Back to the presentation by Nicolas Glady, Professor ESSEC’s Accenture Strategic Business Analytics Chair, who spoke at the Adetem B2B Club meeting last September about CLV, Customer Life Value.
Only 17% of the B2B marketers surveyed in the survey use CLV as an indicator of marketing performance in France. It is, however, an indicator that has the enormous advantage of being able to calculate the performance of marketing campaigns in the future, even over several years – unlike the other indicators, which are reduced to the duration of the fiscal year. Admittedly, it is not easy, and we can see that we need analytical and even mathematical skills. But the good news is that there are experts like Nicolas who know how to do it and practice it in companies.
Nicolas, can you define the CLV?
The CLV (or Customer Life Value) is the discounted value of future profits that a customer will generate. By analogy, a customer’s CLV can be compared to the NPV of any other type of asset: this is the financial value that an individual customer value has for the business. Because these are future profits, and these are very often uncertain, predictive models are needed to estimate future business and profitability of customers.
Does the Customer Life Time Value make sense and can be calculated in B2B?
Distinguishing customers who will be more profitable from those who will be less profitable is as relevant in B2B as in B2C. In addition, understanding the dynamics of this future profitability is particularly enlightening for the decision-maker. It should be noted that Customer Life Value is also used in many contexts to estimate the impact of marketing campaigns …
What are the problems faced by these B2B companies?
CLV is often more complex to predict in a B2B context than in B2C for three reasons:
- The uniqueness of some customers,
- Non-linearity of some effects
- the particular relationship a client has with his relationship manager.
First, modern techniques (such as Bayesian analysis) can be used to model the behavior of individual clients, even in the case where there are few data. Then, some models (eg decision trees) are particularly suitable when the effects are non-linear (eg when threshold effects are expected). Finally, it is essential to take into account the relationship between the account manager and his client in a quantitative way. For example, it is possible to measure the impact of past visits on the future profitability of a customer …
Which model do you recommend?
It must be possible to take into account the issues highlighted above, but also to guarantee that the model will be understandable by the decision maker and that it will be simple to use in practice. This often requires design efforts, and this brings us to the issue of talent. The “Business Analytics” expert must have four sets of skills:
- be able to use computer tools,
- be able to analyze the data, but also (and these points are too often forgotten by those who are unfamiliar with the field)
- understand the business problem in which the Analytics solution fits
- to simply explain to the various stakeholders the ins and outs of the proposed solution.