Customer Portfolio Management: An Interview with Michael D. Johnson, Marketing Department Chair, Wisconsin School of Business

MJ: Yes, but it also gets into what we mean by who and what is a customer experience officer and what do they do? What is it that's really appealing to people in the C-suite? We had this interesting conversation last week with a bunch of academics and practitioners on what actually happens. As one of our German colleagues summed up is, you do all this rigorous research on the importance of being market focused and taking a market orientation. And it often comes down to people at the highest levels saying, well, what do you think? And someone says, well, I think this. And I say, well that sounds good, let's do that. As opposed to really sort of knowing what drives, uh, the profitability of a firm.

I think one of our biggest frustrations today is we have all this research that shows the payback on customer satisfaction. It has all these impacts on business performance metrics based on literally 30 years of research, because for the first time we have time series data on customers that we didn't have before. So we're able to do all this research, but it's pretty lost. For example, one of the things I'm doing a project right now with Claes Fornell, who I mentioned, who's retired from the University of Michigan, and with my old colleague Gene Anderson, who is a dean at University of Pittsburgh in business school, showing the returns, the stock returns to firms with high customer satisfaction. And it's astounding. Market leaders in customer satisfaction, if you put a portfolio together, it beats the S&P benchmark by fourfold, over the last 20 years. I mean the average rate of return is really high.

Now, it doesn't work in all situations. Markets have to be functioning. You have to be in sort of normalized situation where firms are rewarded for treating customers well and punished if they don't treat them well, as opposed to, for example, what goes on in the pandemic. But it's been very hard to get that message across and again, we tend to want to iterate to simple solutions and what's best in the near term. At the end of the day, a market orientation focused on customers is a long term strategy. (9.50)

SS: Certainly with the emphasis today on customer experience, there's osmosis going on here and that thought is beginning to infiltrate the C-suite. But certainly that's a topic we're going to come back to in this conversation. I do want to pick up on your comment. We're going to again come back to satisfaction and loyalty a little later in this conversation. But I do want to point out that the American Customer Satisfaction Index has effectively been flatlining for a while. Is that largely because all the basic fixes have been done for the most part and that customer expectations keep increasing, therefore there will be always this gap between what customers expect out of the relationship and what they're actually experiencing?

MJ: Well, it actually took a big dip during the pandemic and came back. But you do see these periods of flatlining for exactly the reasons. One of the reasons you suggest is, people's expectations adapt. So one of the first studies we did years ago using similar data from Sweden, because we developed the Swedish index first, Claes Fornell was the leader on that. And we found that people have very adaptive expectations. As quantitative performance goes up, expectations adapt to that over time. So you really have to see sort of how things are changing in real time, but eventually they will adapt.

But I think the other thing that that flatlining indicates, because it's basically at the same level it was 10 years ago. There's other studies that have shown, for example, that just people are leaving a lot of money on the table, that again, the leaders in customer satisfaction find this to be very, very profitable. It's an overused example, but I love to teach it to my students, is to really get into the details of Amazon and how Jeff Bezos built the company. But he was obsessed with customers. So I suggested to one of our doctoral students the other day, you should look at the annual reports of Amazon over the last 15 - 20 years. And the content analysis would be interesting because annual reports year in, year out, when he was in charge, were focused on customers.

Now you see them straying a little bit from that. You see them a little bit more focused on employees and other things and policy and things that may be spreading their focus a little bit. That doesn't mean those things, other things aren't important. But I think their success can be traced back to, you know, Bezos had an unrelenting focus on customers...

SS: No question.

MJ: …so there are firms that get it right.

SS: I mean, they're still few and far between. I think the problem today is all of the oligopolies, the law of three that have formed in corporate America, frankly, around the world, where there's no pressure on - the telcos would be a good example, banks would be another - where there's nobody's nipping at their heels so they can get away with as much as they can get away with obviously. So I think that that part is a corporate thing. But again, that's a topic, you know, maybe I'm going to reserve for a little later on as well. I just want to ask you one other question. Just in terms of your academic experience today, talking to these kids. They've chosen marketing potentially as a career option. Are they anxious about that now as a career option? You consider the impact of AI and effectively could be wiping out all of those intern jobs that they used to graduate into. What are the kids' perspectives? I mean they're using ChatGPT probably every day in their studies. What's their attitude, what's their feeling today toward marketing as a career choice?

MJ: Well, I think what we see here at the Wisconsin School of Business is students looking at marketing as part of an overall training. A lot of double majors, we even have some triple majors. So they're looking at things like one of the very best students in my class last semester, I should say in the fall semester, is a double major in supply chain management and marketing. And he wants to see his experiences over both of those over time because they connect. And a lot of my own focus over the years on a customer orientation was by working from people who came from the supply chain channels management area because they wanted really good customer data to drive product and process change.

So, you know, the book you showed me with Anders Gustafsson(3)? Well, Anders was, uh, I joke with him, I say he's one of my recovering engineers, but he was a Swedish engineer who really transformed himself into a marketing professor. But he came to Michigan as a postdoc and spent every afternoon in my office and we started. He talked about it from a product part process change standpoint, and got me involved in some case studies, including one with Volvo. And we've been working together ever since. (14.48)

SS: Yeah, which is highlighted in that book. It's interesting. I interviewed Chris O' Hara, who's sort of in charge of customer data at SAP. And their big thing is integrating supply chain data with demand data. And he said that's the magic elixir. Because if, because right now there's a disconnect between demand levels and your ability to actually meet the demand. So he said it could solve a lot of problems if they managed to crack the code on that. So it's so interesting that you talked about the supply chain side of things.

So let's move into the theme of this podcast obviously, and of your book, which is the idea of customer portfolio management. You are a renowned expert on customer satisfaction. What led you to tackle this particular theme, make it the subject of this book? I mean, CPM has been around a lot, the concept has been around probably 20 years or so. What made you want to take this on as the focus of this particular book of yours?

MJ: Well, this goes back to a long partnership with my friend Fred Selnes(4), who's at the Norwegian Business School BI. I've known Fred for a long time and we've literally been working on this for 25 years. But you sort of have to go back to the early research on satisfaction and the importance of customer loyalty and its impacts on business performance. So these are a couple of stories that I told. We tell very early in the book. But one is I was teaching a seminar on the impacts of loyalty on performance and in an executive setting at the University of Michigan Business School, now called the Ross School of Business. And it was a very interesting discussion. So after this discussion, the executives all got into commenting on what I was talking about and one of the executives at a large U.S. Telecom company said, well, I'm convinced by Professor Johnson's lecture, we should increase satisfaction and lower churn, but I'm only going to want to lower it by 5%. So put this into the context, you know, 20 plus years ago when cell phone contracts were turning over every six months, people were just looking for the best deal. And the other executive sort of looked at the guy and said, well, didn't you just hear what he said? Shouldn't you lower it more than that? And he said, well, you got to understand what might happen if I do that. If I do, you know, I've got tens of millions of customers and tens of thousands of employees and hundreds of offices around the country. I wouldn't have a very big business.

Which sort of takes you to what we call the watertight bucket. So you can sort of go to a simple solution if you take the customer satisfaction loyalty idea too far and say, well, I'll be better off with a smaller watertight bucket of customers. So I used to teach with a guy named Chris Hart(5) who was at the Harvard Business School at the time. And Chris used to talk about these water watertight buckets. You want to plug the holes because replacing customers that you've lost with new customers is not profitable. Well, sure, but the question is, what holes are you going to plug? And the firms that did this ended up years later, you know, saying, we're going to focus on those 10 or 20% of the customers that are 80, 90% of our profit. Fine. But they wake up five years down the line with a smaller portfolio, a smaller business. And the telecom executive said, I'll have a smaller business if I try to push this too hard.

So then I started, you know, Fred and I started, Fred Selnes and I, we started thinking about this. We really need to put this satisfaction logic into the broader context of what a business is trying to accomplish. And that led to some work. A case study we did on a company called PanFish, now part of a much larger company called MOWI(6). In fact the chair of the board of his school at one point was the CEO of MOWI who was part of the integration that grew. But this company has about 25% of the global salmon and trout market. It's really huge. They had a problem which was they couldn't control the size and quality of a catch when they were farming salmon in the North Sea, in the fjords of Norway. So big fish farming company.

And so they started thinking about us and said, you know, we have really three different kinds of customers here. We have, you know, and the translations. This is where we got the language: Partners, friends and acquaintances. We have these partners who want these customized solutions but they're willing to pay more for it. You know, they need the best salmon in the catch because they were salmon smokers. And the salmon smoking mechanisms, the infrastructure by which they sort of smoke the salmon, needs a particular size salmon which happens to be the best tasting salmon. If a salmon is too big, it tastes old. If the salmon is too small, it doesn't have enough fat, it doesn't taste good. So they wanted the medium, medium sized one. And so PanFish started saying we could come up with a customized delivery system, supply chain, and product for these partner companies, which were sort of these salmon smokers.

We also have these companies that buy in regular large volumes. And these were the friends, they're good customers, they buy good volume, they buy regularly and they will buy much of the rest of the catch - a little bit bigger, a little bit smaller fish. The rest of the fish we just were going to sell, we can sell at a much lower price with fewer benefits attached to it, to the transactional customers which were the acquaintances.

So this got us sort of taking those two examples and putting them together and saying this is the way to start thinking about how to manage a whole portfolio of customers. We really start have to thinking about long term investments in relationships in customers. Because the transactional customers are not always poor investments. They can grow over time to become the friends and the partners. Some of them will always be acquaintances but they also help us spread our costs. So that's what got us thinking about the need to start studying customer focus, customer centricity, if you call it from a portfolio standpoint because it really gets into the broader goals of the firm. (21.44)

SS: In writing this book, did you see there was a gap in the market in terms of the literature that really hasn't connected the dots very effectively for corporate management to say this is how you ought to think about it, investing in a portfolio of customers? Because the concept of customer portfolio management has been around 20 years. It's never gone mainstream, which has always mystified me to the point is, it's very logical, right? It just makes sense, particularly in today's day and age where first party data is gold. Why do you think that's the case? And I think you referenced a little bit in the book that there's an orthodoxy to marketing, we stick with what we know. And so it becomes a bit hidebound and very difficult to absorb new streams of thought. And even though CPM has been around for a long time, it just doesn't, hasn't seemed to have penetrated, you know, the consciousness if you will, of most marketers - why do you think that is?

MJ: Well I think there is a uh, disconnect and we have been teaching marketing as if it's just needs based segmentation for 50 years and it's been slow to change and needs based marketing is still very well embedded in the framework. But I think it comes back to the complexity because Stephen, one of the first problems we encountered is this was a much more complex problem than we thought. It involves so many different variables of the firm including bringing their cost structure in. You know, so it builds on work in the corporate strategy as well including I would say especially Michael Porter's work on differentiation versus cost leadership. It's not one or the other - companies are doing both of them all the time. So it affects the fundamental cost structure uh, of the firm.

So the way we attack the problem, we started doing this, we said well we're going to do it with case studies and application, which build on everything a large community of scholars has built over time. There's a lot of people in this field, you know, you think of the work that Pete Fader has done on customer centricity, is very well embedded in this and the work on customer uh, lifetime value models which we also build on. So in the practical applications we use all of that. But it was so complex we decided we have had to take a different approach to understanding the nuances of portfolio management.

So there's a 2004 article in the Journal of Marketing that we published with a model called Customer Portfolio Lifetime Value. And as opposed to it being sort of an empirical model, a database model, it's a simulation model. It's straightforward, it's very similar to what finance people do when they're evaluating uh, real estate, or other type of investment. And they have, you know, 100 variables that they throw into a “what if” model, to see what valuations will be years ahead. So we found ourselves using that model and we started coming back to it three years ago and simulating more things and simulating different strategies. So I worked with Fred on this to simulate, well, what happens if people, if the focus is just on relationship conversion, if all you want to focus on is turning acquaintances into friends or friends into partners. Let's compare that to an offensive strategy because some would say, well, it's all about volume. Again, simple solution, it's all about volume. So let's just put, put as much water in the leaky bucket as possible.

And then we compare that to a strategy where now the focus is going to be more on defensive marketing, which is more of a focus on customer satisfaction than on, you know, growth per se. And the interesting thing we found is that defensive strategy is the most profitable strategy under most circumstances through these simulations - that it really was very consistent with what researchers were finding in the customer satisfaction area about all the ties to business performance. Because what it did is it not only grew the size of the portfolio, adding new acquaintances, converting, increasing the profit margins on customers, it sort of did what both a conversion strategy and a volume based strategy were both trying to do at the same time. I was able to build the whole portfolio. So that sort of showed us and led us to what was the, sort of, you know, the picture on the front of the book was there's more value in a large leaky bucket of customers than in a small, tight, watertight one. And that doesn't mean - you're still plugging holes in the bucket. You're still finding where, you know, applying resources to those customers who appreciate a more complex, a more customized, a more expensive solution, versus those that are the steady, you know, the partners versus the friends who are going to be the more steady customers, versus the acquaintance because they all help you spread in the short term, they all help you spread your costs. In the long term, if you don't continually put some water in the bucket and allow some of it to leak and some of it not to leak, you're not going to grow the portfolio. The portfolio will end up shrinking over time. One of the metaphors we use in the book is even a small water tight bucket, eventually the water evaporates out at the top. (27.42)

SS: Yeah, I like that concept of evaporation. So the idea of progressing the value of a customer over time, increasing the value of a customer over time is really at the heart of what you're talking about. I do want to throw in one contrarian note. Not from me, believe me, because I fundamentally believe in the concept of relationship management, investing over time, adjusting that investment based on expected value, et cetera. We'll come back to that subject too. Optimizing the value of the portfolio, which is language that you use.

So two things, one is, I haven't heard you mention profitability. So, in a book that was written a number of years ago now, about the concept of demon customers(7), which is customers who are a drain on profitability, who buy on price, who are brand switchers, is that a consideration here? That there are some customers that effectively, to use the expression, I don't like it, but the expression that some customers need to be fired simply because they're not good customers and have very little probability of increasing in value over time. And that was Larry Selden's book, which I think was written in 2003 when he made that case, business case, if you will. But it was largely on the idea of profitability as opposed to what you're talking about, which is really increasing revenue over time. Is that a fair comment?

MJ: Well, yes and no, because in one of the chapters of the book we get into sort of one of the approaches that have been around in time, and that's the need for predictive modelling, you know, in a customer satisfaction framework. And one of the most important things there, is satisfaction is an optimization problem. It's not a maximization problem. So there are optimal levels, and there are, those optimal levels vary by relationship segment. So there are customers. And we got a sense of this. I was doing a study with my colleague Chris Hart on hotels years ago, and we looked at, and this was interesting because it involved working for a brand company, as opposed to the owners. And there's a big difference between whether you own the hotel or you're the flag on the hotel. If you're the flag on the hotel, you want to maximize satisfaction because you're selling a brand, but if they're the owner, you want to optimize it. And we said, well, if there's going to be a payoff in satisfaction, it's going to be higher at hotels where there is a future revenue stream to capture, which basically in our application there's going to be business travelers, vacation travelers, who are going to go back to the same locations as opposed to the transactional customer, or transient customer, you know, who's passing through a small town on the only trip they're going to make from point A to point B in the next 20 years and happens to stay in a hotel. So a small market, rural hotels, you know, the optimization point is going to be much lower. And that's exactly what we found.

But in that satisfaction modelling, you have to connect satisfaction to a business performance metric that will show you that it is profitable. We do that, for example, and the American Customer Satisfaction Index research does that, by understanding that relationship between the loyalty metrics that we capture in that data and actual retention. But it is a profitability metric. And you go back to your comment about demon customers, you know, your relationship segmentation should also show you that you're not going to invest in some of those acquaintances customers the way you are in others.

So we did, um, I mentioned my other colleague Anders Gustafsson, who I've done a lot of research with. We did a big study for Telia which was a big telecom in Sweden. And we found a very interesting relationship between prior churn. We were using panel data, we had prior churn, future churn, and in the middle we had customer satisfaction surveys. So we were connecting this to actual retention. And we found this interesting interaction between prior churn and satisfaction, where it was the customers at the higher end of the satisfaction scale who you could increase with an investment, you could increase their loyalty even more and it would be very profitable. There were customers at the lower end of the satisfaction scale who are always just going to search on price. So you had to be very careful in your investment in those customers because they weren't going to pay off as much. So ultimately you have to connect your metrics to a business performance metric that ultimately gets to profitability. (32.43)

SS: Yeah, and to some extent the creation of loyalty programs were kind of designed to address that issue, to give somebody incentive to actually stick with the brand Even though they might have a marginal relationship, there was something in it for them to not to exit to the competitor. And I do want to come back to loyalty as a subject momentarily, but just for the audience here who may not have already read the book, obviously, maybe just provide your overview if you can. You've referenced partners, friends to some extent, but just provide a quick maybe matchbook cover definition of the relationship segments and then I want to understand how those are identified. So, just if you can, just a short definition of what each of the segments are.

MJ: Well, you can think about it in terms of how a relationship grows over time between buyers and sellers. Now they start out transactional and if you're a customer you're looking at different, you know, options to buy and they all seem the same. They're more or less commodities. You may not see any differentiation and you're going to buy mainly on price. I'm still that way for the most part when it comes to buying, for example petrol, if the gas station's on the correct side of the street going home and the price is attractive, I'll stop and fill up the car. But it's not that I'm particularly loyal to any particular brand. But as firms differentiate and provide greater value to the customer, they can start charging a higher price for that.

So it goes back to classic differentiation strategies. And this is where you can start to develop, you know, a friendship with a customer where you can charge a slightly higher price. The price elasticities are such that you can get more profit from those customers. So those are, so we called the first group the acquaintances. But you can create a friendship where someone is going to not just see you every once in a while, but will want to come back and visit you more often. But among those friends are also people who are going to appreciate these more customized solutions and in which, they're going to want to more connect with you, more in more ways, work with you, and to be honest Stephen, we got some of these ideas again out of the B2B relationships.

One of the things Fred and I did when he was with me in Michigan once is we went and visited suppliers in the Detroit area to the auto industry. And many of them were very connected to Ford and General Motors and Toyota and companies where they had highly customized solutions in which they had to invest a lot in the relationship. Then we started seeing that this can also happen in a business to consumer setting, is that, you have these large digital ecosystems now, where customers are connecting in multiple ways. We're working on this, Fred and I are working on a study now with another colleague where we look at this brand adaptation in these service ecosystems and they not only increase satisfaction but they increase the switching costs. So you can get very connected to the hip. So you're starting to see these B2C relationships that are not unlike some of the B2B relationships we've seen for many, many years.

So it's really about do you have a more standardized product that's bought mainly on price, which is the acquaintances. Do you have a more differentiated product that you can charge a little more for and customers are willing to pay, which are the friendships? Or you have a highly customized solution that someone's willing to invest in, that each side, the buyer and the seller, are willing to invest in to create a true partnership. (36.40)

SS: Well, and it comes back to value proposition. Some brands create flanker brands. In your book you do give the hotel example, Marriott I think, and it's interesting, their approach is effectively creating different experiences for different relationship segments within their base. I do want to understand something though. And so conceptually I get it. In terms of how you go about it, and I guess this varies by sector, by the amount of information you might have available, but just in terms of the segmentation variables that might be used. So I think if you're a company that collects a lot of transactional data, you might be able to create a segmentation around obviously current value, potential value as well, what's the growth potential. But there are other variables that obviously can be rolled into that as well. When you think about relationship segments, what are the types of variables that go into the stew mix, if you will?

MJ: Right. Well there's a whole list, including satisfaction and other things. But there's a couple of them that we have found particularly valuable. One is brand preference. So literally, if all things equal, would you prefer this brand to another brand? The other is share of wallet. So if you think of, you know, Tim Kenningham has done a lot of work, including work that links satisfaction to share of wallet. Where is the customer's purchases actually going, which you can get out of your CRM system, your customer relationship management systems.

So, one of our favourite case studies which we put in the book is about a home materials retailer, a Nordic company, that has these 30 stores. And we went to the management and really focused on, okay, how can we segment these customers into these three groups? And so, we used these two variables and it actually yielded four groups, but two of them were acquaintances. And so this is important because it shows you, you can have low volume acquaintances and high volume acquaintances. And the danger with high volume acquaintances is they can switch in a minute if they get a better price. But we basically use these two variables to come up with acquaintances, friends and partners. So we looked at brand preference with a survey and we looked at share of wallet out of the CRM system to develop these. So that was a fairly straightforward approach. But it also, that case also illustrates the importance of just doing the segmentation itself. Because once we had those segments, we did a simple plot, and there's a figure in the book, there's just a simple plot across all the stores in this chain of what the percentage of friends and partners was versus the percentage of acquaintances. And across the stores the difference was remarkable. There were stores with close to 90% friends and partners, which happened to be the most profitable stores in the chain. And there were stores with, you know, 12-13% friends and partners, and the rest were all acquaintances. These were stores with the exact same marketing programs. You know, then when we talked to management, it was more of, okay, so what's driving this? And there was a couple main things. But a big factor was store management. But another was what's the competition? What's the nearby competition? Are there competitors nearby that are better serving the needs of the customers?

So that gave, long story short, that gave the management team a lot of ammunition to go and say, how can we take the lessons from the best performing stores and translate them to the weaker performing stores? And it dramatically increased the profitability of the company. It's not unlike sort of, a hotel case I used to teach on Four Seasons hotels, which you can appreciate where they used to, if they opened up a new hotel in Rome, for example, they would had something called the sourdough starter. They would take the culture of the hotel, you know, in Paris or Toronto and transplant it in Rome and let it grow. So these are the kinds of ways that I like to teach it to my students, is that, you know, it's about borrowing best practices. But it starts with a simple segmentation scheme by relationship strength. (41.22)

SS: So the one thing that doesn't change is the framework, the labels, if you will. But you're suggesting that what, what will vary, is the variables themselves simply because of the nature of a company, the availability of data, et cetera. And you mentioned need segmentation earlier, still being crucial today, and obviously to identify opportunities within your existing customer base as well as in the broader marketplace. How do you see those two segmentation frameworks working in combination? This is a matter of creating the relationship segments and then overlaying need segments to see what the distribution of the need segments are against each relationship segment?

MJ: That's a great question. We get into it just a little bit near the end of the book in terms of where we think marketing is going. And it really should start in our view, with this relationship segmentation and underneath that would be the needs based segmentation. So if you think about it, you know, for those listeners out there who may be familiar with the house, of course, quality function deployment, it's how you translate customer needs into product and process design. There's an analogy there that you sort of need a matrix approach.

So you can think of in that matrix, you know, on the X and Y axes, one is going to be your segments organized first by relationship strength, acquaintances, friends and partners, and strangers who are future acquaintances, friends or partners. And then under that you're going to have different needs based segments and then on the other axis is going to be all your brands. And so it's really about mapping your brand strategy into that relationship strategy.

So if you have, you know, gold or platinum customers in Marriott's system, there's a range of hotels that they might use and this touches on needs based because, needs based is very dynamic. It changes over time and it changes by use occasion. If you're out with your kid's soccer team, you may stay at a Fairfield Inn. If you're going to a wedding, you may stay at a Ritz Carlton. It depends on the context. So the needs based needs to be sort of underneath the relationship based segmentation. But it's about mapping those two together that we think is where this is going. And I actually think some of the European schools have been sort of starting to teach it this way, you know, before some of the schools that are here in the U.S.

SS: It’s hugely insightful. We do a bit of that work here because we, for a client, any, we're sitting on a pile of transactional data. And what gets really interesting is when you do the valuation segmentation and then connect it back to the need segments and just see the disproportionate skew in numbers that line up to those various tiers. So it's a fascinating bucket of insights. It's interesting to hear you say that. It hasn't really, that hasn't really become a standard practice you would think ultimately it has to be. It's the path to proper value proposition development I would argue. Which takes me to, I wanted to cover off satisfaction and loyalty briefly before getting into strategic planning because you offer a model there relating back to what we were just talking about. But before we leave that subject of satisfaction, you do talk about a couple of concepts which is the idea of cumulative satisfaction, meaning I think over time as opposed to point in time. And then you also talk about the term impact performance analysis which I really found interesting, which is, I think correlating SAT to actual performance metric of some kind. Can you perhaps elaborate on those two concepts? (45.13)

MJ: It sort of gets into the evolution of our economies and service research. Let's talk first about transaction specific satisfaction and what we call cumulative satisfaction. Because this was really the main innovation in the work that Cornell did in Sweden and that we did as a team for the American Index. We changed the way satisfaction was measured. Transaction specific satisfaction is sort of the common surveys you get when you had the last trip on Air Canada or Delta Airlines, or last visit to a Marriott hotel, you know, how was that trip? And it's not that they're not valuable, they are valuable, but they're more valuable from a service process design standpoint. And these journey mapping, they're part of journey mapping. And these journey maps have become more popular as the digital economy has grown. Journey maps have been around for 30-40 years. Disney and SAS Airlines was doing this a long time ago. So the transaction specific measures are good at point in time measures and they can tell you where the kinks are, the moments of truth are in these process journeys that then you can then fix. So very good for service process design, but they're not going to tell you necessarily what the customer is going to do next.

So we started measuring some of our early work, looked at the economic psychology of customer satisfaction and said, what you really need to ask is overall, how satisfied are you to date with this product or service provider? And that's, you can think of it statistically, the statisticians out there, as a Bayesian updating function. It's Bayesian(8) statistics. You know, you have an impression, it gets updated over time, your expectations adapt, as we talked about earlier. And that's what's more likely to determine what customers are going to do next. You know, so that whole distinction is important depending on what you're trying to accomplish with the customer data. But when you get into predictive models, customer satisfaction models, so you can really understand the profitability of satisfaction, you should be using these cumulative measures where the American Index Measures, which are all publicly available, have become the standard for that kind of measurement.

The impact performance analysis comes out of the output of satisfaction models where you know what is actually driving that satisfaction. So we use a very generic model in the book. But you know, you have service quality attributes and benefits, you have product quality attributes and benefits, you have price and value, you have all sorts of different things to drive satisfaction. Well, what's the impact and what's the performance?

So part of satisfaction measurement and part of the book that Anders Gustafsson and I published back in 2000, is on this impact performance analysis where you're looking at, okay, to be profitable, I have to understand where I'm going to get the biggest return on investment. The biggest return on investment is likely to be those areas where the impact is high. It's going to have, you know, the big, if I make an improvement, an investment in this area, it's going to increase satisfaction, loyalty, profitability, and performance is low. These are my, if I go back to the competitive strategy sort of language, these are my competitive vulnerabilities. This is where I'm going to lose customers in the near term and long term.

In contrast, the high impact, high performance items tell me where my competitive advantages are. If I can sustain those over time, I have a more sustainable advantage. You know, those are areas that I may, you know, want to continue to invest in if I'm building a point of differentiation there. But they're not necessarily the things I need to invest in first. It's those vulnerabilities that I need to invest in first.

SS: Just so I understand, I think what you're saying is that you're relating each satisfaction attribute, if I may put it that way, to a performance measure which could be revenue. Like, you see a connection between high satisfaction and incremental revenue - is the dependent variable revenue?

MJ: Or you see service quality having a high impact, but a lower high performance level. So we want to separate these two and think about what they mean when you put them together.

SS: But any impact on overall satisfaction or impact on revenue? That was the part I wasn't quite getting.

MJ: Both because it's a causal chain.

SS: Gotcha. (49.55)

MJ: So we can get into this notion of causality. As I increase service quality, I'm going to increase overall satisfaction, which will ultimately increase profitability. So I need to understand that whole causal chain, but I need to sort of separate out, if I invest on something, what's the impact? How much effect is it going to have versus where am I? Where is the actual performance? So if the impact is high, but the performance is low, I better fix it. If the impact is high, but I'm actually doing pretty well on that, I may not have to worry about that right away. If the performance is high, but the impact is low, it can mean a lot of different things. It could mean that I have just over time driven the variance out of that problem, that process - everyone, including my competitors, perform equally well. You know, so it becomes a non-differentiator. It's what gets called a “must be” quality. So if you go back into the quality literature, there's something called the Kano(9) model, which I've always liked, which says ultimately, over time, a surprise and delight attribute becomes a performance attribute. It becomes an expected attribute.

SS: Table stakes, in other words, yeah.

MJ: Yeah, you drive the variance out of it. So you have to be careful. You don't want to take investment out of that. The interesting one though, when I'm teaching the impact performance analysis is the low impact, low performance one because it goes back to your profitability question. These are things if they're quality things, quality related variables. If something's low impact, I don't want to put good money after bad. I don't want to necessarily be, you know, investing in these things if it's not going to have a payback, if it's not going to have impact. So in one of the case studies in the book, we talk about, it's a tire retailer, independent tire retailer, and we talk about the tchotchkes in the store. Well, the tchotchkes in the store are great, but don't put too much money in them because there's no return on investment. Is it a quality tire and is it at the right price?

SS: No, every, every customer wants free stuff. But yeah, it doesn't, doesn't explain their brand choice. So that was helpful. I do want to talk about loyalty measurement because you, you actually don't mention NPS in the entire book, which is amazing.

MJ: Yeah.

SS: Was there a rationale behind that? And just with your overall, because there's SAT measurement and there's loyalty measurement, they're not necessarily one and the same, obviously. Can you just offer your perspective on that?

MJ: No, it's a good point. And again, we deal with it more in one of the earlier books which is sort of linking satisfaction to loyalty and profitability. The question is, are you investing? Are you measuring loyalty as a behaviour or are you measuring as an attitude? And that gets into context and cost and database availability and all of that. And there's been an evolution. A lot of the research over the years has used loyalty as a behavioural intention. And we've actually found good ways of translating, um, those behavioural intentions into actual loyalty. Because it's a nonlinear relationship. That what you actually see is the biggest impacts of satisfaction on retention at the highest levels of loyalty.

There's a guy named Rich Oliver whose research we build on. But Rich was, you know, he was actually one of the more senior guys when I was coming into the field unfortunately passed away a few years ago. But he talked about true loyalty, which exists at very high levels of satisfaction. And that's very much the case empirically. That's why you get such a high return for being a market leader in customer satisfaction. But when we can, Stephen, we usually sometimes avoid that step only because we have the actual data, we have the actual retention rather than the attitudinal loyalty. And that is with the emergence of panel data, and panel data studies where we can look at actual churn. That may be the only reason why we're not measuring loyalty as an attitude.

So but in this research we're doing right now on brand adaptation, we measured as an attitudinal measure to get started, to understand that. So it's still important. But if you can connect it to real behaviour and you can connect that real behaviour to actual profitability, that's what your causal chain and your predictive modelling should do.

SS: Sure. Well, we could have a whole conversation around this subject alone. For sure. I did have the privilege of interviewing Fred Reichheld when he came out with his latest book. And so the whole issue of NPS, some people view it as a SAT score, not a loyalty score. But I want to, I don't want to take us there because I want to. There's not a lot of time left and I need enough airtime to explore I thought one of the more intriguing aspects of the book, and there are many, but the one thing I really glommed on to, if you will, is your brand relationship market matrix. Because I think to some extent it cracks the code on this challenge between managing a product portfolio and category management and the whole concept of relationship management brings these two things together. Can you explain how your matrix works? (55.17)

MJ: Yeah, and we touched a little bit on that earlier in the sense that, you know, you can think about this as how I'm not just looking at a particular needs based segment and how that, you know, relates to development of a new brand. I'm looking at companies that now have large brand portfolios, for multiple needs based segments and how do I market those effectively to different relationship segments? And again, this is just, we're literally planting the seed in that book and starting to focus on that question now.

But how does a Marriott, for example, take its 36 brands and market them to six different segments in their loyalty program? How does a Delta airline, I think has four or five levels in their loyalty program and they don't have as many brands, but they get into constellations of different brands and partnerships with other service categories. So how do you market those? You know, so it sort of takes the problem of needs based segmentation and takes it to the next level of, you know, this market matrix, where you're taking a whole brand portfolio and mapping it into these basic relationship segments as a start. But admittedly we're still early in that work.

SS: And there's a whole set of planning tools and methods that will fall out of that. And it does crack the code, frankly, if you can figure it out. Because that's the chasm that exists today as far as I can see is, you've got product manager, market manager over on one side of the aisle wanting to market as many customers as they possibly can. On the other aisle are the folks in the customer experience department saying, hold on now, let's be a little more measured in who we're talking to and what we're talking to them about.

So I'll be anxious to see how far you can push that thinking because again, the whole idea of a marketing budget being zero based and really coming out of the analysis that you're talking about to me is at the heart of the new planning model down the road. But let me then push that a little bit further in terms of how you think marketing should be organized. Because go back to our conversation earlier about marketing being a bit hidebound. It's still largely product and channel based organizational structures. How do you think if you embrace the concept of relationship segments, maximizing the value of the customer portfolio over time as key precepts here, how do you see marketing organizing itself around that new model?

MJ: Well, I think in the simplest way it needs to be flatter, it needs to cut across. It can't be a silo in the organization. It needs to cut across everything from supply chain to finance. The CFO needs to understand how this is going to affect long term revenues for the organization, and the CEO, the supply chain people, the chief operations officer, they're going to need to know how what they're doing is going to be driven by what we're learning from a customer focus.

You know, so it really goes back. I'll tell you Stephen, one of the first slides I show any class or any executive seminar that I teach is a 71 year old quote from Peter Drucker which says, “marketing is basically looking at the entire organization from the perspective of the customer”. If you start siloing it, it's just going to break down uh, one of the very best literally. There's one sentence of the book I don't get into details on it but one of the best relationship segmentation schemes I've ever seen and best sort of predictive modelling I've seen in a company was at a large chemical international chemical company that developed great strategies for their partners, friends and acquaintances. And you know where the whole thing fell down. The sales force. The sales force was still selling on price. (59.39)

SS: Somehow that isn't altogether a surprise.

MJ: I had an interesting lunch discussion a couple weeks ago with two colleagues here which gets into the alignment of incentives. If marketing is going to be flatter in organizations, you know, incentives have to be carefully aligned so that they really do drive ultimately the performance for the whole organization.

SS: You know to some extent in this conversation, I have one more question to ask but uh, to some extent we've answered the question I started with which basically, what's the future of marketing? And the answer is the future of marketing is related to customers and somebody's got to speak for the customer. So that's got to be marketing's job. But I do want to close out our conversation today and we could have taken this many other directions because it's a very rich topic, certainly I will add, deserving of a follow up book to this one. And this is back to the AI induced job disruption that's ahead. So, is marketing going to be called something different going forward that will put it more at the center of corporate strategy than it is today? Do you see that dramatic an evolution in the cards, over the next say, five years?

I mean I'll just go back to one other comment that OpenAI CEO Sam Altman predicted - this is not too long ago - that when AGI arise, artificial general intelligence arise, pick your time frame, 5, 7, 10 years, it will wipe out 95% of marketing is his prediction could potentially even move into strategy. So is the salvation of marketing here really to embrace the core principles that you've outlined in this book because effectively that becomes the strategy of the entire business, not just marketing.

MJ: You may have said it better than I. I think, when I think of AI in this context, it's going to be what I sort of look forward to is the transition from GenAI to predictive AI because then we're going to what used to take us weeks or months in order to get the data, analyze it, you know, structural equation modeling and really understand what the drivers of satisfaction are. How a predictive model will then predict loyalty and retention and profitability after that. That takes a lot of data, a lot of time. Eventually you're going to hit a button, you're going to have to replace some of that data, you know, in real time. And that data doesn't always update in real time. But it's going to change things dramatically.

But today, you know, we get into the framework of the book, into the need to combine really good data with predictive modeling and causality. And there's a professor I was chatting with last week from Yale, his name is K. Sudhir and he told all these young researchers, he said, avoid the temptation of taking your models, your predictive models and inferring causality in them. Because AI still has, now I'm going to speak for myself. AI still has challenges separating true score from error, statistically real information from error and, you know, in causality. So it's going to be researchers who continue to understand how customers behave, what drives their behaviour, what drives customer satisfaction, which involves quantitative and qualitative research. People like you and I doing this research, combining with those tools where it's going to become really, really powerful.

But without the notions of the causality, you're going to hit a predictive model and not be able to know, okay, well how do we fix it? How do we actually improve the prediction? Or how do you go, you see these, as I told my students, you see these ads on TV or what's the probability of the football player catching the touchdown pass in the end zone and it was only like point 0 something percent and they caught it. I go, well that's great to know, but how does that relate back to, you know, both the training of the receiver and the training of the quarterback and the offensive line that protected him in terms of driving it back into this notion of product and process change that I mentioned earlier. So that's still going to be there. So it is sort of an exciting but scary world ahead. But you know, marketing's not going to go away. But we need to think of it as a customer market orientation.

SS: Yeah, for sure. And after 20 years of talking about it, finally maybe it will actually happen. Well, your book will certainly, I think, contribute to the thinking for sure. As I said, a book like that is overdue because of the end to end aspect of that, we do get caught in our little specialty disciplines and whether it's database marketing versus experience management, et cetera. So the book is a very refreshing take on all of it and hopefully you'll extend that thinking into, well great, now, here's a planning model that we can all embrace that will help us realize this ambition to be customer first. So I want to thank you for your time today, Michael. It's an absolute joy talking to somebody who is obviously very steeped in knowledge about all of this - SAT loyalty and obviously customer portfolio management. So it's been a delight.

MJ: Well, I appreciate it. It's been a pleasure.

SS: Yeah. And I'll look forward to, uh, your next book, for sure.

MJ: Let's have another conversation.

SS: We definitely will have another conversation, that's for sure. Things are moving at such light speed right now. It's just unbelievable. So a year from now, it'll be another very interesting conversation, I'm sure. So thank you, Michael, very much appreciate it.

MJ: You're welcome.

That concludes my interview with Michael Johnson. As we learned, marketers continue to overinvest in the acquisition of new customers. Some also make the mistake of focusing excessively on the most valuable customers for fear of losing them. Instead marketers should just accept the fact that it is impossible to plug all of the holes in the customer “bucket” since organic churn is unavoidable – even amongst seemingly loyal customers. Instead marketers should selectively cultivate closer relationships with those customers who have the greatest potential to increase their “share of wallet”, using lifetime value modeling to estimate incremental revenue growth. They can then progressively grow the relationship with each segment over time – converting strangers into acquaintances, acquaintances into friends, and friends into partners.

1 - Berry introduced the term in a paper titled "Relationship Marketing" in a presentation to the American Marketing Association’s Services Marketing Conference.

2 - Roland Rust, Valarie Zeithaml, and Katherine Lemon are credited with coining the term customer equity in their book “Driving Customer Equity” (2000)..

3 - Anders Gustafsson is Distinguished Professorial Fellow at the University of Manchester's Alliance. Fred Selnes is professor at BI Norwegian Business School. He is co-author of the books “Customer Portfolio Management” and “Marketing Management: A Customer Centric Approach”.

4 - Chris Hart is an adjunct professor of marketing on the executive-education faculty at the University of Michigan's Ross School of Business.

5 - Mowi, known as Pan Fish prior to 2007, is a Norwegian seafood company with operations in a number of countries.

6 - The term "demon customers" was coined by Larry Selden and Geoff Colvin in their book "Angel Customers & Demon Customers". It means a segment of customers who are unprofitable and can even lose a company

7 - Bayesian statistics is a method of statistical inference that updates the probability of a hypothesis as more evidence or information becomes available.

8 - The Kano model analyzes customer preferences and prioritizes product features based on their impact on customer satisfaction.

Stephen Shaw is the Chief Strategy Officer of Kenna, a marketing solutions provider specializing in delivering a more unified customer experience. He is also the host of the Customer First Thinking podcast. Stephen can be reached via e-mail at sshaw@kenna.