CLV as the yardstick for successful customer relationships

Companies can only be successful if they manage to establish successful and profitable long-term customer relationships; Customer Lifetime Value (CLV) is the KPI that enables companies to focus on this aspect. What is meant by CLV is the value (measured in terms of sales or profit) that a customer generates for the company throughout the relationship; in practice, this is often for a specific period of time.

In simple terms: If a customer stays with a company for 10 years and spends CHF 1,000 annually on its products or services, their (sales-related) CLV amounts to CHF 10,000. Thus the CLV comprises a time component as well as a value component. It’s important to stress that “lifetime” refers not to the time during which the customer is actually alive, but the length – and in other words the lifetime – of the relationship itself. When analyzing the CLV, one often looks at a limited period of time because doing so over an excessive period serves no purpose either methodologically or with regard to any decisions that have to be made.

The advantages of CLV, together with opportunities to use it

Why is CLV so relevant for companies? In short-term assessments of value which are frequently undertaken (for example, by examining current sales/profits figures based on the previous year), the fact that customer relationships can develop in quite different ways is neglected: one refers here to various “relationship models” (cf. Fig. 1). Some customers make many purchases at the start of the relationship but then break it off again, while others are initially cautious to give it a try but then turn into valuable long-term customers. Many customers only become profitable after a certain period has elapsed – if the marketing and/or acquisition costs are correctly allocated.

CLV variants and how they are applied

Sales vs. profits

As with single-period approaches to customer value, the value component of CLV can be based purely on sales or alternatively on contribution margin and/or profit (while simultaneously taking costs into account). The latter (profit) is particularly useful if the costs per customer differ greatly (e.g. product, support and marketing costs). However, many companies find it difficult to allocate costs to individual customers.

Constant vs. variable contributions to value

In our initial example, we assumed constant contributions to value (CHF 1,000 annual sales). This assumption isn’t unrealistic in some business models such as relatively straightforward subscription models (e.g. Netflix). In such cases, CLVs vary solely as a function of the length of the relationship. In many other cases (e.g. the retail trade), the sales attributed to a given customer can vary greatly over the course of time, so one needs to look at the individual sales per period, as shown in Fig. 2.

Past vs. predicted CLV

The explanations above are based on a past-oriented approach to value (past CLV) which is suitable for retrospectively rating individual customer relationships as well as the entire customer base. It allows one to make statements about the contribution to value that individual customer relationships have hitherto generated.

The intention might be to use CLV to make investment-related decisions (for example, which customer relationships merit more investment, and which less?). If this is the case, past CLV is of only limited significance because of course one should invest in those customers who show the most potential for the future. This is where the future-oriented approach to value (future or predicted CLV) comes into its own because it uses data science to estimate future customer value. In order to do this, one uses predictive factors (i.e. indicators) which allow one to make statements about future value. For example, these factors include:

• Previous customer behavior: Even if one should assume that customer relationships don’t always develop in the same way, previous customer behavior is nevertheless relevant.
• Personal customer characteristics: A variety of personal attributes can help to predict customer value; they include age, income, place of residence, number of children, and so forth. This information isn’t always available, so it’s beneficial for companies to obtain such information wherever possible.
• Acquisition channel: A variety of studies have identified the predictive power of the acquisition channel. Thus, for example, acquisition in the context of a discount promotion can point to lower future value, whereas acquisition due to recommendations by friends and acquaintances has been shown to impact positively on customer value.
• And there are many other factors…

It’s often the case that the time and value components of CLV are predicted separately, and the predicted CLV is derived from this.

In order to comply with investment theory, when determining future CLV one should - strictly speaking - discount future values in order to derive the so-called present CLV.

Individual customers vs. customer segment

The core competency with regard to CLV is of course to determine individual customer values. However, for some decisions it might suffice – or even make sense – to look at a customer segment. For example, if one knows that a specific segment manifests a very high average CLV, this information can help companies to prioritize this segment in the context of marketing. When it comes to new customer acquisition, there’s often little individual information about potential customers. A value-oriented focus on specific segments can therefore already allow companies to make important progress.

Challenges and future perspectives

Companies face a variety of challenges when it comes to ascertaining and then utilizing CLV. Some of them are as follows:

• It isn’t always sensible to make customer-related decisions based upon CLV. An airline’s Customer Value Manager reported a situation where the check-in agent was able to award an upgrade which should have been undertaken in accordance with CLV. Two passengers stood in front of her: an elderly CEO and a young junior executive who, purely due to his age, had been assigned a higher CLV in the system. But who should receive the upgrade?
• Ascertaining CLV (and in particular determining future CLV) are associated with significant data and analysis requirements, yet it’s worth making the effort to fulfil them. Moreover, CLV analyses also benefit from new methodological approaches in the field of machine learning and AI.
• If decisions are made on the basis of predicted CLV, they should normally impact upon customer behavior and thus in turn upon predicted CLV. Strictly speaking, these dependencies that are inherent in the analysis must be considered in CLV modelling.

These and similar challenges and issues make CLV a promising area of research where numerous questions have yet to be addressed.

Would you like to learn more about the topic of customer value? If so, please contact the experts in person: georgi@customer-metrics.io; adriana.ricklin@hslu.ch

Further reading

• Bruhn, Manfred; Georgi, Dominik; Treyer, Mathias; Leumann, Simon (2000): Wertorientiertes Relationship Marketing. Vom Kundenwert zum Customer Lifetime Value in Die Unternehmung, Vol. 54. (2000), No. 3, pp. 167–187.
• Bruhn, Manfred (2022): Relationship Marketing, 6th Edition, Munich.
• Georgi, Dominik (2000): Entwicklung von Kundenbeziehungen, Wiesbaden.
• Georgi, Dominik (2007): Werttreiber in Kundenbeziehungen, Postdoctoral thesis, University of Basel.
• www.hslu.ch/de-ch/wirtschaft/institute/ikm/
• www.customer-metrics.io