever a new important document was added, or something had to be rearranged. This kept everyone informed as to the current progress and where they could access certain information.

Another important way that I made sure that communication was effective was making sure that all emails went through a shared group email so that all emails were seen by everyone concerning the group project. This approach to keeping information public and making sure everyone could follow along made sure everyone in the group knew everything was going through clearly.

Finally, the last tool I used to keep information clear was making sure that we met the deadlines needed. By assigning one person to write our weekly tasks, one person to communicate with marketing, another to work on analysis, we each stayed aware of what our personal task’s deadlines involved and therefore were better able to inform the entire team.

Development Areas

When it came to reflection, I learned that I have trouble taking command of the situation and while it is good to be horizontally structured, there were times I could have been more assertive in order to reduce inefficiency.

One thing we could have done better was analyzing what went right and what went wrong at the end of each week. This would better help us determine how we could better improve team dynamics and reduce wasteful activities. It would also give us the chance to determine based on overall team feedback what went right and where we should better focus ourselves.

I personally feel that such reflection on myself would also be beneficial. For instance, not showing up late to group meetings was a positive thing that helped team efficiency. Treating this project as a serious endeavor with other people counting on me was key. This helped build team rapport and kept everyone on track by always working together as a team rather than individually. While certain conflicts do arise, keeping the team together in one office setting not only kept everything professional, but on task as well.

But, not updating our Asana board reduced team clarity. While we followed many agile principles, “working together”, “giving members support and trust,” “face-to-face conversation” what was missing was the principle, “the team reflects on how to become more effective, then tunes and adjusts its behavior accordingly.”

As such, behaviors like working on outside projects during the team meetings could signal to others that it is ok to lose focus on their work as well. If we had taken the time to reflect weekly, we could catch and correct these counter-productive behaviors. Being a role model in my actions, like showing up on time, was a social cue to others on how the team should behave. When I would show up late or not work, reflecting back I realize it would also influence others. As a team member I should have been less lenient on being late or neglecting activities, and rather enforced a culture of taking responsibility.


(8) “12 Principles Behind the Agile Manifesto.” Agile Alliance, October 6, 2019.

(2) Ben Chu Economics Editor @Benchu_. “This Is What Nudge Theory Means — and Why You Should Care about It.” The Independent. Independent Digital News and Media, January 13, 2018.

(10) Cockburn, Andy. “Psychology of Referral: What Really Motivates Customers to Refer-a-Friend.” Psychology of referral: what really motivates customers to refer-a-friend. Accessed November 20, 2019.

(12) DesMarais, Christina. “3 Secrets to Getting Your Product to Go Viral.” Inc., December 7, 2016.

(6) Efti, Steli. “Why Referrals Are The Most Valuable Form Of Marketing (And How To Get More).” Forbes. Forbes Magazine, June 7, 2019.

(7) “Infographic — The State of Referral Marketing for 2017.” SaaSquatch, May 7, 2019.

(1)Layton, Mark C. “What’s Different about Agile Communication?” dummies. Accessed November 20, 2019.

(9) Moffett, Tina. “Drive Customer Obsession With Journey Analytics” Forresster, November 30, 2015.

(4) “Overheard at a TSA Checkpoint the Other Day.” Return to main page, July 12, 2011.

(5) Slyker, Karin. “Texas Tech Participates in ‘Economics of Loyalty’ Survey.” TTU. Accessed November 20, 2019.

(3) StackPath. Accessed November 20, 2019.

(11) “What Is Confirmation Bias?” Psychology Today. Sussex Publishers. Accessed November 21, 2019.

I write digital marketing strategies and trends to improve my own approach to customer acquisition. Visit to read my strategy research.