Maximizing Customer Value Through Segmentation
Customer segmentation allows us to create groups of customers that behave similarly in response to marketing treatments. By marketing to homogeneous segments, we can customize the persuasion process and maximize the lifetime value of customers.
This is segment two of our talk about database marketing and this segment we’re going to focus on Customer Lifetime Value Maximization through Segmentation.
What follows is a lightly edited transcript of Episode 2 of the Inevitable Success Podcast with Damian Bergamaschi and special guest Gary Beck. (Listen Here)
Damian: To drive game-changing growth marketers must be focused on understanding their customer. Today, resident expert Gary Beck shows us how.
All right, glad to have you back Gary.
Gary: Thanks for having me, Damian.
Damian: This segment we’re going to focus around is getting to the root of customers, how to understand the customers in your database, and what you could do once you know who they are. The very first thing and we hear this a lot when we sit down with clients. Being in the database space for BuyerGenomics brands don’t know, even big ones, don’t know how many customers they have. They think they do when they sit and think about it they don’t. Have you ever experienced that?
Gary: I have and it’s a great point. One of the fundamental challenges that many companies have is knowing the quality of the data that they have and being able to, to your point, count the number of customers – so, how many customers do I have? If there are data hygiene issues, there can be a real challenge in understanding what the customer base looks like.
One of the first things that many companies will do when they ask that question, how many customers do I have, is they will first clean up their database. That will address issues such as misspellings in a customer’s name or incorrect address information that may not have been put into the system correctly. That cleanup process is usually one of the very first things that we will do.
There’s also a fundamental question of, how many households are on the database?
Damian: Well that’s a great point. Could you go into detail more about that, what is a household versus a customer?
Gary: Of course. So, when we look at customers we frequently will find customers who are, let’s say, a man and wife or a couple of some sort.
They will use one email address for an account to conduct transactions yet we have two different customers in the universe. They will frequently show up as two different customers in a retail database that has not been processed through a traditional name and address hygiene processing system. So, one of the first steps is to make that decision concerning counting individuals or counting households, or counting accounts. There are implications in terms of how we market to people after that. For example, many traditional mailers will seek to only send one expensive catalog mailing per household if we suspect that everybody in the household will share that collateral.
Typically, there’s really no need to incur the additional expense associated with multiple catalogs to the same household. So traditionally catalog marketers have marketed based on a household basis. Digital marketers who can see different behaviors by the individual might choose to organize their database based on individuals, giving them the added leverage of understanding how to communicate to one person of a household versus another person of a household who might be purchasing different products.
Damian: I can totally see the value of not getting the same catalog to your mailbox. That ultimately probably it’s really one person that householder’s who was driving that purchase decision. At least I know that’s how it is in my household and I’m not the purchasing decision maker on most of those things. Some people may be very familiar with you know how you would go and dedupe let’s say your transaction database to a customer. So, you say that wow this person I have 10 transactions but it’s five customers. You also have this other process of oh I have you know 100,000 e-mail addresses and some of those are from the same person. Then you must go one step further to say now that I deduped my transactions on my e-mails to individuals. How do dedupe it to the household? How do you even go about doing that?
Gary: Great question, the intelligence that has been built into matching algorithms is one that allows us to look at multiple factors that tie database records together. So, for example, if John Smith and Mary Jones happened to be living in the same household. We might be able to tell that based on the physical address that’s perhaps the easiest. We might be able to tell that since they use one email address for any transactions that they conduct with us. We might be able to tie that information together based on telephone numbers that they have provided as part of the transaction process. So, there are multiple ways of linking individual records together and the software that’s available today gives us options in terms of how to combine those records.
Damian: I guess you know you are growing a certain area, once you do that you know you’ve figured out for example how many households have this how many customers I have. What do you do next with that type of information, now that you know these are customers?
Gary: Once we have these customers identified, the next question that we have is one of, all right, are these customers created equally and should we be treating them the same way? So, for example, if we know that one customer is a very high-value customer versus another customer that has been less valuable to us in the past, you would likely treat them differently. Similarly, somebody who might have a future value with us that’s very high versus somebody who appears now to have a low likelihood of purchasing with us again, how do we treat those customers differently?
Damian: I think you’re hitting on one of the more valuable points of kind of deduping all your silos into one central customer database and that is what’s the value of a customer. Do certain groups have higher values than others? You mentioned future value. How do you even begin to come up with a way of thinking about what that could be?
Gary: Great question. There are many ways of assigning values to customers and that’s something that I think we can knock around a little bit here. You know one of the simplest ways for companies to assign value is based on what a customer has purchased from you in the past. So, something that is commonly referred to as historical value. If we look at all your previous transactions, we can add up all the revenues that we receive from you and that’s your historical value.
Customer Relationship Management
Gary: Of course historical value may or may not be an indication of future value. So, the next question that one might ask is knowing what your historical value is what percentage of your purchases within this category have been with my company versus my competitors. In other words, what is your share of wallet with this customer and immediately what that does for you is it gives you some sense of loyalty. It also gives you some sense of potential value. So, if I’m only getting 25 percent of your business I know that 75 percent is still available. So, how do I go about getting that if I know that potential is there?
Damian: So the concept of you know the share of wallet. Sounds like that’s a piece of data that I think every marketer kind of wants to know about you know their customers. My sense is that that data does not or couldn’t exist in your own database or even your own first-party data that you have, you’d probably have to have some competitive data set. How do you do that, is that something that is inferred?
Gary: There are a couple of ways of acquiring this information. One is to simply survey the customer and ask them.
Damian: So you could get it yourself in that case, right?
Gary: So you could communicate, develop a relationship with the customer, and acquire that information yourself.
If the customer volunteers that information, that’s powerful information, it’s powerful in the sense that they’ve taken the initiative to talk with you, to spend the time to tell you that information and the odds are pretty good that that information is accurate. So, that’s a great source. There are also other ways of getting that information. There are syndicated data sources that will survey customers and identify how different groups of customers typically consume your products. So, by looking at the snapshot of what your customer looks like and comparing it to the snapshot of what syndicated data tells us that customers like you are consuming, it gives us the ability to estimate potential within the category.
Damian: All right so now that you’ve kind of had a look at the potential of your existing customers what are other different ways to kind of slice and dice your customer base? There are lots of different dimensions that customers have. What are some of the ways that you think about it? I think we can kind of share some experiences here and how we look at customer databases.
Gary: Sure, there are many ways of thinking about customers and it really varies by type of industry as to what seems to be the best way of segmenting customers for the kinds of communications that work for your business. One of the most common segmentation schemes is new customers versus existing customers. So, for new customers, for example, we typically have or we will frequently see an onboarding program that helps begin to set the relationship on the right foot with a new customer, educate them about the benefits of our company, and our products and services, and guides them to help make the relationship that makes sense for them.
It’s a process of creating a relationship that will work for that customer and set the stage hoping for a relationship that maximizes the lifetime value of that customer over time. So, new versus existing is a very common starting point. Once we have an existing customer, we will see many companies segment based on value.
Gary: The example that I like to talk about is the frequent flyer programs of the major airlines. If we look at the frequent flyer programs there are multiple levels; we can think about platinum, gold, and regular customers. One of the things that is interesting about this segmentation scheme is that the biggest rewards are given to the most frequent flyers.. A platinum customer receives more bonus miles for flights than the regular customer does for example. The platinum customer also can upgrade and receive many other services that regular customers do not.
What that does is it does create loyalty for the existing frequent flyers. Why would I want to lose points or why would I want to fly another airline where I don’t have status because I won’t be receiving the same level of miles and benefits in return. I just feel good about being recognized as a platinum customer. So that platinum, gold, and regular customer model has proven to be very effective for the airlines. It’s also a way of segmenting services. As a regular customer, if I call American Airlines, for example, I may receive an automated answering service and I may find myself on hold for five minutes or some period. As a platinum customer, a human being will frequently pick up or the automated service will give you the option for a human being and you are bumped higher in terms of having that kind of interaction.
Damian: Yeah, I think even in the e-commerce space there are probably some ways to apply that you know in the click to chat or even these concierge shopping experiences. I could see, for example, we know that this person visiting the site falls into one of those cohorts of extreme value or it’s a very specific customer type like an influencer. Everybody else gets the chatbot but that person gets the hotline right to the best curator of product there, or vice versa, or personal buyer. It’s an interesting topic and you know I think things like that really stimulate the people who are in marketing teams trying to figure out how do I get value from that data point like a little thing like that can make a big difference.
Gary: Absolutely and in a world of finite resources we cannot provide a personal concierge to everybody.
Damian: That was deep Gary. I couldn’t agree more.
Gary: Thank you, Damian. I am trying to provide deep insights.
Damian: Yes. So, you know it also reminds me of. So, you were talking a lot about I think like loyalty programs and there’s a book I’ll remember what it is in the show notes and I’ll put it in there but it was basically about the principles of persuasion. Loyalty kind of hits on one of them which is this concept of like reciprocity that you know I do something for you. There’s a real economic benefit in the loyalty points. But there’s also this other persuasion point where it’s like a buyback. You’re doing something for somebody and they’re more likely to kind of you know stick with you. To basically return that. There’s one story in there that I thought was interesting and I’ve seen a couple of brands starting to knock around this concept. It’s an un-loyalty program. So, in the context of a loyalty program, you’re keeping track of everything. Both parties are very aware that there is no loyalty system, that it accumulates points. But the un-loyalty concept is well if you have all this data you can still use it but not in a traditional loyalty sense.
So, for example, if you went to like a coffee shop and they give you a punch card and every time you went there they put a hole punch in it. On the 10th one you got a free coffee. Both parties there were in a loyalty relationship through a program. But what if you knew that this was their 10th visit and you just spontaneously said hey this coffee is on us. Thank you. In that way, it’s kind of organic. It’s a surprise and because we’re marketing data we’re marketing to humans using that data to get that extra mileage out of this could be powerful. Because how you do something is could be just as important or at least the multiplier to what you do.
There’s a segment from that study where they talk about a waiter who handed out the check in the control where they didn’t give any mints to anybody. Then they said they looked at what the tip size was for the control. Then in the first cell, the waiter would drop the check and leave two mints on the table. In that group, they didn’t mention anything but they would get about 3 percent higher tips. So, it’s already better. In the next group, they did something like four mints and then it went up a little bit more but not much. In the final test group, what they did is they would drop the check with the mints and then come back two minutes later and say hey do you guys want some more mints I got some more for you. It was just kind of surprise and it was like a 20 percent lift in in in the tip size. It was not just doing something it was how they did it. It was like this multiplier effect. You know you got a little bit of a bump from giving mints to loyal customers as a thank you. But not nearly as much with how you did it.
Gary: I think that’s a great story. The biggest opportunity for marketers is finding ways of engaging customers. When you can surprise and delight customers, that is just awesome. Because, there will always be that spillover effect where customers will tell their friends. I think the value of word-of-mouth, the value of it creating those scenarios where customers become advocates, is something that we always strive for. One of the things that we’ve noticed in just about every industry is that when we create engagement with our firms, with our companies, that only good things happen as a result, assuming that engagement is a positive engagement of course. But when we have positive engagement we can correlate that with increased purchases and increased revenues.
In packaged goods, for example, when we see customers downloading coupons, it’s a form of engagement and that form of engagement translates into higher revenues. We have seen that. So how do we create those types of opportunities for engagement? How do we surprise and delight customers in a way that they spread the word, become our advocates? All of those are opportunities for us as marketers, it’s something that we are constantly striving to identify.
Damian: There are so many different dimensions of how you can categorize you know customers and some of them you can kind of combine with each other. One that I know that you know here that really comes up a lot is the concept of your buyer lifecycle. So, kind of how you mentioned you know you could be a new or an existing customer. Now that you’re an existing customer you’re going to have a purchased cadence should you have a loyalty to us. Knowing where you’re at in that cycle whether you’re in the market. recently purchased. or feeding, and soon to be at risk of loss, and inactive to never purchase again. Statistically knowing where you’re at in that cycle is kind of a dimension you could combine with value, or what cluster you’re in, or what kind of type of customer you are.
The Buyer LifeCycle
Damian: Some examples are, you have a key influencer that’s a customer type and they’re in-market and you know that you can treat them special and they have a lot of power in your brand. Maybe you have your most valuable customers as a group and you know that this month three or four of them reached that statistical envelope that said hey they’re at risk if we don’t save them and they have very high historical and future values if we do. Do you have any experiences or stories to that maybe relate to combining different dimensions not necessarily the lifecycle but just different dimensions like of slicing your customer database up?
Gary: The buyer lifecycle is a very powerful tool and understanding when customers are on the verge of leaving the company in some way is intelligence that all companies can benefit from. I did some work with a cellular service provider and one of the things that they were successful in doing is something that we referred to as attrition modeling. So basically, when somebody was about to leave the company we could watch patterns and understand when those customers were potentially at risk, suggesting a preemptive strategy. So, the question is, once I’ve identified that a customer is likely to change cellular services, for example, is there something that the company can do to keep you as a customer
Damian: I’m curious when you were coming up with I’m guessing a model for what was predicted that they’re likely to leave I could see time-based being something. Were there other factors that you guys could track or monitor that were predictive or did you find that some sort of time horizon was helpful?
Gary: A time horizon, particularly for the cellular business, was one of the most significant variables and it was also the ability to just look at how a price plan compared to other price plans in the marketplace. So, if a customer had signed up for a price plan that was less than favorable for them vis-a-vis what the competition, you know obviously that customer could be at risk. So, what this provider did was a proactive outbound telemarketing campaign where they would call their customers based on their perception of risk, their perception that a customer was about to fade and they would ask the customer “how is my company doing?” “Are you happy with your cellular service?” and frequently for this modeled universe, the answer was “funny that you should call, I’m really unhappy. I happen to know that I can get cheaper service with your competitor.” The well-trained customer service person would then respond, “That’s exactly why I’m calling. I want to let you know that we have a new price plan that I think you’re going to appreciate.” And she would go into the details about how the new price plan was going to make a difference. The number of customers that were saved because of this program was extraordinary.
Damian: Yeah, I know that sounds amazing.
Gary: Having the intelligence into when a customer is at risk and looking at other customer behaviors to inform those insights is really one of the benefits of database marketing; by understanding relationships, and keeping your finger on the pulse of transactions in the marketplace, we can improve marketing programs.
Damian: So one of the things that struck me it’s very interesting about that story is if you’re that brand they executed or had a significant amount of self-awareness and honesty about the value that they were bringing versus what was out there in the marketplace. Then looking at the segment of customers that there is not as strong a relative value. You must be honest with yourself. I think sometimes brands they think of themselves as the only option and in reality, is the customer has many options. I think just going that step of saying “Ok we’re comfortable looking at what’s the value that we’re bringing and what’s the value in the marketplace” and finding the cohort or group that has the biggest dislocation there. Then giving them extra attention at the right time and then having a plan to address it when the customer does sense that that is an issue. It’s interesting and I think that’s a great skill for a brand. So, you either learn or adopt.
Gary: Absolutely and a little bit more background for this example might be worth sharing here. Our cellular provider suddenly had started to lose share. So, they saw that they were losing customers at a rate that was not acceptable long-term. They were forced to be introspective. They were forced to ask themselves the question “Why are we losing customers?” While that answer might seem obvious, “I’m losing because somebody else is offering a comparable product or possibly even a better product at a lower price.” They were put in that position of asking themselves the question of, how do I stem the losses? There were really two components to that, one is keeping your finger on the pulse of your customer base and when you begin to see an outflow of customers can you identify who they are, and second, what are the marketing measures you’re willing to enact to stem the tide.
Damian: Well Gary thanks again. This is incredibly insightful and I hope that you know all our listeners learn as much about how to think about who’s in their customer database and how they can be kind of grouped together and segmented as I did. We’ll have you again for the next episode.
Gary: Damian thanks for having me. It’s a pleasure.
Damian: If you enjoy today’s episode we ask you please leave a rating and write a review. Or better yet share with another marketer. Be sure to subscribe to the podcast for new episodes. Also, check out the show description for complete show notes and links to all resources covered in today’s episode.
Damian: If you’d like to speak to someone about any topics covered in today’s episode please visit buyergenomics.com and start a chat with the BuyerGenomics team Today.
*Book referenced: The 6 Principles of Persuasion by Dr. Robert Cialdini
Host: Damian Bergamaschi
Special Guest: Gary Beck
Gary’s background includes over 30 years of analytics & database innovation for several leading Fortune 500 companies and Madison Avenue advertising agencies. Gary has been a frequent lecturer and author on the topics of database marketing and applied statistics. His articles have been published in DM News, Direct Marketing and the Journal of Direct Marketing. He recently was President of the Direct Marketing Idea Exchange and currently serves on their Board. Gary received his M.S. in Industrial Administration from Carnegie Mellon University.