More Dollars per Dollar!

Introduction

TEG Analytics has already talked so much about CLV (Customer Lifetime Value), at EMC 11, webinars and through publishing thought notes, that you would have picked the cues and have already started implementing it. Remember to incorporate factors such as weather, usage forecast, demographics data, plan preferences and attrition profiling among few more to get accurate value. CLV assists business in decision making at each stage of customer lifecycle depending. Below, we present 3 CLV applications in retail energy market setup.

Fig. CLV Application in retail energy market

Look-Alike Analysis

CLV provides a means to divide customer base from most profitable to least profitable segments. There are various ways of clustering namely – Hierarchical clustering and Gaussian Mixture Models which can be applied on top of CLV based segments to understand customer characteristics better. Moreover, clusters can be used to construct persona at different levels like channels, usage, plans, zip codes and utility to differentiate characteristic of profitable customer segments from the rest. It thereby enables a forward-looking approach for business to realign its marketing spend better.

A retail business can utilize those personas to acquire the desirable set of prospects who will yield better margins and will stay for a longer duration. This exercise is known as look-alike analysis. It gives a good predictive power to business for mapping prospect behavior across multiple variables and to later service them as per their cohort’s needs. Taking acquisition costs in consideration, CLV based profiling helps a retailer to know how much it can spend to acquire a similar customer while maintaining profitability.

Personalization

Personalization refers to the ability of a business to spot its customers’ preferences and to engage with them accordingly. Personalized offers should be made to upsell or cross-sell, depending upon the cohort and lifecycle stage of customer. High CLV cohorts should be given preference for them to stick longer with the business. This enables business to create dependency for such customers and ultimately leads to longer retention period. Factors like demographics, CLV, customer interactions help devise various types of personalization such as Time of Use, Behavioral, Channel/Mode, Geographic and frequency among others.

Cost of servicing is another important metric which needs to be tracked in a subscription-based model. Cost to serve comes down as the tenure of customer increases but if renewal rates decline then average cost to serve rises which adversely impacts the bottom line. Therefore, for a retail business it is imperative to determine important customers for aligning available resources to garner maximum returns and to keep their cost to serve under control.

Pricing Corridor

In subscription-based business such as deregulated energy market where there is minimal switching cost, customer experience makes a retailer stand out. Here, CLV helps retailer to devise plans with features which are preferred by profitable cohorts across geography. On top of it, classification models such as Random Forest Regression or CART come in handy to discover what factors impact attrition rate the most.

As part of calculating CLV, customer-level churn probability can be calculated which can identify most vulnerable set of customers. Considering cost of servicing, margin, usage, bill rate and other similar factors for such customers, pricing corridor can be developed for different plans. Retailers can leverage pricing corridor to price their plans at optimum rate to avail maximum gross margin.

Conclusion

Predictive customer analytics empower every retail energy provider to utilize its internal data along with publicly available sources to acquire/service desired set of customers. As these customers continue to stay with the business, in the long run, it will accrue better margins. This is the key to making more dollars per dollar.