Is your organization customer-centric? Does your product team dive into the demographics of your customers to figure out what features will make them as happy as possible? If so, then you’re doing it all wrong! Perhaps. On this episode, the gang chats with Dr. Peter Fader (@faderp) from The Wharton School and Zodiac, about putting customer lifetime value (CLV) front and center when it comes to developing and executing marketing strategies.
References Made During the Show
- (Book) Customer Centricity: Focus on the Right Customers for Strategic Advantage
- Customer Lifetime Value (CLV or CLTV or LCV or LTV)
- (Book) How Brands Grow: What Marketers Don’t Know by Byron Sharp
- The NPS Loyalty Forum
- Angela Ahrendts (Apple)
- Fred Reichheld and the Net Promoter System (NPS)
- CPG / FMCG
- Jeff Bezos
- Atlas by Facebook
- Electronic Arts (EA)
- Net Present Value (NPV)
- Customer Centricity Simulation
- gmailr R Package
- Women in Digital
- Alaina Shearer
0:00:04 Announcer: Welcome to the Digital Analytics Power Hour. Tim, Micheal, Moe, and the occasional guest discussing digital analytics issues of the day. Find them on Facebook at Facebook.com/analyticshour, and their website analyticshour.io. And now, the Digital Analytics Power Hour.
0:00:28 Michael Helbling: Hi everyone. Welcome to the Digital Analytics Power Hour, this is Episode 70. Do you believe in life after love, or do you believe in loving a brand for life? Or if you are a brand how do you measure the love that your customers have for you? This is all a round about way of asking, how do you think about customer lifetime value, and maybe more importantly, how do build your organization into one with Customer Centricity? I’m pretty sure, in all of that we’re going to find some analytics to talk about as well. As always, I am joined by my co-host, Moe Kiss. Hello Moe.
0:01:10 Moe Kiss: Hi, Michael, how you going?
0:01:12 MH: I’m going great. Wait, what’s the appropriate going greeting?
0:01:18 MK: I feel like one of these days, I need to come in with “Good day, mate.”
0:01:21 MH: Well, no, “How you going”, that’s not a typical United States greeting so I’m gonna attribute that to the Australianess.
0:01:28 MK: Sure.
0:01:29 MH: We’ll work on it later. All right. And Tim Wilson, welcome back to the show, Tim.
0:01:34 Tim Wilson: Where you doing?
0:01:37 MH: What are you doing? Why are you here? And I’m Michael Helbling, and to help bring a little light to this topic, we went straight to the top of the food chain to find our guest. He is, Dr. Peter Fader. He’s one of the world’s foremost thought leaders in customer lifetime value. His research and mathematical models are valued by both startups and Fortune 500 companies. He’s a graduate of MIT, he’s a professor at the Wharton School. He has more than 25 years of experience as a consultant to many of the world’s leading brands across many industries. He’s also the author of Customer Centricity: Focus on the Right Customers for Strategic Advantage. Welcome to the show, Pete.
0:02:22 Dr. Peter Fader: It’s great to be with you and I guess I’m obligated to give you the Philly greeting, how you doin’?
0:02:29 MH: How you doin’? See.
0:02:31 DF: That one I’m familiar with.
0:02:33 MH: I’ve actually traveled enough to New York City and Philadelphia areas, like I can pull off a little bit of it, and I’m walkin’ here.
0:02:39 TW: And you just equated New York and Philly, so I’m sure you just alienated a [0:02:42] ____.
0:02:43 MH: It’s all the same thing. Boston, New York, Philly. Basically.
0:02:48 DF: Not to us.
0:02:50 MH: Yeah. We’ll get some comments about that probably. [chuckle] All right. So, let’s get things going. So, I think maybe the first thing that would be really great for us to understand or talk to about our listeners is, sort of, the definition of Customer Centricity.
0:03:05 DF: Sure thing. ‘Cause it’s not what the listeners think. And I think it was a mistake on my part to use that as the title of the book, and the label overarching of everything I do, ’cause to a lot of people, Customer Centricity means, “Super duper customer service, the customer’s always right, the customer is at the center of everything we do.” And no, that’s not what I’m talking about at all. I think that’s naïve. I think that’s disrespectful both to the managers within the company as well as the customers. It can’t be implemented. We have to acknowledge that not all customers are created equal. That’s just the reality of it. Some of them are really good, some of them are “Eh”. In fact a lot of them are “eh.” And so our ability to understand those differences and leverage them. And in fact in particular, to put, I hate to say it but I mean it, disproportionate attention on the really good ones, that’s Customer Centricity. So it’s not just being centered around the customer, it’s being centered around the right customers, while at the same time being mindful of those others out there that might not constitute a lot of revenue on an individual basis, but collectively they’re still pretty important to us.
0:04:20 MH: Yeah. I love that ’cause picking the right customer actually defines you as a business too.
0:04:26 DF: It really does. And the same ways in the old days, it used to be “Are we gonna be all about the product or are we gonna be all about efficiency?” In this new era it’s gonna be which customers are we gonna be all about? And then the products that we develop, and the service that we offer, and all that stuff is auxiliary to that choice of who those focal customers are.
0:04:47 MK: On that point, I guess, I’m interested to hear a little bit about if you’re focusing on the customers that are the highest value, “your good customers.” Are you potentially missing some learnings about what leads to a bad customer experience, so that you can then fix those things as well?
0:05:06 DF: Yeah. I’m glad you’re putting that upfront ’cause sometimes this general point doesn’t come up in a conversation, or sometimes it’s at the very end. I seem to be laying out this black and white, good and bad customer dichotomy and I don’t mean to do that at all. I am serious about putting disproportionate amounts of attention resources on the good customers, but that doesn’t mean it’s zero for the other customers. And for most businesses, like I said before, if it’s an 80/20 rule then those 80% of the so-so customers are collectively still important. So, maybe you’re not gonna be quite as responsive to them. Maybe when you’re developing new products and services you don’t wanna take their pulse quite as much as you wanna really focus in on those good customers. But you do want to know what’s up with them, and you do want to know if there’s anything systematic that’s gonna chase a lot of them away. And if there are some cost effective ways to keep them happy, I think it’s worth doing that. Its just that, you don’t wannna prioritize those so-so customers over the really valuable ones.
0:06:07 DF: Somewhere in the book and in all this stuff that I talk about, I discuss the paradox of customer centricity, and it kinda goes like this. The more that we zoom in on those really valuable customers, the more that we really have that insight and the ability to know who they are. In a weird way, the more that we need those so-so customers, just to have stability, just to have predictability, just to keep the lights on in the business. So we kinda need them, we need to be having a cost effective conversation with them. We’re happy to have relations with them, but it’s just gonna be a bit more on our terms than theirs, when we compare them to the more valuable customers.
0:06:46 MK: So if you were making decisions about a product roadmap or a marketing initiative, or say a loyalty program and you up weighed I suppose feedback from those good customers, is the goal about, by focusing on them to increase the overall number of good customers? Because I noticed in one of the articles that you’ve written, you mentioned about not wasting time trying to move people that have had a bad experience to making them have a good experience. But how do you increase the number of good customers if you’re not at some point focusing on these bad customers, whatever, quote, unquote, didn’t have a great experience and these are the reasons why, and this will help us get more of those high value customers?
0:07:35 DF: Sure thanks. There’s actually two really good points there. One is this quantity versus quality trade off, and then I wanna go back to what you said the outset about the product road map. But first it’s kinda strange but we need… Look, acquiring customers is rarely throwing spears and catching whales. Usually it’s throwing nets out there, recognizing that most of the customers we get are so-so ones. We just hope that we can get a better yield of better customers every time we throw the nets. It’s just getting a little bit better about how we acquire customers, but it’s still gonna be a quantity game and again most of the customers, not only that we have today but even the ones that we’re gonna acquire tomorrow are gonna be merely so-so. That’s the reality of it, and I think it’s important to recognize that we’re still throwing nets pretty broadly, even when we get real smart because it’s gonna be in most cases impossible to know exactly where those good customers reside.
0:08:37 DF: So for instance, down there in Australia there’s a lot of noise being made by my friend Byron Sharp about how brands grow and how important it is to increase penetration for a company. I agree with that, but for a different kind of reason. For me penetration, having a broad customer base isn’t an end unto itself, but the broader we’re out there searching for the right kinds of customers, the easier it’s gonna be for us to kinda know where they reside in that ocean of customers. So again, we’re always gonna continue to go broadly and as we do so we’re gonna look at those so-so customers and we’re gonna say, “What makes them different than the really good ones?” So we can basically avoid over acquiring them. So we do wanna know about them, kinda what makes them tick, and let me give the positive spin on it to the extent that the so-so customers are aligned with the good customers.
0:09:34 DF: Let’s go back to the product road map now. I don’t like the idea of a product road map, I want a customer road map, the products that we develop will be a natural offshoot of the decisions we make about which kinds of customers we wanna go after. Were gonna go after certain kinds of customers, that is going to drive the kinds of product decisions we make. But we hope that while we come up with products that are gonna be really, really appealing to the good customers, and attract more of them, we’re hoping that the so-so customers like them as well. That would be icing on the cake. We really do care about the so-so customers, but just not quite in the same way. And again when it comes to product road maps, when you say words like that, it shows that you’re not thinking customer centric-ly, or…
0:10:20 MH: Heaven forbid if somebody had a title called product analyst, that would be a nightmare.
0:10:24 MK: I do confess, I work in the product team.
0:10:28 MH: That’s it Moe, you’re off the show.
0:10:31 MK: Gone.
0:10:32 MH: We tried, but not customer centric. So there’s like a fundamental kind of confusion. If you’re casting a net and you bring in all these customers, how do you know from purchase number one, which ones are good and which ones are so-so? Like how do you…
0:10:50 DF: Well for a startup we don’t, so we have to wait some amount of time to really see their goodness manifest. Part of my job is to figure out how little data, and how little time do we need to wait in order to do that. But the other part of the job is what can we leverage from our previously acquired customers that’s gonna give us a pretty good sense about this just acquired customer, to know whether he or she is a good one or not? So for instance, every time we acquire a customer let’s look at every possible characteristic, every possible descriptor that we get of that initial purchase. When did we acquire that person, through what channel, what was the first product that they bought from us? What campaign did they respond to? Even bring in demographics, I’m fine with that.
0:11:38 DF: And on the basis of all those descriptors, see how close they resemble other customers that we have, in some sense come up with a weighted average of all of our existing customers to say, here is our initial best guess about this just acquired customer. If this customer seems to have all the right characteristics, then that’s great, we’ll put them right in, the red carpet, gold medal, blue ribbon club, even though we don’t have a lot of information about them. So I think that’s a really interesting challenge.
0:12:09 TW: It’s like a look alike model, but you have to start by saying take my historical data set, figure out who my best ones are, there may be three or four, 27 different combinations of characteristics but it is truly about a modeling exercise so that then when somebody comes in you can start being probabilistic as to how valuable they will be.
0:12:32 DF: Absolutely, and I’ll be the first to admit that when it comes to all of the fun things that we can do with customer lifetime value, number one on most people’s list would be to use that as the basis for lookalike modeling. And with a lot of the companies that I work with showing how we can move the needle by doing lookalike modeling on the basis of CLV instead of other lesser characteristics, that’s what really opens their eyes and, “Okay, okay, okay, now tell me the 49 other fun things to do.” So that, again, is a great point to be raising early and just helping companies recognize that and to do it the right way is really important because then all the other benefits follow from there.
0:13:13 MH: So what would you say to someone who’s like, “Hey, Apple didn’t follow a customer centric model, [laughter] right? Steve Job is like the product to product. [laughter] No, but but I’m asking a loaded question because I think a lot of people out there would be like, “Whoa.” Henry Ford said, “If they asked customers what they wanted they’d want a faster horse.” So then in a customer centric mindset like there is an answer to this and I think that’s what I want to give people a bridge so they don’t have to feel like there’s a choose a side kind of a sense to that.
0:13:50 DF: Well, you know what? I’ll help you build that bridge but first I’ll pick up on the very last point, Michael, which is to say, I think you do have to choose. I think a company has to decide what its fundamental strategy is going to be. Are you gonna be a product leader, we are the best, best, best of what we sell, and Apple is the exemplar of that. Are we gonna focus on operational efficiency that, man, we just turn the crank really well and keep the cost down. Amazon, Wal-Mart, Zara, I mean companies that are just really good at getting stuff out there. Or are we gonna be customer centric. I don’t think that customer centricity is the right strategy for every company. I think it’s the right strategy for maybe, maybe a third of the companies. And if you are finding great success as an Apple, tremendous, stay with it. But here’s the thing, eventually it does commoditize, eventually it does plateau, that’s what’s starting to happen to Apple today. And the problem is too many other companies emulate them when they shouldn’t be. I’m gonna tell you a specific story. I’m gonna name names over here. So… [laughter]
0:14:51 DF: At some point in this conversation we’re surely going to talk about net promoter score, and we can get to that for listeners who aren’t familiar with it, we’ll fill in the blanks on it. But I was giving a key note talk for Bain Consulting, the originators of NPS. I was giving a key note at their NPS loyalty forum a couple years ago and I’m giving all this stuff about [0:15:14] ____ picks up on this, the early part of the conversation, we have to understand what makes the promoters different from the detractors, and instead of trying to turn the detractors into promoters, let’s just find more promoters. Let’s just get smart about it. And people are getting on their heads and they’re with me and then the woman sitting over in the front row speaks up, and it’s Angela Ahrendts, the person who runs Apple retail. And she says at Apple that’s not how we do it. At Apple we want to delight everybody, or something like this, we’re going to find what ails those detractors and we’re going to fix it.
0:15:50 DF: And so all of a sudden she’s completely turned the room. Everybody is now nodding their heads at what she said. But I’m not gonna give up on this. [laughter] So I say to her, I say, “With all due respect,” and I really do mean that. I mean ’cause I have tremendous amount of respect for Apple but every other company in the room, you’re not Apple, and you never will be. And so trying to follow the Apple strategy and emulating the Apple tactics is not going to work for you. So for this, this is gonna be great for Apple and that’s terrific ’cause you will have customers lining up around the block every time you launch a product and they don’t even know anything about it.
0:16:26 DF: But the rest of you don’t have that, and you’re never gonna have that, or until you have that, you gotta to listen to me instead. And it’s going to be focusing on those better customers, recognizing that those so-so ones are born that way and are probably gonna stay that way and you really gotta focus on those differences instead. So for Apple, great credit to them, but it is interesting that I’m starting to get a few more emails and things every now and again, and it’s clear that they’re reading some of my papers and are asking me some technical questions, and are asking me about some of the dance steps of customer centricity. So that doesn’t mean they are going to abandon product leadership but it means they do need to shore up some of their weaknesses on the customer centric dimension.
0:17:11 MK: Just on the net promoter score, I’m curious to hear your views about why customer lifetime value. If we’re talking about being customer centric why isn’t a customer happiness metric appropriate? Because doesn’t a happy customer mean someone who is more likely to be brand loyal, someone who is more likely to come back, I’m just, yeah, curious to hear…
0:17:32 DF: Well, I have, I don’t even want to call it a love hate relationship, but I have probably more love for NPS than any other academic. Most academics hate, hate, hate NPS because it is such a simple measure. And a lot of my colleagues in the field of marketing have been developing these elaborate very carefully validated customer satisfaction schemes over the years with 100 questions. And then this one question comes along, “Will you recommend us?” And it kinda blows them all out of the waters. Everyone just loves to hate NPS. I actually love it for a number reasons.
0:18:06 DF: Number one, no particular order but what the heck. Number one is the fact that it celebrates heterogeneity. The fact that instead of looking at an average satisfaction score we’re looking at the differences. We recognize there are some who love us and some who don’t. And it’s the difference across them that really indicates how well we’re doing, so I love that. Number two, I love the fact that the simplicity makes it very comparable so we can really see how we’re doing for different kinds of segments of customers, different geographies or different product lines. And number three, maybe the most important reason is that it is the first time that a customer metric has really gained the attention of CEOs and other C level people. So for me, NPS is kind of a foot in the door to get them to actually care about customers. To get them to appreciate differences across customers and to get them to ask other kinds of questions about it.
0:18:57 DF: So I’m fine with Net Promoter Score and really to your point Moe, having an attitudinal metric as a compliment to a financial behavioral CLV metric, I think it’s a great thing. Again, I developed the world’s best CLV models but I’m the first to admit that they’re still incomplete and there are still aspects of what’s going on inside the customer’s head that aren’t necessarily reflected in those metrics. So the ultimate thing is to have both. To have the behavioral metric that’s just looking at transactions but having some kind of attitudenal metric whether NPS or something else that might either align with it or pick up some other aspects of it. So I really spent a lot of time talking to companies about trying to, “Now that you got NPS, now let’s just slide CLV in there as well”, not necessarily as replacement for NPS.
0:19:48 MK: I completely agree with you. Our number one company KPI is our MPS and what I do find difficult about that is that it’s really an output metric. It’s so far down the funnel. We send our MPS survey quite late after a customer’s purchased. Having something a little bit more tangible and also I agree in having some kind of monetary value is important. So I think that’s a really good suggestion to have the two compliment each other.
0:20:18 DF: And I’m so happy to say, but first of all I’m glad that you agree. I’m so happy to say that a lot of companies on their own have been coming to me saying, “We’ve got the NPS thing, we’ve been drinking that Kool aid. We understand how good it can be. But we know there’s more. Can you help us bring CLV in as well?” I think that NPS has done a really great job of just elevating companies’ awareness of what kinds of metrics they need and it’s created demand for CLV that just wouldn’t exist if the folks at Bain, Fred Reichheld in particular, hadn’t done all that great work 15, 20 years ago.
0:20:54 TW: But is there that fundamental that the way to move the NPS score is to increase promoters and decrease detractors and you generally think really weight more towards the get more promoters and do more with your promoters is that a fair…
0:21:13 DF: Yeah. It is a fair statement and it’s the biggest problems that I find. I’m glad to give you the whole list of issues why people throw things at me and disagree. But this could be number one on the list, the idea that I’m basically being so fatalistic about and saying “They’re bad customers. Live with them.” A lot of companies take that personally. It means it’s a failing of them as people or as corporations that we can turn these ugly ducklings into beautiful swans.
0:21:43 DF: Now, let me first clarify my position by saying I recognize that customer development, the improving of our existing customers is absolutely part of the holy trinity of customer centricity, customer acquisition, customer retention and customer development. We are going to try. All the time we’re going to be asking questions. “Do you want fries with that? Do you wanna super size it?” To cross-sell and up-sell. But in most cases those development activities are gonna be kind of transient, they’re gonna be opportunistic. They’re not going to be transformational. They’re not going to be turning the ugly ducklings into beautiful swans.
0:22:23 DF: And you know what? Even at times when they do appear to be transformational, my view is and the data seems to support this, that those customers were actually born good. They were good all the time. But you just didn’t put the right kinds of products in front of them. You didn’t respond to their needs. You were undervaluing them because of managing the relationship poorly and you finally opened up and unlock the value that existed with them all the time. So too often companies are patting themselves on the back for those transformations when they occur.
0:22:55 DF: Look, I will acknowledge that there will be some customer dynamics over time. The classic story in financial services where someone starts with the student loan and they get the car loan and they get the mortgage, and then the home equity line and the boat loan. I mean, there will be some growth in customers over time but it’s not as easy to orchestrate that or to make it happen over a short period of time than what a lot of companies believe.
0:23:23 TW: So, another nagging question I have is how… Of the companies, the organizations you’re working with that are embracing or even have the potential to shift, where does CPG, where does package goods or FMCG for the European…
0:23:41 DF: CPG, FMCG they make me so sad.
0:23:45 DF: Because there are a few sectors out there that have the real desire, the real willingness to do this and they’re always trying initiatives here and there. Whether it’s “Let’s get some kind of social media thing going on so we can start developing direct relations with our customers, hey let’s get a loyalty program going, hey let’s put in store tracking this or that or hey let’s go out there and spend a billion dollars and buy a dollar [0:24:10] ____ that’s probably not even worth half of that. Let’s just do things so we can start to get in the game as well.” And it’s really hard for them because putting aside their current customers, retailers, to the extent that they can be customer centric with their end users which is what they aspire to do is really, really, really difficult.
0:24:32 DF: So they have all these smart people and they’re really trying all these hard things but then I turn around and look at the idiots that we call retailers, who have good data, who have the capability to offer different kinds of products and services, who have the ability to just do all of this stuff and they’re just too lazy, risk averse, cheap, stupid to actually do it. So I really feel for the CPGs who want to but aren’t yet in a position, and it just makes me even more angry about the retailers who just have this golden gift that the CPGs would die for and they don’t take full advantage of it.
0:25:11 MK: So at this point I might confess that I work for a retailer.
0:25:15 DF: Not all retailers are lazy, dumb, and stupid, and I’m not saying that about their employees.
0:25:18 MH: It’s just US retailers Moe.
0:25:20 TW: Or the product team as the product analyst.
0:25:23 MK: Just giving you a hard time. But I actually…
0:25:27 MK: Funnily enough I wanted to talk a little bit about online retail, particularly in ecommerce. Does this whole view of customer centricity and this approach, does it really start with single customer view because I know that we’re totally online. If someone’s coming into a store it’s a little bit different, it’s a person, they have a loyalty card or whatever it is. In the online space particularly people will have multiple accounts, multiple email addresses. Is part of understanding their value to the business starting with that outset of how do we actually get that single view of a customer across devices, across accounts, what are your thoughts on that?
0:26:10 DF: So US specifically is at the starting point so let me get to that in a second. But is that single view of the customer essential? Yes, it is. Is it the starting point? Not necessarily. So it could be… The starting point is just getting the mindset around all this stuff that this is what we wanna do, here is how we’re gonna design the organization to do so. Yes at some point we’re gonna have to have that single view of the customer to really act on it. But you know what, if a company wants to start with some kind of proxy measure for customer lifetime value, even if it’s net promoter score that’s fine with me. Let’s just get some place holder in there just so we can get going with it. And let’s think of, of course, the king of online retailing, Amazon.
0:26:55 DF: When Jeff Bezos got that company off the ground, his objective was to find rich people and he admits this, there was a really nice interview with him recently. He said he wanted to find rich people cause he had the sense that that would be a good proxy measure for what we call CLV. He’s not that right about it but he had nothing. So let’s start with a proxy measure and then let’s think about products that will disproportionately appeal to the rich people, books, and that’s really what led to the whole Amazon story.
0:27:24 DF: So I’m fine if a company doesn’t necessarily start with that single view but if that really does have to be on the roadmap pretty early on. And I really worry because a lot of companies grossly under invest in that single view, in CRM systems. They’re trying to look at let’s say CRM as an end unto itself. Will this system pay for itself? And the answer is well no. But will clean bathrooms pay for themselves? No. It’s just something that a company just has to do. And the value of it will show up once we start to see and leverage the CLVs. Yeah we have to get there, yeah we have to get there pretty early. Yeah we have to do it in a smarter way, it’s not just some kind of function that’s way down in the organization. The CMO himself or herself should really, really care about that instead of worrying so much about just the brand. I think it is mission critical in a way that most companies, not just retailers don’t fully appreciate it.
0:28:28 MK: You touched a little bit before on demographics and…
0:28:33 DF: You should be keen to avoid it, avoid it.
0:28:35 MK: The reason I say that is because people continually ask analysts for, what is our average customer? What is the persona of our typical customer? And how does that really conflict and what are the dangers here if you’re trying to be customer centric but lots of people are coming with that mindset of, oh we’re trying to be customer centric so the first thing we’re gonna do is look at demographics. We’re gonna look at our typical customer. Do you have any advice on that?
0:29:04 DF: Yeah, I do. In fact you see the hairline here, you see how it’s all working it’s way back, that’s ’cause of demographics. I’m just pulling it out from all these companies that A, don’t get it in the first place but B, even after I educate them about it, it really is about what they do, not what they look like. They say, “Okay that’s great that’s good, now tell me about the demographic.” No, let’s stop. In the 1950s the only thing we could measure about our customers is what they looked like that made all the sense in the world, that’s the only way that we could put people in buckets. And we would start with those demographically oriented buckets and we’d say which of those buckets is best? Which of those buckets is the best fit for us, we’re gonna produce these products, we’re very product centric. But which one of those demographic buckets just seems to be the best fit with the products we are going to produce anyway.
0:29:56 DF: That’s the way companies operated. That’s the way companies still operate, it’s just a shame, that’s just this legacy mindset that persists today. Here’s the role of demographics, put them aside, at least for starters, let’s start with meaningful measures or indicators of customer value. First and foremost it’s what they do. It’s recency, frequency, monetary value, it’s behavior, it’s transactions. I have no problem bringing into that mix attitudinal variables like net promoter score or just other things about wants and needs and frustrations that might not be picked up in the behavior. I have no problem bringing in social type variables like where are you located in the social graph, and if you’re surrounded by high value customers, then we should maybe treat you well too, even if you haven’t manifested that value.
0:30:49 DF: After we do all that, after we come up with our CLVs, then we can bring in the demographics, then we can we can bring in the personas, then we can bring in all the other old school variables to say, now that we see the CLVs, are there demographic differences between the high and low value ones, the answer is usually not but if there are then I’m totally fine to cease upon them.
0:31:13 MK: At least it’s not Tim for once.
0:31:14 DF: And last piece of this rant, before I pull some more hair out.
0:31:21 DF: I recognize that even after we wave our magic wand and see the CLVs across all of our customers, we need to go to the salespeople, and just to even to the folks coming up with the marketing messages. And we can’t just say, “Go after those people and these are the ones to leave alone.” We need to come up with some kind of tangible description. So I’m actually, in a weird way, okay with personas, as long as they’re driven by customer lifetime value. So what is the stereotype that makes these customers different than these customers over here? And even if it’s not perfect, even if it doesn’t perfectly describe those customers uniquely and exclusively, if it’s just gonna be an approximation that steers some of our resources in the right direction, I’m okay doing that as a starting point. ‘Cause again we’re chucking nets not spears, and so as long as we can just get a little bit better about the direction of them, demographics can help us do that but it shouldn’t be the starting point.
0:32:21 TW: It seems like the persona exercise, if you have that customer lifetime value data to develop your persona, that’s forcing you to think through who are the customers that are most valuable to you, which ultimately the whole value of describing personas so that the creative team has something tangible in their mind. But that does seem like a shift, you’re not describing this 15 to 27-year-old female living in rural areas, you’re describing something more from a lifetime value perspective.
0:32:53 DF: That’s right. Exactly right. So again I’m fine layering the personas on to bring the CLVs to life but that’s very different than the way a lot of companies pursue it, which is they just make stuff up, that we want: Busy Betty, working Wanda, professional Paul or whatever it is.
0:33:08 DF: They just make these things up. Based on a purely stereotypical guess about their customer base. And the problem is by my legitimizing personas at all, even when I say it’s this limited application for them, once I say, “It’s okay to use personas at all,” people take that as the wrong idea and then they run the wrong direction with it. People are so consumed by demographics, and don’t even get me started on the whole snake person thing. I mean, uhh. I would have really thought that, let’s take it back to e-commerce, that thanks to e-commerce, we really do have the ability to tag and track individuals, and really understand how they’re different from each other in meaningful ways, that a lot of the demographic and persona stuff would have shriveled up and started to go away by now.
0:33:55 DF: It hasn’t, it’s going as strong as ever before. And I think it’s really a problem, that e-commerce companies should be leading the way and changing practices, and some of them are. But for the most part, when the company grows big enough and it’s time to hire the CMO, they’re bringing that person in from PNG or Unilever, and they’re bringing the same bad dumb practices with them. And I don’t mean disrespect for all those companies, ’cause for those companies, that’s the only thing they can do. But to bring CPG practices into e-commerce is generally not letting you play your strongest suit.
0:34:32 TW: It can be like Walmart and just keep bringing e-commerce into your company.
0:34:37 DF: Walmart is really, really, really interesting in that way because in my book, I singled out Walmart. Again, not being negative about them but saying they’re pursuing an operationally efficient strategy, so they don’t have to worry about the customer centric thing. But to their credit, we might question the means that they’re doing it or the amount of money that they’re paying for it, but the fact that they are trying to learn a little bit more about it, and move in that direction. It’s not only with the acquisitions but it’s even some of the technologies that they’re using that I think get much less attention. So for instance, Walmart does not have a loyalty program, and if you ask them why, they say, “‘Cause that would be overhead. And that would raise our cost and we don’t wanna do that kind of thing. What we wanna do is to try to get the benefits of a loyalty program without actually having one.”
0:35:25 DF: So for instance, let’s come up with this, I forget what they call it, “Scan and go” or something like that. A mobile app where you actually scan the products in the store and then when you check out, instead of having everything. Pull it out of the cart, put it on the conveyor belt, you just hold up the app and it says, “Here’s all the stuff that you bought,” for them, it’s labor savings. People getting out of the store faster, they can hire fewer people, it helps them be more operationally efficient. But at the same time, it gives them the data that they would have from a loyalty program, so Walmart to their credit, they are taking steps in this direction and part of it does come to the acquisitions as well.
0:36:00 DF: So, they’re acknowledging the need to beef up that side of the business. They’re not there yet but I’m interested to see how it’s gonna play out for them.
0:36:10 TW: I wonder if the demographics bit is also being somewhat…
0:36:14 DF: [0:36:14] ____ still talk about demographics.
0:36:16 MH: Well, [chuckle] I was gonna say the other…
0:36:18 S?: We’re over it.
0:36:19 TW: Well, from the media side if, like, Facebook when they’re selling advertising, they have the highest volume of gender and age, they know that. And I feel like there’s that piece too when it comes to… When you’re setting up a campaign, they’re saying, “Oh, this is the easy thing for us… For you to pick or it’s the thing we feel most comfortable inferring.” So, there’s a little bit of kind of a unhealthy feedback loop of where they’re saying, “We can target these demographic characteristics much better than we can target… Certainly direct customer lifetime value but also other characteristics.” So, it’s a shortcut.
0:37:00 DF: Sure. So, there’s no question, Facebook is facilitating bad practices. They’re not promoting them but that’s what their advertisers want. They just gotta be there for them. But at the same time, let’s give Facebook some credit too. Some of the things that they’ve been doing. Some of the algorithms that they developed are really, really good. So, first on the technology side, I’m amazed at how few people are aware of the Atlas technology that they have which really has Google and all the other big tech firms worried. ‘Cause Facebook, it’s an amazing piece of technology to really link together behavior across different platforms.
0:37:37 DF: And then on the algorithm side, what they’re doing with their Facebook look-alikes is really great. It’s actually a lot better than the equivalent service that Google and the others offer. So, I don’t know what’s going on inside that black box but when I’m working with companies and saying, “Hey, here’s the CLV people, let’s go to Facebook and come up with other customers who resemble them,” they’re doing a pretty good job of that. And so part of that might be demographics, sure, but I think a lot of it is based on other characteristics and then they’re kinda finding a genuinely good look-alike customer. So, Facebook, they get it but at the same time they can’t completely move away from demographics, they’d be out of business tomorrow.
0:38:22 TW: Well, but it’s just… I’ll give you… I think that’s one where, like, they’ve got the offer… The product where you can upload a bunch of email addresses and they’ll do the matches and it’ll all be hashed and protected and they’ll sorta do a look alike. But it seems like companies are like, “Oh, that’s hard and we don’t trust it, we’re like uploading our customer data to Facebook.” So, that’s where it’s like, they sorta enabled it. I also think they’ve got some brilliant data scientists predictive analytics modelers over at Atlas developing people. And then they’ve got the people who are doing what annalists get to where they can’t actually count a video view correctly and they keep having egg on their face. So, it’s clearly like two different buildings where one is like slow stuck in 2002, we can’t even count a video view and on the other end they’re doing this amazing stuff but there’s my rant.
0:39:13 MK: Got one in, got one in.
0:39:14 MH: So we’ve got time for one more question about demographics… No, I’m just kidding.
0:39:21 MK: I actually do have a kind of not demographicy but related question…
0:39:27 S?: No, you cannot.
0:39:29 MK: So, Pete, if you’re advocating for looking at what customers do instead of who they are and that’s the D word. I just wonder if there’s a slippery slope, and I again have my product hat on here, where you go, “Okay, good customers with high CLV, they do these things on our product,” right? “So, let’s try and make all of our customers do these things.” And I’ve seen that play out with really dire consequences. How do you reconcile that with instead saying, “Yes, these are the things that our high value customers do but we acknowledge that not everyone wants to do them so let’s just build great features that support people to do the things when they wanna do them.
0:40:15 DF: When it comes time fo product development, let’s not aim for that average. Let’s not say, “What products can we develop that will be widely loved.” I’m gonna give a big shout out to one of my favorite companies that does this oh, so well, Electronic Arts, the gaming company. I’m overstating things a little bit here but philosophically, when they launch a new game for them it’s not, “How many units of the game did we ship?” Because, you know what? Suppose all of the people who buy that game are low value customers who never bought from us before and will never buy from us again. The lasting value of that game is nothing. On the other hand, if we can develop a game… And maybe fewer people will buy it but those who buy it will either become high value customers or high value customers will become even more valuable, then the value of that game could be a thousand X over the number of sales of it.
0:41:08 DF: So, we shouldn’t really be worrying about just what products or what features will sell, we should be saying, “What are the things that we can do to enhance value or attract more valuable customers?” And so, again, that takes us to the product road map is really the customer road map. So, it’s not so much how can we get those so-so customers to love this product too. It would be great if we could. If we could do that in some really cost effective way to get them to wake up and enjoy this great product, that would be fine. But in most cases those same resources will be better allocated to appeal to… To either come up with features or new products that will be more appealing to the high valued customers. And if the low value ones wanna buy it, like I said before, icing on the cake, we’re happy to have them buy it too.
0:41:55 TW: So, there’s… I don’t know that they’re emerging, but there are newer industries around service design and experience management. They’re coming at this customer-centricity maybe from a experience or user kind of a perspective. In your view, how do those intersect with this customer centricity model?
0:42:19 DF: Michael, for the most part I disagree with the premise. I think that most companies that are going heavy duty into the CX thing are doing it in a product centric way. So what experiences can we wrap around the product that we’ll get people to buy it instead of going to Amazon. I don’t think they’re doing it in a customer-centric way. If they’re doing it in a customer-centric way they’d be saying, “Okay we’ve got these high value customers over here. What experiences would be particularly appealing to them?” It might appealing to the others as well. If they’re celebrating heterogeneity as they design and evaluate those experiences, then I’m good with it. But if they’re doing it in just some kind of one size fits all way, what will make the customer happy? Which is really what’s going on, then we’re just in the same place that we were.
0:43:06 DF: And the corollary to that is how they do the measurement. And anytime you’re doing any kind of CX campaign, it should be just like the product example that you just mentioned. What was the CLV of the customers who participate in that campaign before and after? And that’s how we’re gonna find if it’s really bringing value or not. And instead [0:43:25] ____ just saying did it help us sell more product or not. I don’t see CX as being aligned with customer-centricity unless companies are doing it that way, which most of them aren’t. If anything they’re using it just as a way to almost stick their heads in the sand and it’s the next shiny object that they can chase after so they don’t have to deal with the data and the analytics and the CLVs.
0:43:53 MH: Right. No, that’s great. I love that perspective. And you keep using this word CLV, what does that mean? I’m just kidding.
0:44:03 DF: You know what? I gotta say this, everyone understands what the letters mean, but people don’t appreciate, A, how accurate we can make these forecasts. That’s frustration number two, maybe behind the demographics, is people don’t trust it. People would rather run their business on backward-looking metrics, ’cause they can measure them precisely rather than use forward-looking projections ’cause they’re just not sure about it. This is my academic thing, I wanna teach people not just to trust it, but how to validate it. How to know whether these models are accurate or not. Here are the pictures you wanna look at, the tests you wanna run so the models work really, really, really well. And let’s say if we’re talking about retail in a non-contractual setting.
0:44:52 DF: So whether we’re talking about retailers, or airlines, or hotel chains, or pharmaceuticals, or industrial distribution, or mobile games, or console games, where there’s no contract. It’s just people doing things over time and then whatever reasons they just stop doing it. It’s the same model. So you don’t necessarily need to develop a custom model for each different sector. I can just take my non-contractual model off the shelf and apply it to the next non-contractual business, B2B, or B2C, Product or service doesn’t matter. I’m trying to convince people that CLVs are more than just a bunch of letters. It’s something real. It’s something tangible and validatable, it’s broadly applicable. To me, I’m obviously biased, CLV is actually much more practical than the hot topic these days in the analytic space: “Attribution modeling”.
0:45:49 MH: You have come to the wrong podcast if you’re somebody who loves attribution.
0:45:52 DF: We have got to talk about that. I’m not gonna be critical about attribution. I think it’s an important test to wanna do, except you can never validate an attribution model. And I’m developing in my research too. We can develop the model itself, but we can never know if the attribution, if the assignment to value, is correct. We never know what truth is. Whereas with CLV, we can validate it. We can say, “Over the next two years, how many purchases will this person make? Did they or not?” So I think we wanna start with the stuff that we really can measure and validate. And once we’ve really exhausted all of the validatable metric, then let’s start bringing in the stuff that’s important, but a little bit more of taking it on faith. So again, I’m not criticizing attribution but I’m saying I wish companies would first start with CLV. In fact, when you’re doing your attribution model do it for CLV, not just do it for did we have a conversion or not. So anyway, you get the idea.
0:46:53 TW: Now, I’ve got to ask two quick, quick CLV calculation questions.
0:46:56 MH: Look at this guy.
0:46:57 MK: Seriously.
0:46:58 TW: You say you can use it… So is it a fixed time frame or is that vary by business? And two, are you typically discounting? Is there a net present value where you’re discounting? You definitely do that and then what about the timeframe and the pickup.
0:47:14 DF: With the discount rate, absolutely. And if you read any of my papers, you are going to hate me because I always, in my papers use 10%. Why? Because that’s how many fingers I have. I mean, it’s just it’s an arbitrary number. I think it’s really really important for companies to think very carefully about what that discount rate is. At the very least it should be the weighted average cost of capital. But you might even say, depending on which geographic area we’re looking at, or which customer thing we’re looking at. Maybe we should be using a differential discount rate based on the inherent or the expected riskiness of the customer base. These are conversations I was never having until about two years ago. Until I got my own CLV startup off the ground, and now we’re really trying to educate companies about using the right discount rate. So that was question two. What was the first one on that?
0:48:02 TW: Well, the first one was gonna be the timeframe, which the discount rate will, seems like it will effect when it becomes meaningless because your discount rate’s gonna give you a natural horizon.
0:48:10 DF: Sure. Absolutely, but to the timeframe goes back to an earlier question that you asked, again, I want it to be as short as possible, but if I have a really good CLV model and if I can leverage the data across other the customers I’ve acquired in the past, I’m not gonna need a long timeframe. So, but to your main point, it does depend that for those new customers that I don’t know much about them and I am gonna wanna go through several purchase cycles to see if they’re staying with me or not. Yeah know, if we’re talking about mattresses or refrigerators, it’s gonna be a little bit longer than if we are talking about a mobile game, where it takes me maybe a week to know what their CLV is going to be. So, the timeframe does matter but having said that if we stretch or shrink that timeframe, again recognize that for refrigerators it’s long, for mobile games it’s short, the basic patterns of customers doing things over time, and how it differs across customers, and how they tend to slow down on an individual level, the overall patterns are still remarkably similar.
0:49:12 TW: Cool. Okay, now I’m done.
0:49:15 MH: That’s awesome. All right. So now we do have to, sadly, we do have to start to wrap up. This has been a great conversation and I think will probably spearhead a number of conversations in the analytics community for sure. One of the things we do on the show is we go around the horn and we share what we call a last call, which is, anything that’s interesting that we’ve observed recently, can be analytics related, not analytics related, or if you have a new book coming out, that would be great to talk about too. I don’t know if you do but…
0:49:46 TW: But you should. [laughter]
0:49:48 MH: Pete, we’ll start with you if you have one.
0:49:51 DF: Yeah, the cool new thing, I mentioned very briefly before is they got this customer centricity simulation coming out. It exists right now and we are happy to share it. It is a great way, first of all, to get people to understand the difference between products and customer centricity. There’s just nothing like jumping behind the driver’s seat and say, “Our job is to build this profitable customer base as possible.” So, what kind of customers do we want, how much are we willing to spend on them, how much are we gonna allocate to acquisition versus retention and development, are we gonna develop a loyalty program, a CRM system, all of this stuff, to really bring these principles to life, but in a really fun, experiential, and competitive kind of way. In early 2018, we are gonna have this thing out there broadly for not only for companies to use as an in house training exercise but even for individuals to kinda buy it and play the game and just do it for fun. So customer centricity simulation, you heard it here first, and all the cool kids will be doing it soon enough.
0:50:52 TW: Well, is that when the companies are using it there will be some tweaking to make it more real to fit with them or is it a generic scenario?
0:51:03 DF: At first, it’s generic. At first, it’s some 3D printing company, it doesn’t even matter. The parameters that are in there are pretty realistic, but I think once you get people to appreciate it then, of course for me it’s a trojan horse to lead to some of the academic work, then it’s okay company, now give me your data and we will build some of the same kinds of models for you, to understand how long your customers are staying around and how do they differ from each other and how much are they spending and how responsive are they to different marketing activities. So, it’s not so much that the sim itself can be tweaked, although I suppose it can be, but it’s more to then get companies to just want to, not play the game but to actually do this for real on their own with their own data.
0:51:49 TW: Got it.
0:51:49 MH: Nice.
0:51:51 TW: All right. Going down the line, Moe, do you have a last call?
0:51:56 MK: I do indeed. I am stealing Tim’s thunder because mine this week is actually an R package which I just discovered and I am so excited to try out. So, it’s the Gmail R package, now the reason I think it’s really cool and I want to mention it is because lots of analysts and this is the use case of how we discovered it, lots of analysts get CSVs that are sent to them automatically to their email, they then download them, then they do a bunch of stuff with them. With the Gmail R package, you can now get it to read your email, download the CSV, do all the stuff that you wanna do and then resend it out via email. So, you don’t even actually have to open Gmail or Outlook or whatever it is, oh Gmail, which is just, yeah, it blew my mind.
0:52:39 TW: If that CSV’s just a bunch of demographic data, will it just delete the file automatically?
0:52:47 MK: Auto-delete? Yeah, so that was my really cool find and I saw it being used really well the other day in that context.
0:52:55 MH: All right, I am gonna do one but it’s silly. Well, no, it’s not silly. Actually my last call is a podcast, specifically episode 70 of the digital analytics power hour. So, here is why though, sort of an analytics power hour first to recommend people to go back and relisten to this episode. So, I’ve been taking notes this whole time and there’s been probably eight or nine really interesting things to go explore from here that I think our listeners should go back in a time when they can actually jot some of these things down so they don’t lose them. There is some really critical marketing ideas and analytical ideas that will advance your life as an analyst pretty significantly. And a lot of it had to do with us as the hosts of the show and then also our guests. [laughter] So equal measure, a little bit of balance. That’s my attribution. I can’t validate that.
0:54:05 MH: I’ve literally written down about eight or nine things that I’ve just been like, “Oh yeah that’s… ” and some of it’s really validating too. I’ve been telling our clients for a long time, “Use the behavioral data to inform persona builders”. Build them off of behaviours and then bring in this other stuff. Cause the people build them just to make, yeah, like soccer moms and, whatever. Anyways sorry. Tim, what’s your last call?
0:54:35 TW: So first I wanna call out that Moe actually did more talking than I did as a host this time.
0:54:42 MK: I can’t believe it, I can’t believe it.
0:54:44 TW: If ever that, it’s been, it’s thought that it could not have been done in 70 episodes. It has happened, that I was not the one constantly cutting people off. My…
0:54:55 MH: We can fix that in post editing, so…
0:54:58 TW: So my last call is a… I’m gonna be at Columbus, Ohio homer, and call out the Women In Digital annual conference, October 25th through 27th. So, a past Web Analytics Wednesday Columbus Attendee Elena Cher started this like a year and a half or so ago. And I was oblivious to it. But, it’s two things to call out: One, it’s a Women In Digital. There’s all this women in analytics, kind of, momentum going on and a lot of good stuff happening from the DAA, which is awesome. The Women In Digital, seems like it’s pretty closely… There’s women in technology, women in tech as well, kind of West Coast, I think, driven. But, one thing that I just thought was really neat, every time there is a women in something Slack channel, forum, mentor program, there is the “Can guys be there or not?” And it’s always a binary, and I just thought it was super clever that the Women In Digital, the annual conference is the only thing where guys are allowed, but it is limited to 4% of their registrants.
0:56:02 TW: Which I thought was like, that’s the way, that’s like a great way to nail it. Like, yep, guys can come, but you are gonna be in the extreme minority. So, check it out. They’ve got some pretty cool speakers lined up and if you’re up for heading to beautiful Columbus in the mid-fall, the weather’s usually pretty good about that time.
0:56:23 MK: And, hey, you won’t have to line up for the bathroom.
0:56:31 TW: Well put.
0:56:32 DF: Guy’s never do.
0:56:33 MH: That’s true. It’s, not as common.
0:56:37 MK: I suppose for women though, with analytics conference, it’s normally the only time we don’t have to line up, which is the one positive, so…
0:56:45 MH: We’re working to change that, Moe. We’re all working to change that. That’s right. Anyway this, I already kinda made a little bit of an homage to this episode but I really have enjoyed this and I think our listeners will probably enjoy it as well. If you’ve been listening and you’ve got thoughts, ideas, questions, feel free to reach out to us. Pete, do you use the Twitters or the social medias or the ways people could contact you?
0:57:15 DF: I am big big into the Twitter, yes. In fact, that’s, all the time now I’m finding articles or whatever, blogs, about people doing particularly smart or usually dumb things about customer centricity and analytics. So, FaderP on the Twitter. Happy to link in with people, email, whatever else. I hope you can tell, I’m pretty passionate about this stuff. I’m on a mission to bring these ideas and the metrics associated with them to practice. I appreciate your help in facilitating that. And I’m real glad to engage with anyone else who wants to explore the possibilities.
0:57:51 MH: Feel free to reach out to us with questions. Follow Peter Fader on Twitter. You can find us on our Facebook page, on our Twitter page, and at our website, analyticshour.io and on the Measure Slack, which we are all part of. And, so, again, Peter, real pleasure to have you on the show.
0:58:11 DF: And a privilege to be with you. Let’s do it again!
0:58:14 MH: This is some really good stuff, so, thank you very much. And, for my co-hosts, Moe and Tim, out there, everybody, keep analyzing.
0:58:29 Announcer: Thanks for listening and don’t forget to join the conversation on Facebook, Twitter, or Measure Slack group. We welcome your comments and questions. Visit us on the web at analyticshour.io, facebook.com/analyticshour or @analyticshour on Twitter.
0:58:49 Announcer: So smart guys want to fit in so they made up a term for analytics. Analytics don’t work.
0:58:57 TW: Moe, you’re part of it. You have to own it.
0:59:00 MK: I had no idea what was going on though. I was just like, uh huh?
0:59:06 MH: We did discuss one potential sound effect. We’re just wanna gauge your potential interest level. So, given your last name, Fader, we thought maybe just like a DJ scratch right after your name.
0:59:19 TW: Yeah, if it feels like I’m dropping a gratuitous f-bomb, it’s purely so that we can keep that explicit tag.
0:59:25 DF: Well, you know what, whatever. Hit me with your best shot.
0:59:29 DF: Wow, you really did your homework Micheal.
0:59:38 MH: Actually, I’m not gonna be able to do that.
0:59:39 MK: Why not?
0:59:42 MH: I just have this complex. I don’t know what it is.
0:59:48 DF: That would be weird. That would be, like a family therapy program.
0:59:50 TW: Who’s got the siren?
0:59:58 MH: I know, I hear a siren. It’s okay. That will get [1:00:01] ____. I was like, “Should I stop? Should I not?”. Sounds like it’s a siren.
1:00:09 DF: Some silence, I’ve ended the conversation right there.
1:00:16 MH: Yeah you’re gonna install the fader R package then delete all the demographic data automatically.
1:00:28 TW: Michael I have a question for you, where’s Pete located?
1:00:32 MH: Philly.
1:00:33 TW: Where is it always sunny?
1:00:35 MH: Philly.
1:00:36 TW: Rock flag and CLV.