Interview: Dr Daniel McCarthy

Columbia Road: Can you briefly explain the CBCV model and its benefits?

Daniel: The CBCV framework uses historical data and a collection of state-of-the-art statistical models to predict the lifetime value of each customer. It does this by forecasting the future stream of revenue and variable profits per customer over the customer’s projected lifetime. Performing these calculations across all existing customers, factoring in new customers that will be acquired in the future and accounting for other standard financial factors, for example capital structure and cost of capital, we arrive at a CBCV for the company. Of course, doing so also gives us a number of other key financial KPIs “for free”, such as total revenue, customer lifetime value (CLV), CLV relative to customer acquisition cost (CAC) and how these quantities have been varying across acquisition cohorts.

The beauty of this framework is that although the model gives us an overall valuation, it also provides us with individual-level estimates for the value of each and every customer in the customer file, so if other company stakeholders, for example marketing managers, want to break it down — even to the level of individual customers — then they can, with everyone working off of the exact same model, building a bridge between marketing, finance, and other divisions within a company. Ultimately, it offers a new way of defining company valuation that is derived from the existing and potential customer base along with the cost to serve each customer, rather than looking at market size and historical revenue and profit figures.

CR: Do your clients automate CBCV to track predicted lifetime value and costs for individual customers, and can they use this to track the overall company valuation?

Daniel: Yes, some companies do, at least for the revenue and cost data. Having that available in near real time isn’t easy because of the difficulty of attributing things like product returns to specific customers. We’ll often just do a one-shot analysis, but as you allude to, a company can really take this to the next level by building it into their company’s DNA. If it’s helpful, I can take you through an example of one of the analyses that we did on AT&T to get a better sense of what the valuation process looks like.

CR: Yes, please tell us more about the AT&T case...

Daniel: For AT&T we ran two separate valuations, one for all postpaid customers and one for all prepaid customers. Although the prepaid segment was growing rapidly, the expected lifetime value of the customers within it was much lower, meaning that most of the company’s value was coming from postpaid customers. That bigger-picture view showed us that the success or failure of the company is likely going to be driven by what happens in the postpaid segment.

Breaking down a firm’s value using the expected lifetime value of both existing customers and those yet to be acquired allows us to peel back the onion and analyse the company’s value in a much more granular and diagnostically relevant way, helping us to understand what might happen to overall valuation if some of the drivers were to change. To use a hypothetical example, if AT&T’s prepaid segment were to see a 10% drop in the expected lifetime of new customers, we could quantify exactly what that would mean for AT&T’s overall valuation. If we were inside the company, we could then examine what drove this reduction in expected lifetime and what remedies we could consider to get this back on track. Management can put together a customer dashboard to track these predictive KPIs, in addition to the historical ones, to know how the health of their customer base is evolving over time.

CR: How does a CBCV-inspired approach affect the work and focus areas of sales and marketing teams?

Daniel: The main way is that we could move prospects through the adoption funnel more efficiently, so we’d be acquiring more new customers for the same amount of marketing dollars, resulting in a lower CAC. Ideally, we would be bringing in higher-value customers, meaning their post-acquisition value would rise, too. We would then set up a workbench through which we could start to optimise marketing spend allocation in this long-term customer value-oriented way. That’s the sort of thing we’d recommend. It can be used to get the CFO on board with marketing initiatives because CBCV gives them a more direct view of how customer-focused activities affect the overall company valuation.

CR: Can the model help with investment decisions across the whole company, or is it just for sales and marketing?

Daniel: Yes, the CFO already thinks about proposed initiatives — typically capital projects — in terms of measures like return on investment, payback period and internal rate of return. The more we can translate our marketing activities into their language, using these same sorts of measures, the easier it will be to sell marketing projects to them. An aspect that’s often ignored is short term versus long term — there are many initiatives where the return is positive in the short term but it won’t be in the longer term. An example of this could be routing customers into an interactive voice response (IVR) system when they call up. Although this can offer significant savings in the short term, it can also lead to a lot of frustrated and unhappy customers — so it’s really important to estimate not just the reduction in costs, but also the impact on customer retention.

CR: Does your work lead to a more data-centric approach to everyday sales and marketing operations for your clients?

Daniel: Facilitating AB testing is one very valuable thing that this framework can offer. When we run an AB test on an initiative we can establish its net effect on the average customer then use that information to work out the return on investment. For example, we might want to find out if implementing call centre automation positively affects the customer value. If we can randomly route some customers calling in to the more automated solution, while routing everyone else to the incumbent solution, running the right predictive models on both groups and taking the difference between the two enables us to get the impact of that automation initiative on the long-term value of the customer, and how it may vary across customers.

CR: Does doing this kind of analysis help people to see sales and marketing initiatives in a different way?

Daniel: Yes. For each initiative you have a base case valuation and a full potential valuation. Without any fancy instrumentation, you can still manually run the numbers. Pilots such as this can allow you to generate some early wins and in turn some excitement about this new way of looking at the world. Once you’ve successfully executed this process a few times you can clearly demonstrate to management why it’s important to have an experimental platform and continue running these “test and learn” experiments with a broader set of activities. Netflix and Stitch Fix are great examples of companies who’ve made it easy to run lots of experiments to see how their initiatives perform. They’ve spent a lot of money to do this, but companies can achieve similar results with less extravagant methods — you don’t always have to re-invent the wheel as there are many existing solutions on the market. For example, Optimizely is purely focused on AB testing, while Pega Systems has relevant workbench capabilities. Using existing tools like these effectively, you can create your own “laboratory” in which you can run the experiments and present the results in an easy-to-interpret way.

CR: Have you seen companies turning their public reporting and valuation principles into a CBCV-inspired model?

Daniel: The whole disclosure question is interesting. We’ve been working with a major bank that’s now planning to disclose customer data for the first time and we’re very excited to be a part of it. They see disclosure as a way to close that valuation gap and get the credit they feel they deserve for their valuable customer relationships. If companies hold themselves accountable in this way it also adds transparency and puts pressure on them to demonstrate that they can maintain or improve their CBCV over time, which is good news for everyone.

---

DANIEL MCCARTHY is an Assistant Professor of Marketing at Emory University's Goizueta School of Business. Alongside his colleague Peter Fader, Daniel developed the customer-based corporate valuation (CBCV) model, and the two founded Zodiac (which has since been acquired by Nike) and Theta to make the model’s predictive customer analytics commercially available to companies.