Customer Lifetime Value

Graphics courtesy of GmavenOpens in new window The hallmark of CRM ideaOpens in new window is that a customer should not be viewed as a set of independent transactions but as a lifetime income stream. In the auto industry, for instance, it has been estimated that a General MotorsOpens in new window retail customer is worth $276,000 over a lifetime of purchasing cars (11 or more vehicles), parts and service. Fleet operators are worth considerably more.

When a GM customer switches to FordOpens in new window the revenue streams from that customer may be lost forever. This makes customer retentionOpens in new window a strategically important goal for GM.

Customer lifetime value (CLV) is even more important if you consider that a small number of customers may account for a high proportion of the entire value generated by all customers.

Tukel found that the top 28 per cent of customers generate 80 per cent of the total value of all customers.

Customer lifetime value (CLV) is a measure of a customer’s profit-generation for a company. CLV can be defined as follows:

CLV can be estimated at the level of the individual customer, customer segment or cohort.

This type of cohort analysis is useful, for example, to find out whether certain channels are more effective or more efficient at recruiting high value customers.

A European motoring organization knows that it costs an average of $105 to recruit a new member. However, recruitment costs vary across channels.

The organization’s member-get-member (MGM) referral scheme costs $66, the organization’s direct response TV campaign costs $300, and door-drops cost $210 per newly acquired member.

The MGM scheme is most cost-effective at customer acquisition, but if these customers churn at a high rate and cost significantly more to serve, they may turn out to be less valuable than customers generated at higher initial cost. In fact, customers acquired through the MGM referral scheme remain members longer, buy more and also generate word-of-mouth referrals.

The focus on free cash flow rather than gross margins is because a customer that appears to be valuable on the basis of the gross margins generated can become less profitable once cost-to-serve the customer is taken into account. Companies that do not have the processes in place to allocate costs (e.g. ABC — Activity Based Costing) to customers cannot use free cash. They must work either with gross margin or sales revenue data.

Research by Reichheld and Sasser indicates why it is important to look forward to compute CLV. Their data suggest that profit margins tend to accelerate over time, as shown in Table X-1. This has four causes.

Causative factor #1

Revenues grow over time, as customers buy more. In the credit card example in Table X-1, users tend to grow their balances over time as they become more relaxed about using their card for an increasing range of purchases. Also, a satisfied customer may look to buy additional categories of products from a preferred supplier.

An insurance company that has a loyal car insurance customer is likely to experience some success cross-selling other personal lines, for example home, property and travel insurance.

Causative factor #2

Cost-to-serve is lower for existing customers, because both supplier and customer understand the other. For example, customers do not make demands on the company that it cannot satisfy. Similarly companies do not communicate offers that have little or no value to customers.

Causative factor #3

Higher prices may be paid by existing customers. This is partly because they are not offered the discounts that are often promised to new customers, and partly because they are less sensitive to price offers from other potential suppliers because they are satisfied with their experience.

Causative factor #4

Value-generating referrals are made by satisfied customers. Every customer not only has their own CLV, but also, potentially, a Customer Referral Value or CRV. That is, satisfied customers can generate additional value for their supplier by giving positive word-of-mouth to their friends and associates. Word-of-mouth can be powerfully persuasive when it is regarded as independent and unpaid.

Computing CLV

The computation of CLV potential is very straightforward in principle, but can be complicated in practice. Several pieces of information are needed. For an existing customer, you need to know:

1. What is the probability that the customer will buy products and services from the company in the future, period-by-period?
2. What will be the gross margins on those purchases period-by-period?
3. What will be the cost of serving the customer, period-by-period?

For new customers an additional piece of information is needed:

4. What is the cost of acquiring the customer?

Finally, to bring future margins to today’s value, another question needs to be answered for both existing and new customers:

5. What discount rate should be applied to future net margins?

Some commentators suggest that CLV estimates should not be based on future purchasing only, but on word-of-mouth (WOM)Opens in new window influence too. The logic is that a satisfied and retained customer not only buys but also influences others to buy. Lee and colleagues show that incorporation of WOM effects increased customer value significantly.

Not all customers with the highest CRV at a telecommunications company were not the customers with the highest personal CLV, but those in the mid-levels of CLV. For example, customers in the top decile with an average CLV of $1,933 generated an additional $40 of value through referrals, whilst customers in the 5th decile, having a CLV of $313, generated $1,020 of referral value for their telecommunications. As more and more customers interact with each other over social media, the impact of WOM is likely to grow substantially.

Computation of a meaningful CLV estimate requires companies to be able to forecast customer buying behavior, product and service costs and prices, the costs of capital (for determining the discount rate) and the costs of acquiring and retaining customers. This is very demanding, especially at the level of the individual customer, but is a challenge that analytical CRM implementations often take on.

A number of companies have developed models that produce approximate CLV estimates. US Bancorp, for example, calculates a customer profitability metric called Customer Relationship Value (CRV) in which they use historical product ownership to generate “propensity to buy” indices. Overhead costs are not factored in to the computation. Within their customer base, they have been able to identify four CRV segments, each having different value, cost, attrition and risk profiles.

Each of these segments is treated to different value propositions and customer management programs: product offers, lending decisions, fee waivers, channel options and retention efforts.

North Carolina’s Centura Bank has two million customers. The top customers receive special attention from service staff and senior management, including an annual call from the CEO.

For situations where the cost of generating accurate CLV data is thought not to be prohibitive, Berger and Nasr have developed a number of mathematical models that can be used in CLV estimation.

Table X-2 shows how to compute CLV for a cohort of customers. In year 0, the company spent $10 million in marketing campaigns to generate new customers. The result was 100,000 new customers added to the customer base at an acquisition cost of $100 per customer.

In year 1 the company lost 40 percent of these new customers, but the remaining 60 percent each generated $50 contribution to profit. If this is discounted at 15 percent, each retained customer’s profit contribution is $43.48. In year 2, the retention rate rises from 60 percent to 70 percent, and each of the remaining customers contributes $70 ($52.93 at discounted rate) to profit.

You can see from the right-hand column in Table X-2 that it takes nearly five years to recover the costs of acquiring this cohort. The data demonstrate a couple of well-established phenomena. First, profit per customer rises over time — for reasons set out earlier. Second, customer retention rate rises over time.

It is feasible to use data such as these to manage a business for improved profitability. Several strategies are available:

1. Improve customer retention rate in the early years of the relationship. This will produce a larger number of customers to generate higher profits in the later years.
2. Increase the profit earned per customer by:
   a) reducing cost-to-serve;
   b) cross-selling or up-selling additional products and services.
3. Become better at customer acquisition by:
   a) using more cost-effective recruitment channels;
   b) better qualification of prospects. Customers who defect early on perhaps should have not been recruited in the first place.
   c) careful nurturing of prospects with high CLV potential; d) recruiting new customers matched to the profiles of current customers having a high CLV.

You should not leave this discussion of CLV by believing that if you improve customer retentionOpens in new window, business performance will automatically lift. It depends entirely upon which customers are retainedOpens in new window and how you manage those relationshipsOpens in new window. We have more to say about customer retention here.