Use of Lifetime Value Data

Lifetime value has been adopted by companies of every type throughout the developed world as a means of valuing their customers and directing their marketing programs. It is not that difficult to compute if you have a database that contains transaction history, and if you can determine the revenue and costs involved in those transactions. The real value of lifetime value is not in the numbers, but in the use that is made of the numbers. Without imaginative strategies, the LTV numbers are useless.

Throughout this book, we will be using lifetime value tables to evaluate the effectiveness of various relationship-building strategies. Most lifetime value tables cover only 3 or 4 years. It is only in the automobile and similar industries (farm vehicles, trucks, capital equipment), where there is a lease or loan arrangement that typically lasts 3 or 4 years, that we have 9-year tables like those shown in this chapter.

Once lifetime value has been calculated, it can be put into every customer’s database record. The numbers will be different for each individual. Here is how they are calculated.

Owners of 2002 B Model cars will have one value. Owners of 2003 B Model cars will have another value. This reasoning applies to the owners of every Asian model in every year, and to every used Asian model. It applies to Asian lease and loan holders. People who own two or more cars have the values of these cars added together. The final result is a unique LTV number for every customer based on everything that customer owns.

None of this takes into account the demographics of the customers. Is the LTV of a 50-year-old customer different from the LTV of a 30-year-old customer? How would you discover that? The answer lies in the different repurchase rates for different age groups. Instead of looking at only the LTV by model, we can look at it both by model and by age group. We can look at the LTV of women versus the LTV of men. We can look at the LTV of people living in the Northeast versus those living on the West Coast. We can look at the differing LTVs of white, black, and Hispanic owners. Where do we stop in our calculations?

We have to consider two things:

1. Do the demographics result in significant differences?

2. If so, can we use these differences to craft different relationship-building strategies that will improve the repurchase rate?

The first question can be answered in short order by an analysis of the database going back 6 years. We can tell pretty rapidly whether older people have a higher repurchase rate than younger people, whether there are differences between whites and blacks, whether there are differences between East and West, and so on. In most cases, there will be only marginal differences, so we forget about those factors. You will have to look at the numbers to decide whether you can create a marketing program that recognizes and uses the differences.

For each possible program, we have to decide whether the cost of the program will be sufficiently low and the benefits sufficiently high for the resulting LTV to actually go up when the programs are put into place. In many cases there is little we can do. There are more possible database marketing programs that drive LTV down than there are that drive LTV up. That is part of the reason why CRM has been such a failure in so many industries.

Creating LTV Segments

Once we have LTV in everyone’s database record, we can begin to think about segmentation. We have already segmented customers by model and year, and we may also have segmented them by demographics. One important segmentation method is by total value to the company. We can arrange all customers by LTV from top to bottom and create segments that we call Gold, Silver, Bronze, Steel, and Lead. Our Gold customers are the most valuable ones. They have the highest lifetime value. They probably own more of the company’s products, and they have a higher spending rate and retention (or repurchase) rate. They are cheaper to serve, and they tend to buy higher-priced options. They are less price-sensitive than other segments. We really want to retain these Gold customers. We must devise programs to keep them coming back for more. We will add their segment designation to their database record so that we can recall them easily when we want to communicate with them.

Multiple Product Ownership

Most companies have more than one product that they can sell to their customers. Statistics from a variety of industries, such as banks, insurance, and travel, show that the retention rate is often a function of the number of products owned. The more of a bank’s products you own, for example, the less likely you are to drop the bank’s credit card when you receive a competing offer. Because of the importance of multiple product ownership, many companies have put major resources into selling their customers a second product—not only for the profit they make from the second product, but also because the second product shores up the retention rate for the first product. Since multiple product ownership is important, when you create a LTV table, you can segment customers by the number of products owned and create separate retention rates, and hence LTVs, for owners of multiple products. The chart that a bank developed is shown in Figure 4-1.

Figure 4-1: Gain in Retention Rate

Potential Lifetime Value

As stated earlier, lifetime value is a forward-looking concept. We don’t care about the past. What we want to know is, how much profit will we make from this customer in the future? The number of years is important, of course. If you knew that Arthur Hughes would be a customer for the next 40 years, whereas Bill Anderson would be a customer for only 5 years, you might treat Arthur Hughes and Bill Anderson differently, and perhaps treat Arthur Hughes better. But you cannot know that. Neither Arthur nor Bill knows it either. What you can know is that Arthur Hughes is a member of a segment that has a retention rate of 60 percent. You know his lifetime value after 3 or 4 years based on his current rate of spending, the number of products that he owns, and the retention rate of his segment.

There is something else that you can predict, however. You can look at all the additional products that you can sell to Arthur Hughes and estimate the increase in his lifetime value that would occur if you were to sell him each of several possible products. You can also predict the likelihood of his buying each of these products, based on the response rate of members of his segment when they are offered the products. You can draw up a table for each customer like that in Table 4-13.

Table 4-13: Next Best Product

Possible Product	Probability of Buying	3-Year LTV	Potential Value
A	45%	$ 400	$180
B	5%	$1200	$ 60
C	85%	$ 300	$255
D	35%	$ 200	$ 70
E	20%	$1000	$200

The most profitable product you could sell this customer would be product B, with a potential LTV of $1200. But the product he would be most likely to buy would be product C, with a potential LTV of only $300. Potential value is computed by multiplying the 3-year LTV by the probability of the customer’s buying each product. Putting both together, it looks as if the next best product for Arthur Hughes right now is product C. If you can sell him that, then go for product E, which has the next highest total value to your company. Banks do this analysis today. The next best product is put into every customer’s database record. Bank employees are asked to look at these records when they are communicating with customers so that they can discuss the next best product. This is using a lifetime value table to increase bank profits.

Prospect LTV

Thus far, we have discussed the lifetime value of customers only. We have not discussed prospects. Of course, prospects have a potential lifetime value. How to go about determining this is covered in Chapter 13, “Finding Loyal Customers.”

Using LTV to Sell Your Marketing Program

One of the most powerful uses of LTV is internal to the company. Every marketing organization in every corporation has to come up before a flinty-eyed chief financial officer every year to get a budget for its next year’s activities. Before LTV analysis, this was not a very easy sell. Advertising can point to the colorful TV ads that it produces. Sales can point to the number of customers that were signed up last year and the number that are projected to be signed up in the year to come. What can marketing point to?

Here is where LTV comes into its own. In the Asian automobile case, the database marketing manager can come to the CFO and say, “I can bring in $56 million more profits during the next 9 years if I can have a budget for a welcome kit and a series of strategic customer communications.” When asked to prove this statement, the marketer can trot out the projections shown in Table 4-7 and back them up with detailed tables like the ones shown in this chapter. Assuming that the presentation goes well, the CFO will retain these charts and refer to them in future budget sessions. She is likely to ask, “What was your actual repurchase rate? What were your actual marketing costs?”

The great thing about LTV is that these numbers are all actual projections that can be audited by PricewaterhouseCoopers or Deloitte & Touche. They are not smoke and mirrors, but projections of real events that will actually occur and can be verified. Database marketing, therefore, becomes a combination of science and art. The art, of course, is thinking up the strategies that will produce the improved customer relationships, and hence the improved repurchase rate. The science is putting all these projections together in an LTV chart that makes sense and proves the value of your efforts.