A disruptive new model shows firms and investors where the money is

Today’s consumers are living in an ever-expanding web of contracts. The typical consumer of the 21st century is bound by agreements with mobile phone services, internet providers, streaming services, home security companies, health clubs, appliance maintenance firms, auto leasing and financing, and the list goes on.

These businesses all share a common goal: they want more customers, but more importantly, they want them to stay for longer periods of time. Firms typically weight their budgets more heavily toward customer retention than customer acquisition.

That allocation is wrong-headed, says Professor Peter S. Fader of the University of Pennsylvania’s Wharton School. Over the past decade, Fader and his colleagues have come to believe that the traditional investment focus on enticing existing customers to stay longer needs to be upended. In a soon-to-be-published paper, Fader and four co-authors unveil a methodology that supports their contention that the key to finding the customers who are “born” with greater propensities to be loyal, instead of trying to nudge them in that direction repeatedly over time.

In “’How to Project Customer Retention’ Revisited: The Role of Duration Dependence,” slated for publication in the Journal of Interactive Marketing, Fader and his co-authors – Bruce Hardie of London Business School, Yuzhou Liu at Man Numeric, Joseph Davin at Two Six Capital, and Thomas Steenburgh at University of Virginia’s Darden School of Business – introduce their beta-discrete-Weibull (BdW) distribution model, which provides the technical foundation that substantiates their theory.

An earlier model, the beta-geometric (BG) described the phenomenon of why cohort-level retention rates increase: much of the so-called “retention dynamics” arises as a cohort of heterogeneous customers “shakes out” and sees an ever-increasing proportion of good customers over time – even if each individual customer doesn’t change with time. The BdW model advances the discussion by allowing for differences among people as well as changes over time in a discrete time setting. The managerially relevant conclusion, according to the authors, is that “accounting for cross-sectional heterogeneity [among customer cohorts] is more important than accounting for any individual-level dynamics in churn propensities.”

Fader explains, “Common wisdom says that the longer a customer stays, the more we should spend on locking them in and finding ways to make them love us. But that’s not always true. In fact, I’ve never seen a case where that can be true, and I doubt it will happen very often in the real world.”

Instead, he says, people are inherently different, in that some are “sticky” and others “flighty.” Companies would do better to spend a greater share of their budgets on acquiring not just quantities of customers, but quantities of the right kind of customer – i.e., the sticky kind.

Finding and recruiting the kind of customer who sticks around will mean investing more in predictive analytics than in retention promotions, Fader says. In this way, firms can grow their customer equity by increasing the proportion of sticky customers.

While the BdW model demonstrates that sticky customers tend to become slightly less loyal over time, this effect is “very minor,” according to Fader. What is extremely clear is that “in a heterogenous market, it is practically unlikely that customers will become more loyal over time.”

The practical implications of this research are massive. The obvious application for firms is to revamp their budgets. The current practice is to spend relatively little on customer acquisition and the lion’s share on customer retention and development. Dollars would more wisely be spent on finding the kinds of customers who are naturally inclined to be loyal, Fader says.

These insights not only apply to marketers, but also to finance professionals. In other related projects, Fader and colleagues are focusing on the emerging notion of “customer-based corporate valuation,” in which they perform “bottom-up” analyses on publicly traded subscription-based companies (such as Dish Network and SiriusXM) to demonstrate how customer renewal/churn patterns can be directly linked to market valuation. In such an exercise, it is vitally important to clarify the differences across customers from the differences within customer over time. “The key is to get below the surface of the data and sort out these underlying behavioral patterns. The CMO and the CFO should both be interested in knowing what the company is worth based on the kinds of customers it has,” Fader says.