Increase customer profitability with machine learning
The complexity of Customer Lifetime Value (CLV) -calculation limits practical implementation. Can we use machine learning to overcome this challenge? Read new blog by Stian Daazenko.
Service innovation changes the way we measure profitability, with Customer Lifetime Value (CLV) as the key number for businesses to track. However, the complexity of CLV-calculation limits practical implementation. Can we use machine learning to overcome this challenge?
Service innovation puts customers in the center of profitability analysis
Service innovation thinking requires us to take the customer’s perspective (outside in thinking) when business decisions are to be made. At CSI we believe this also changes the way we measure and evaluate profitability.
This was the topic of a 2016-article written for Magma that I co-authored with a fellow researcher (article in Norwegian). Some of the key recommendations in this article are:
- The value of your customer relationships, and not the value of individual products / services, should form the basis for investment decisions
- It is the future value of the customer relationships, and not the customers' historical profitability, that matters
- Customers have different profitability potential and should be segmented accordingly. Increased profitability is achieved by targeting customers with the largest economic potential
Based on these points it is recommended to establish Customer Lifetime Value (CLV) as the key number to track, which is a well-established concept in the service literature.
CLV is a powerful number, but difficult to implement
CLV can be defined as the future cash flow from a given customer relationship. The sum of CLVs for all current and future customers is called Customer Equity (CE) and represents the total value of a company’s customer base.
CLV has several fields of application, including:
- Identify the right customers to spend money on and where the potential for increased profitability is largest
- Deliver personalized customer service and marketing to improve customer experience and boost sales
- Create a business case for investments. ROI is then represented as the change in Customer Equity
However, the data needed to calculate CLV does not come for free. In its simplest form CLV still requires estimation of future contribution margins and retention rates, as well as a discount rate to create the discounted cash flow, exemplified by the typical CLV definition:
In practice we often need more fine-grained representations of CLV to take proper decisions. The construct of the calculation itself also vary, dependent on the business environment we are in.
Hence, CLV is not a trivial number to estimate, and requires relatively complex modeling and calculation. We believe this is an important reason why CLV has had limited practical implementation. Model simplification is often proposed as the solution, but is there another way?
Can machine learning be the solution?
Machine learning is a sub-area of artificial intelligence that enables computers to get into a mode of self-learning without being explicitly programmed. When exposed to new data, computer programs are enabled to learn, grow, change, and develop by themselves. With the introduction of more powerful computers and increased availability of data, machine learning has started to gain ground as an important decision tool for business.
We believe CLV-calculations is an excellent case for machine learning. Based on available customer information, the computer can identify patterns, build models and estimate CLV values – much faster and with more complexity than a human can handle. The model can also self-improve as more data becomes available. This represents a new way of working with CLV-calculations, where the model and key parameters are defined along the way:
Too difficult and expensive, you say? Contrary to some beliefs such a model can be created and set up relatively easy, with low investment costs. There are examples of Norwegian companies that already have a fully functional CLV-model running, with very promising results. We therefore believe this is well worth some attention in further CSI research.