Pedro de Arteaga: This is another episode of our MarFinance series. We will discuss Customer Lifetime Value (CLV) and its application in growth strategy.
With the regime shift from growth at any cost to the path to profitability, I’ve seen many marketing teams in the last two to three years get fewer resources even after being successful in driving way more results with less.
I believe this is strongly connected with marketing and finance, usually speaking different languages.
With this purpose, MarFinance aims to equip growth practitioners with frameworks and tools to align a common language with finance, influence the boardroom, and secure resources for long-term sustained growth.
To do this, I have the pleasure of sharing the screen with two of my heroes in the quest of connecting the dots between marketing and finance.
First, we have Peter Fader. He is the Co-Founder and Director of Theta and the Frances and Pei-Yuan Chia Professor of Marketing at the Wharton School of the University of Pennsylvania.
He’s considered one of the world’s foremost experts in the development and application of Customer Lifetime Value.
There are three books out there. I strongly recommend the audience to read them: Customer Centricity, the Customer Centricity Playbook, and the Customer-Based Audit.
He’s the co-creator, together with Dan McCarthy, of the Customer-Based Corporate Valuation (CBCV) methodology, which is one of my favorite methodologies to value both subscription and non-subscription type of companies.
And before Theta, he and Dan co-founded Zodiac, a predictive customer analytics firm, which was acquired by Nike in 2018.
Dan is a Co-Founder and Director of Theta and an associate professor of Marketing at the Robert H Smith School of Business at the University of Maryland.
He has won several Academy Awards, including an MSI, Alden G. Clayton. He has also been featured in many media outlets and academic journals, including Harvard Business Review.
I know both of you have created hype around IPOs in 2019, 2020, and 2021.
He also holds a PhD in statistics from the Wharton School of University and graduated with degrees in economics and system science engineering from Wharton University of Pennsylvania.
Lastly, I will moderate this panel. My name is Pedro, and I am the Chief Business Officer at Winclap.
For those unfamiliar with our brand, we help brands like Wellhub, MercadoLibre, Didi, Rappi, Avenue, and Globo to transform their growth strategy. In other words, we help them to transform their acquisition, retention, and monetization strategies from multiple solutions, including growth consultancy, paid and owned media management, creator-based content for performance, and Martech services.
We combined a consultancy firm with an agency. We call this category growth transformation.
So, with this, we will kick it off.
Peter Fader: For me, I like numbers. I like data. I like to predict things. So whether it’s sports statistics or the movement of songs through the popularity charts, I’ve always been interested in tracking things and seeing where they will go next and how far they will go for how long. There are a lot of different domains where you can do that. You can do it with Wall Street, you can do it with an insurance company, so many other areas.
While I was an undergraduate math major at the Massachusetts Institute of Technology (MIT), one professor told me, “You ought to get a PhD in marketing.” I thought she was crazy because this was the early 1980s. But she painted a picture of what marketing would become, and she was exactly right. The ability to tag and track customers, and to start asking questions like: how long will they stay with you, how often will they buy, and how much will they spend? It was a perfect fit for my interests. She was right that marketing would become an area with tremendous opportunities, new data sources, new decisions to be made, and innovative ways to evaluate those decisions.
It’s been an amazing ride to help observe many of the changes in marketing and, to some extent, to create them, to push companies forward, to be more data-oriented and customer-centric.
Customer Lifetime Value is just a natural consequence. It combines so many different kinds of predictions into one, and it does that over a long range and with very important consequences. So that’s where I’m coming from. I’ve been messing around with it for a long time.
And there was an absolute pleasure when this young PhD student, Dan McCarthy, walked into my office about ten years ago.
Dan McCarthy: It started when I was a straight-up Wall Street guy coming out of Wharton for undergrad. I was pricing rainbow swaptions and all sorts of crazy financial derivatives and then, more broadly speaking, fundamental analysis.
I spent several years working for a hedge fund before returning for the PhD. In the second year of the PhD, another Dan recommended that I chat with Pete Fader, this guy up on the seventh floor. Yeah, he had to cross dual listing in both departments. But I went up there, and the rest is history.
We started working on one project. That project did not convert into anything publication-worthy, but it interested me in predicting what customers would do.
Then, it was about connecting that with my prior background, where I had spent years thinking about questions like: What will the revenue be? What are the free cash flows going to be? How can we use that to determine whether the company we’re looking at is undervalued or overvalued?
There had been some early work in marketing literature that attempted to link these two concepts. However, much of that research focused on Customer Lifetime Value, rather than addressing it from the perspective of finance professionals. This was the approach we quickly gravitated toward, leading to the publication of our first paper on subscription-based customer-centric corporate valuation, followed by a second one.
So, yeah, that became such a fun topic to study, but how we got there was interesting.
Pedro de Arteaga: I agree. It’s a really fun topic to study and work on.
It’s one of my passions, and it’s also what I do here at Winclap. I invest time in helping clients and our customers understand how their companies are projecting and then being able to influence that trajectory a little bit. It’s both fulfilling and fun.
Let’s continue with a simple but tricky question.
Dan McCarthy: It’s basically the net present value of the variable profitability of a customer, considering all the relevant costs and profits associated with the customer.
So, how much you spend to acquire the customer in the first place, as well as all of the variable profits and costs that occur until the customer ends the relationship. So, one critical point is that we want to be very careful about what expenses we include and which ones we exclude.
The costs that we want to include are everything that goes into the revenue that’s being generated. So, if you’re a widget seller, you want to make sure that you’re accounting for all of the costs associated with the manufacturing of the widget, labor, and materials, as well as everything else you know you have to incur to bring that revenue in.
You would probably not want to include things like the CEO’s salary because when you bring in that next order, you’ll still going to have the CEO.
So, there are a lot of costs that do not scale with that next incremental order.
That brings us to some of the challenges. Questions arise, such as: there’s this cost, and we know it doesn’t increase with the next order, but perhaps that cost could scale if the company were twenty-five percent larger.
It falls in this gray zone, where we spend much time thinking about how should we divide this expense into fixed vs. variable.
That’s one thing. The other big thing it highlights is : how much we spent acquiring that customer?, which is a tricky question to answer.
So yeah, I think that there are easy answers, there are directionally correct answers, and then, we try to get better and better from there.
But conceptually speaking, what it should represent is how much money has been spent to bring in that customer in the first place.
So you would want to take into account things like:
It shouldn’t be a one-size-fits-all number that applies to every acquired customer.
Peter Fader: And if I can add, I love Dan’s definition and how he elaborates on it. Listen to what he said. It’s not all just about revenue. We have to care just as much about the cost.It is about profitability.
If you look at some of my older work, it was overwhelmingly focused on revenue alone when we started laying these models out. Because you know what? Projecting revenue, how long will the customer stay with us, and how often will they buy? That’s all the stuff I love to do. On the other hand, trying to project and understand costs is less exciting and less sexy. It’s less about math and more about accounting. But if we as marketers want to build the bridge to finance and accounting, we have to care as much about that stuff and be as clear, careful, and respectful of those details as we are about the revenue-oriented ones.
That’s where Dan McCarthy elevated the discussion. It’s not just about doing a bit more math; it’s about changing the way we talk about these concepts, the inputs, and the outputs.
And the discount rate. It doesn’t have to be ten percent.
Pedro de Arteaga: Before we jump into this conversation, I know that both of you teach different courses related to the topic.
When you describe CLV, it sounds simple. I take a customer, figure out how much profit this customer will generate over a long period of time, the period of time I can foresee, and consider all the costs incurred to either serve this customer directly or acquire these customers. And there it is. I have CLV, post-acquisition value, and CAC.
Peter Fader: Let me start with a few, and I’m sure Dan will have many more to add.
Of course, he highlighted some of the key issues, such as determining which inputs we should use in a CLV analysis. One of the challenges, however, is that when we talk about CLV analysis, it often implies there’s a fixed formula or a specific approach. Frequently, when you read a white paper or listen to someone discuss it, they’ll present it as ‘the’ way to calculate CLV.
As if there’s one formula or one approach that works for every business, which clearly isn’t true. You mentioned early on the distinction between a subscription model and a more discretionary model where there’s no contract involved.
And there are lots of subtleties beyond that.
I often talk about the six fundamental business models, but even those don’t capture everything. The complexity goes beyond what the company sells—it’s really about the nature of the customer relationship. Do we actually observe when the relationship ends, or is there latent attrition? Do revenue-generating activities occur on a regular basis, or are they more sporadic? Do spending levels remain consistent or fluctuate over time? There’s so much more nuance involved.
And you said it yourself, Pedro, that people think about customer lifetime value as a simple concept. That’s our job, not so much to complicate it as to make it realistic for people to identify some of the challenges in talking about it and, therefore, some of the challenges in implementing it. That’s whether there can be plenty of conceptual and technical complexity. That’s just for starters. I’m sure Dan has more to add.
Dan McCarthy: There are quite a few common mistakes we often see in calculating CLV. One of the most frequent is the revenue-only method of computing CLV, which we touched on earlier when discussing costs. This method often relies on dividing by the churn rate, which has multiple flaws.
First, it tends to group many customers together, and second, it assumes a retention curve that doesn’t hold up in real data. You should never compute CLV simply by dividing revenue by a churn rate.
Pete also mentioned the distinction between calculating CLV at the individual level versus the aggregate or overall customer base level. Ideally, you want to calculate CLV at the individual level, or at least at the cohort level, to capture differences in how customers are acquired and behave over time.
Another issue we frequently encounter is companies failing to apply any discount rate. You might think this is a minor detail, but in reality, it can significantly distort the results. For example, in Peloton’s pre-IPO prospectus and with another company called Aspiration Financial, the lack of a discount rate led to an overstatement of CLV by a factor of two or three. That’s because they used a formula that implied customers would still be active and paying fees thirty, fifty, or even a hundred years into the future.
This is exactly where the discount rate comes into play. Without it, cash flow projected a hundred years from now is valued the same as cash flow next month. But even a small discount rate dramatically reduces the value of future cash flows. This can lead to highly inflated CLV estimates, and it’s often intentional when companies want to present more favorable numbers.
So, it raises the question: how far out should we go when calculating CLV? Ideally, as far as possible, since all future cash flows are valuable. However, projecting too far into the future can be problematic, especially for young companies that have only been around for a couple of years.
Do we really feel confident projecting ten or fifteen years ahead for such companies? It takes experience and careful consideration to arrive at the most accurate answer, and having done this many times certainly helps.
Pedro de Arteaga: In summary, we’re focusing on profits versus revenue. The key is to project based on cohort data rather than relying solely on historical averages, particularly when analyzing churn rates.
We need to ensure that a discount rate is applied. When I read your post about Peloton and realized there was no discount rate, I thought, ‘What the hell? How is that even possible?’ I’m sure there were a lot of discussions around it, but it’s crucial for the audience to understand the concept of a discount rate and make sure it’s always included.
Lastly, we have enough data to foresee and project the future.
That takes me to the next question. The audience is widely heterogeneous regarding the type of businesses, the type of companies, and the stage of the companies.
I have two questions that usually I get asked a lot:
Dan McCarthy: Yeah, good questions.
There was this piece that I had looked over the shoulder of someone; his name is Rodrigo Fernandez. He suggested that even if you don’t have any historical transaction data, it’s beneficial to start thinking about your business unit economics.
When you think about growth planning and building a sustainable business model, you have control over your pricing and a good sense of your margins and what they could be. You likely have benchmarks to guide your understanding of whether your retention is average, below average, or above average, and how well you can improve it. With these insights, you can work backward to determine the economics of the business once you get it moving.
And I think that sort of exercise can be tremendously valuable, even in the absence of transactional data. Now, we have the most fun when we have the transactional data, and we can start running the models, making predictions, and replacing assumptions with just directly observable historical behavior.
But you should do this from day one, even before you have any data. The nature of the analysis you’ll be able to do will vary over time.
In terms of business models, I believe there are two main constraints you often encounter. First, can you actually observe the end customer? If not, that can lead to challenges.
An example of this could be a B2B2C model where, for various reasons, you don’t have direct access to the transaction behavior of the end customer. Take Hershey’s or another CPG company, for instance. Depending on the business model, they might struggle with tracking and tagging the same customer across different channels where Hershey’s products are purchased. This lack of visibility into the customer journey is a significant challenge.
Another challenge is that even with good, trackable customer data, if you only have a small sample, like ten customers, it can still be difficult. You need to know exactly what those ten customers are likely to do, and predicting their behavior accurately can be tricky.
There’s a minimum sample size requirement. We often say that, for our models, we need to find patterns across customers. It doesn’t have to be huge, but at least a cohort of 200 to 500 companies is necessary to identify meaningful trends.
For most B2C businesses, this isn’t usually a problem. However, for some, like very small businesses or B2B SaaS companies with a few large customers, it can be challenging. The same issue applies to CPG companies—if you’re selling something like pickles and Walmart is your main customer, you’re selling through them and can only observe the sales indirectly.
Peter Fader: I agree entirely with the second point, and I also see some of the limitations of companies that will have difficulty getting the data or leveraging the models.
On the first point, when should you be using it? Let’s say you have plenty of data and so on.
I worry about companies trying to leverage CLV too early. They should be doing the calculations, developing the discipline and muscles to continue updating those calculations, and starting to think about what to do with them.
In the early stages of a company, especially for a startup, it’s all about achieving product-market fit. The focus is on getting the product out there. When we talk to investors, we often discuss the Ideal Customer Profile (ICP), but the truth is, most companies don’t have a clear sense of it—they’re essentially guessing.
It’s often crucial to cast a wide net in the beginning, bringing in as many customers as possible to start identifying who your best customers are. Lifetime Value (CLV) plays an important role in helping you figure that out.
But if companies start analyzing too early, they might get a misleading sense, thinking, ‘Oh, our best customers come from this geography or industry,’ when in reality, they could be missing even better opportunities. Early cohorts often aren’t representative of later ones, so relying too heavily on initial data can create false impressions about who their best customers truly are. So, do the calculations and develop the necessary capabilities early on, but don’t place too much emphasis on them until acquisition growth begins to slow down. It’s all about finding the right balance.
Pedro de Arteaga: That’s a great point, Pete. That’s why we typically work with companies that are post-product-market fit. Before reaching product-market fit, it’s difficult to assess customer lifetime value because you’re still figuring out who your real customers are.
Dan McCarthy. The question is when to start doing these calculations. Growth planning is still valuable, especially when you can identify clear, distinct segments, each with its own strengths and weaknesses. I often think about the quality of those segments through the lens of CLV, even though we haven’t done a lot of the calculations ourselves, for some of the reasons we’ve mentioned. That framework can be useful, but running the numbers too early and heavily focusing on segments that look good based on historical data can be risky at that stage.
Pedro de Arteaga: To your point, Dan, modeling growth planning and thinking ahead about what the business needs to look like to be profitable over time is a valuable exercise, even if you haven’t achieved product-market fit yet.
I wouldn’t recommend that companies make decisions based on CLV before achieving product-market fit, but I would suggest thinking ahead about how the business will eventually turn a profit. Most business models out there—though not all—are already well-established. So I imagine that you can look at existing companies, not necessarily the same vertical, but how do they work, and use that as well to complement and think ahead, If you’re not high on the market fit yet, how can you turn a profit?
Dan McCarthy: Yeah, that’s a great question, and it gets to the heart of how to make CLV a useful concept. For many, the discount rate feels theoretical, but think of it this way: imagine you’re a company with access to capital markets, borrowing money to invest in customer acquisition. What rate of return does the capital provider demand? That cost of capital is highly relevant to the CLV calculation because it represents what you need to pay back to the lender to acquire the customer.
As the cost of capital goes down, it lowers the hurdle rate—the required return you need to justify spending that money on customer acquisition. Even if you have a lot of cash on hand, knowing the rate you’d face in the capital markets is a reasonable figure to use when calculating the discount rate.
And, of course, this can shift in response to rate cuts.
Peter Fader: That’s a great example and a solid answer. Dan mentioned early on some of the older methods, where people would simply calculate lifetime value, add it up, and declare that as the company’s value—without considering the cost of capital or access to capital. This oversight ignores a crucial element in assessing the true value of the business.
Some people say, ‘We’re marketers, we don’t worry about that. We’ll focus on the customers and leave finance and accounting to others.’ But the problem with that approach is you lose credibility. Your recommendations and strategy evaluations will be way off. It’s crucial for marketers to understand finance and accounting concepts—not only to make informed decisions but also to have intelligent conversations and avoid making uninformed statements.
There’s already some skepticism surrounding marketing and marketers, so it’s essential for us to stay informed about what’s happening in the capital markets. We need to think carefully about how those changes will impact our marketing decisions and actively participate in that conversation. It’s important not to dismiss it as ‘someone else’s problem.
That’s why I love the whole MarFinance concept. Our goal isn’t just to build a bridge between marketing and finance but to fully integrate them. This means we need to do our homework more thoroughly than ever before.
Dan McCarthy:
There was a person who commented after we released a post about CLV for private equity. On Twitter, they responded—not quite dismissively, but with a hint of skepticism—saying something like, ‘This is just a way for VCs to dump overvalued IPO shares.’ And, honestly, that sentiment is fairly common. A lot of people view CLV as just another marketing gimmick with little credibility, and that’s exactly what we’re trying to change.
I think there’s a strong incentive for companies to follow the path of Peloton—I don’t want to keep picking on Peloton, but it’s a good example—doing whatever it takes to make the numbers look good. They want high CLVs and low CACs, so the temptation is to manipulate the metrics: take costs out of the equation, remove the discount rate, base it only on revenue, and eliminate non-cash expenses—all to make the numbers appear favorable.
But the problem is, when you do that, you end up with numbers that nobody believes. So, what’s the point? It does a disservice to the whole concept of CLV. Having a third-party, objective view to get the numbers right can go a long way in showing people the true value of these metrics when they’re calculated and applied correctly.
Peter Fader: And while I was critiquing marketers for needing to do their homework and pay attention to finance and accounting, it goes both ways. The Peloton example illustrates this well. Keep in mind, it wasn’t just a marketing pitch—this was in documents filed with the Securities and Exchange Commission for the company’s IPO.
There are no clear standards or rules on how metrics like Customer Lifetime Value or customer retention should be reported or calculated. So, Dan and I are going to the finance and accounting communities, saying, ‘If you’re going to talk about this, here’s the language, the metrics, and the narrative you should use.’ But they tend to ignore it, saying, ‘That’s not our problem.
It’s appalling that these kinds of things happen and that financial regulators turn a blind eye.
Pedro de Arteaga: Moving into the next section, I completely agree—it’s both finance and marketing that need to do their homework. Over the last two and a half years, the burden of proof has shifted from finance to marketing. Now, marketing is the one that has to explain and convince stakeholders. Marketers are expected to justify expenses and demonstrate their value while doing the right kind of work, yet they’re facing shrinking budgets and fewer resources. This could ultimately harm companies in the long run.
I truly believe the main reason for this shift is that we, as marketers, need to do a better job explaining what we do, how we drive growth, and how marketing is a growth driver, not just a cost center for the company.
Let’s assume for a moment that companies have the right CLV models in place. They’ve done their homework and have accurate calculations. I’ve come across companies that are really good at this.
Peter Fader: Exactly. It’s all about acquisition, retention, and development—what we call the building blocks of customer centricity. These concepts aren’t new; every company works on acquiring new customers, retaining them, and increasing their value over time. But the way these activities are executed, the organizational focus, and the accountability around them are often lacking.
In many companies, marketing spend is still divided into ‘brand versus performance,’ which can be misleading. While branding is essential, it’s not an end in itself. Brand investment doesn’t directly drive growth—it supports the key activities of acquiring customers, retaining them, and increasing their value.
What we aim to do is elevate these core tactics—acquisition, retention, and development—to a higher level in the organization, increasing visibility not just within marketing but across departments like finance, accounting, and even supply chain. Everyone should be aligned.
Often, the acquisition team is siloed from the retention team, leading to disconnection and inefficiency. The goal is to elevate and integrate these efforts, making them more visible and better coordinated, so they can effectively showcase and leverage Customer Lifetime Value.
It’s more of a cultural and organizational shift than a technical challenge, and that kind of change takes time.
Dan McCarthy: Then there are the acquisition-related tactics. You can start considering how to improve both acquisition and retention, and being instrumented for action becomes really helpful.
Pete and I co-founded Zodiac before Theta, and at Zodiac, our core focus was on helping with customer acquisition. While it could certainly be used for customer retention, acquisition is often easier to optimize first.
The idea is that different channels come with varying costs and bring in customers with different lifetime values. If you can get accurate estimates of both the costs and the value of customers acquired through each channel over time, it helps you assess performance. This allows you to reallocate your budget toward better-performing channels and away from less effective ones.
The same concept applies when slicing and dicing your business by segments. It doesn’t just have to be about Facebook vs. Google vs. TikTok—it can be about different business units, geographies, salespeople, or other segments. Understanding how your unit economics vary across these different areas can help you adjust your strategy and prioritize where to focus your efforts.
Peter Fader: There’s a great question in the chat about:
This is where the emphasis on lifetime value becomes crucial. If we can make Customer Lifetime Value as visible, tangible, and visceral as the cost of acquisition, we can achieve a more balanced perspective when evaluating acquisition activities. Rather than focusing on cost minimization—acquiring as many customers as cheaply as possible—we shift to value maximization.
Are we bringing in the right kinds of customers who offer long-term value?
Conceptually, it’s simple, but operationally, it’s much harder to get people to shift their mindset from minimizing costs to maximizing value. Even the smartest companies struggle with this change.
Peter Fader: It’s technically straightforward, but it’s really about culture and incentives—you’re exactly right.
We’re quant people, not cultural experts, but we’re gaining a deeper appreciation for the cultural element. So, how do we start that conversation? What experiments can we encourage companies to try? What analogies or metaphors can we create to help people see things differently and rethink how they organize their approach?
Dan McCarthy: We’ve been speaking with a public company where the person heading the division, who believed in this approach, has since moved on. But then, there’s that moment when you get an email from your boss’s boss saying, ‘We’re missing our number. We need to acquire this many active customers, and we’re not there yet.’ What do you do?
You can’t ignore your boss’s boss. Cultural elements, especially around incentive compensation, tend to dominate. You’re going to do what gets you paid, even if you believe another strategy would be better. If following that alternative path might cost you your job or put you at odds with leadership, it’s hard to stick with it.
When I first started, I was mainly thinking from an investor’s perspective. One of the benefits of customer-based corporate valuation is that it changes the conversation. When you start discussing which email to send to which customer, the CEO and CFO’s eyes might glaze over—they’re not that interested. But when you say, ‘This is how it impacts your valuation,’ suddenly they’re very engaged. We often end up speaking with heads of investor relations or the CFO specifically for those types of considerations. It’s like a Trojan horse: it opens the door to get people thinking about value creation in a different way. But without that hook, it’s tough to start the conversation.
Pedro de Arteaga: Dan, you’ve brought up customer-based corporate valuation, and of course, we can’t have this conversation without diving into that topic.
I want to start by reading a quote from your HBR article, which, when I reviewed it for this webinar, I found incredibly compelling. The quote reads:
‘Using customer metrics to assess a firm’s underlying value, a process our research has popularized, is called customer-based corporate valuation. This approach is driving a meaningful shift away from the ordinary but dangerous mindset of growth at all costs toward revenue durability and unit economics, bringing a much higher degree of precision, accountability, and diagnostic value to the new loyalty economy.’
What struck me is that this was written in February and March of 2020, just 12 to 15 months before the bubble we’ve all witnessed in the digital economy.
Dan McCarthy: Yeah, it’s tough, but it’s funny to think back to when we were working on these IPOs that Pete mentioned. We’d run the numbers on company after company, and at the time, many of our implied valuations aligned with the IPO valuations—though not always. Slack and Revolve were notable exceptions.
So, the question was: who’s wrong?
We had a fair number of people saying we were wrong because there was this ‘X factor’ we weren’t accounting for, leading us to seem systematically pessimistic. But given how things have played out, I’d like to think that the numbers we presented have been vindicated.
Peter Fader: Today, we’re almost facing the opposite problem. Many private equity firms are sitting on a lot of dry powder, and there are some great values out there, but now they’re too hesitant to take the risks they would have routinely taken just a couple of years ago.
The pendulum swings. One of the benefits of this approach is that it brings stability and rationality. It can help not only identify ways to make more money and find that alpha, as finance folks say, but also reduce volatility and the herd mentality—where a certain sector or business model is suddenly ‘cool’ and then just as quickly, it’s not.
This method allows us to cut through the noise and simply ask: Do they have good customers or not? period.
Dan McCarthy: Yes, customer-based corporate valuation explicitly ties customer activity to a firm’s overall valuation. This relates to one of the chat questions: How do you connect CLV to the P&L? You can visualize it as a matrix, where every row represents a customer and each column represents a point in time.
In each cell, you have data like the number of orders, the spending from those orders, and the contribution profit generated from each customer. CLV is essentially about looking at the rows of that matrix—discounting the values to calculate the lifetime value for each customer.
On the other hand, finance professionals tend to focus on summing the columns for total revenue and variable profit. If you sum all the revenue across customers within a specific time frame, you get the total revenue for that period. Our models aim to predict the entire matrix. This matrix not only provides the CLVs but also gives finance teams the period-by-period revenue and, ultimately, cash flow projections.
So, the way we connect CLV to the P&L is by predicting the matrix, which in turn gives you everything else—whether it’s summing rows for customer value or columns for revenue.
Peter Fader: Another way to think about it is that people often see revenue as something tangible, an entity in itself. But it’s not. Revenue is simply the result of how many customers we’ve acquired—how many rows we’ve added to that table—how long they stay, how often they engage, and how much profit we make when they do.
By breaking revenue down into its components, understanding and projecting them, and then adding them back up, you can forecast revenue, cash flow, or EBITDA over a longer horizon with greater accuracy and deeper diagnostic insight.
For example, if a company’s sales are plateauing, is it because we’re not acquiring enough customers, they’re not staying as long, or they’re not buying as frequently?
As Dan likes to say, it’s still an accounting identity. It’s irresponsible not to examine this revenue decomposition if you want to uphold your fiduciary responsibilities to your stakeholders.
Dan McCarthy: Often, CBCV is seen as the framework that can’t be wrong—it has to be right. The key premise is whether you believe that a company’s value comes from future cash flows. If we can’t agree on that, we can’t even begin. But once you accept that idea, you have an accounting identity that tells you what the future cash flows will be, and that has to be right.
Now, our predictive model might not always be accurate. We could make a forecast in February 2020, and then something unexpected, like COVID, happens.
That’s why it’s important to distinguish between the framework itself, which is solid, and the predictive model we build based on historical data, which can be subject to unforeseen events.
Pedro de Arteaga: What are the building blocks? Let’s say I want to start tomorrow.
Dan McCarthy: When conducting a customer-based audit, I would start with historical data first, then move on to predictions. We’ve often broken it down into key components: acquisition, retention activity, repeat purchases, spending, and variable profit. These are the main drivers of a model like this, so they’re what we focus on.
Step one is understanding these components. Step two is to analyze what those numbers looked like in the past—by segment, cohort, channel, etc. This provides a clear picture of where the business has been. The predictive model can then project that into the future, helping to inform valuation and prioritize key segments.
Starting with historical data can be incredibly clarifying. The other benefit is that it’s easier to get executive buy-in because you’re not making assumptions about the future. Executives may be skeptical if they’re not used to viewing the business this way, but if you can show them, ‘These are the historical numbers; this is just the truth,’ it can create that ‘wow’ moment.
This is often more compelling than presenting projections based on assumptions that stretch into the future.
Peter Fader: Thanks for the shout-out, Dan. My third book, The Customer Base Audit, goes beyond just being the foundation for predictive models; from a cultural standpoint, it’s a powerful tool for gaining buy-in. By conducting a straightforward, data-driven audit, we signal that we’re committed to transparency and accountability. This ‘boring’ customer-based audit helps us see what’s happening with the company—comparing this quarter to last—and identify areas that might need attention.
It’s a great way to show the organization that it’s not just about executing flashy tactics or marketing strategies. It’s about understanding the true lifeblood of the business, through metrics that everyone should care about.
Starting here is a natural step toward driving both tactical decisions and aligning more closely with finance.
Peter Fader: That’s what’s so great about starting with historical data—there’s no debate over which model to use, no arguments about the discount rate. It’s all data that everyone can agree on.
Then, we can start exploring the insights, like seeing if the more recent cohorts are less valuable than previous ones, or analyzing by acquisition channel or first-order product.
A great example, which we cover in a whole chapter in The Customer Base Audit, is when we line up all the products from best-sellers to worst-sellers. We often focus on the best-selling products, but it turns out that many unprofitable customers purchase them. They come once, buy the popular item, and never return.
Meanwhile, products further down the list, though selling in smaller volumes, are bought by highly profitable customers. When we shine a light on these operational aspects, we can see where the real value is—not just the volume. This gives us immediate, actionable insights on the types of tactics we should prioritize.
Dan McCarthy: We’ve found that, with many of the companies we work with, it’s not as simple as snapping your fingers to get your CAC or cohort data. It usually involves a lot of work. Some mature companies already have this data, but many don’t, so there’s this necessary first step of gathering the correct data to train the models. You need to start there anyway.
Another effective approach has been to begin with charts that provide model-free evidence of the vast heterogeneity in customer value. Often, managers don’t fully appreciate how much of their revenue comes from a very small percentage of their customer base.
With the right charts, you can illustrate this clearly. You take a cohort of customers, sort them by value, and then look at what percentage of your revenue comes from the top X percent of customers. Seeing that visually can be striking, especially if it’s something they’ve never considered before.
Pedro de Arteaga: Unfortunately, we’re nearing the end and running out of time. I could discuss this all day long, but I don’t want to conclude without putting the two of you in the shoes of our audience. I imagine many CMOs, VPs of Growth, and Growth Directors have to navigate the challenges of convincing finance influencers while also working to shift mindsets.
So, I’d like your advice on two key points:
Dan McCarthy: It can vary depending on the situation. My default approach would be to focus on getting the CEO and CFO interested. To do this, we should present compelling, data-driven insights about the customer base and its value. This can serve as a strong entry point and accelerate the cultural transformation within the company.
Starting with the CEO and CFO is likely to be more effective than beginning with others, as their buy-in can drive broader organizational change. The specific approach may differ based on the company’s financials and unique context, but the goal is to present something impactful that they haven’t considered before.
Peter Fader: It’s fascinating how our conversation has evolved over the past decade. When Dan first walked into my office, our discussions were focused solely on models, forecasts, and technical details. Now, we spend a significant amount of time discussing cultural issues.
Personally, I used to think that if I could just provide a “CLV magic wand” to show each customer’s projected value, it would be straightforward to figure out what to do, and the financial benefits would follow. I now realize that this perspective was overly simplistic. The reality is much more complex. Jumping straight into the technical aspects of models and discount rates can sometimes miss the core point.
The real challenge is about getting people on board with understanding the diverse value of their customers. We need to start with broad, qualitative insights that are easy to verify and act on. Once people see the value in understanding their customer base better, they’ll be more inclined to explore CLV in depth.
The goal is for them to come to us, eager to delve into CLV. It’s up to us to spark that interest and make them ask the right questions and seek out the kinds of models we’re passionate about running.