Marketers use several terms-ROI, profitability, lifetime value and RFM-as if they are synonymous. They aren’t. Each has its own special use in database marketing.
Return on Investment Return on investment (ROI) is generally used for marketing campaigns. You invest a certain amount in a specific effort to sell a group of products or services, and make a certain net profit from the activity. The ROI is determined by this basic formula:
ROI = (net profit – amount invested) / (amount invested)
Suppose you invest $40,000 in mailing a catalog to 100,000 people. You get a 2% response rate. Two thousand people spend an average of $100 each, paying a total of $200,000. If the average cost of each order (including the cost of goods sold, fulfillment, telesales, credit, returns, etc.) is $40, then your net profit on an average order is $60, multiplied by 2,000, or $120,000 in all. Your ROI is:
ROI = ($120,000 – $40,000) / $40,000 = 2 = 200%
Unfortunately, most customer-acquisition programs are not that successful; the profit for most DMers comes from subsequent sales to buyers acquired in the first sale. A bank may spend $80 to acquire a credit card customer whose lifetime value at that moment is only about $25. As soon as the customer begins to use her card, however, she may build up a big balance, pay interest charges and an annual fee and increase her lifetime value by multiple amounts.
There are really two ways of computing ROI:
* ROI from the initial acquisition, which can be negative.
* ROI from the lifetime value of the customer, which is usually positive.
Profitability Profitability applies to customers rather than campaigns. For example, instead of relying on traditional measures like deposit and loan balances, leading-edge banks are using profitability to help them understand the true value of their customers. To get a fix on this, they must first calculate product profitability at the individual customer level. Once this has been determined, usually on a monthly or quarterly basis, then profitability is simply the sum of the product values for each customer or customer household. This measurement is useful when comparing the value (or profitability) of customers to each other. It is historical in nature, a snapshot of the value of the customer at some past period of time-last month, last quarter or last year.
Suppose a customer has only a retail loan with Bank A. The original loan may have been for $15,500. As of this month, the balance owed is $9,806.66. Banks borrow money wholesale, and lend it retail. To calculate profitability, the bank’s software will determine the interest the customer paid last month, less the wholesale cost the bank had to pay on the outstanding balance, less the allowance for possible loss and the bank’s overhead costs. The net result, in this case, was a positive $13.76. That was the profitability to the bank of that automobile loan last month.
If this was the only product owned by this customer, her profitability last month was $13.76. If she owned more bank products, her total profitability would be the sum of the profitability of each separate account. With the computing power available today, companies can measure the profitability of every customer every month based on such variables as the customer’s current product ownership, balances, rates, transaction behavior and channel usage.
When properly calculated, customer profitability is an accurate measure of the relative value of customers’ relationships to the company. It allows you to determine how additional product sales may add to or detract from profitability, and to identify essential customers-and those whose departure would be welcomed.
Here’s how one bank segments its customers based on their profitability. The first group is the gold customers, who accounted for 69% of the bank’s profits last month. They represent only 6% of the households. The second group, 14% of the households, provided 31% of the bank’s profits. The lowest group-a quarter of the bank’s customers-cost the bank 20% of its profits. They used up more resources than they paid for. Knowing this, the bank has to re-price its products or find some other way to increase profitability.
Profitability is measured by looking at results from the most recent month, quarter or year. Since we are focusing on the performance of individual customers, often missing from profitability calculations are such factors as retention rate, referral rate, marketing expenses, acquisition cost, and the effect on behavior of various relationship-building activities. For these, we must look to lifetime value.
Lifetime Value Lifetime value (LTV) is computed initially for groups, or segments, of customers. Once determined, it can be attributed to individual members of the group. The graphic above (“Are These Customers Worth Keeping?”) is an example of a lifetime value table, calculated for one segment of 50,000 bank customers who have credit cards. Though hardly the best customers, they are worth retaining and cultivating.
Lifetime value calculations add several factors to a customer’s profitability. Take the 50,000 customers (e) who were acquired at an average cost of $80 each (j), or $4 million. These folks are not static-they tend to drift away. A year later, only 37,500 of them (d) are still using their card. Their retention rate is 75% (c). Of those who remain, the retention rate increases to 80% in Year 2 and 85% in Year 3.
Some of these customers are encouraged by marketing efforts of $25 per customer per year (k) to increase their spending, to stay with the card, and to become advocates and get their relatives and friends to take the card. This results in a referral rate (a) and 3,000 referred customers in Year 2 (b). The marketing efforts also can encourage increased retention and spending rates, although these are not shown on the chart.
The revenue per customer (g) in the year of acquisition is the same as in the profitability example ($389.76). In the following two years, it goes up by $40 because the annual fee kicks in. The direct costs (i) are the same ($256), but in addition we have the acquisition cost of $80 (j) and the annual marketing costs of $25 (k).
First, some chronological definitions. Year 1 is not a calendar year; it is the “year of acquisition.” Year 2 is the year after acquisition and Year 3 is the year after that. Year 1 may contain people who have been acquired in various years (such as 1995, 1996 and 1997). Year 2 would represent the performance of these people in their second year (1996, 1997 and 1998).
The profit (m) is the total revenue minus the total costs. Particular attention should be paid to the discount rate (n). Discounting is required because we will be adding profit generated in different years. Money to be made in the future is not as valuable as money in hand. If we add the amounts from various years together, we must discount future monies to give them values comparable to current funds. The formula for the discount rate is:
D = (1 + i )n
(i = the market rate of interest times a factor for risk; n = the number of years for which you have to wait.)
If the market rate of interest is 7% and the risk factor is 2, then the discount rate in the third year is:
D = (1 + .14)2
D = 1.30
Dividing the net profit in each year (m) by the discount rate (n) results in the net present value (NPV) profit (o). The NPV profit in each year is added to that in the previous year to get the cumulative net present value profit (p). Dividing this figure in each year by the original 50,000 customers gives you the lifetime value (q) of these acquired customers.
Instead of being a monthly profitability figure like $13.76, lifetime value is a set of annual figures that changes each year. Some of these numbers are under the bank’s control, while others depend on external factors such as the market, interest rates, competitive offers, marketing effectiveness, etc. For example, the retention rate (c) reflects the fact that some customers are always leaving. They may die, or defect to another credit card. They may move away or stop using their card. Some of these actions can be influenced by the way the bank treats its customers. This depends on the way the marketing budget (k) is used. The bank can also use its best efforts to increase the average balance (f) or the spending rate, which determines the revenue per customer (g).
>From the table on page 65, we have determined that the average lifetime value of the 50,000 customers in this segment is only $28.76 in the first 12 months, rising to $216.02 in the third year. If we want to convert these numbers into the lifetime value of an individual, we can do this by attributing the expected behavior of an individual to the average of the group to which that person belongs.
In evaluating a marketing strategy, companies typically compute the base lifetime value of a customer segment. Then they estimate the effect of a proposed marketing strategy by creating a second lifetime value table that shows the changes in the retention rate, referral rate, spending rate, plus the increased costs of the marketing strategy. These four factors will produce a revised lifetime value for the segment. It will be either greater or lower than the base lifetime value. If it is lower, then the proposed strategy is unlikely to be successful and should be revised or scrapped. If it is higher, it shows that the proposed strategy has a possibility of success. Whether it will be successful depends on the execution-which could, of course, be flawed, resulting in a reduction of lifetime value.
The fact that lifetime value varies with the number of years that the customer remains with the supplier raises the question: How many years out should you calculate this value? The lifetime value table shows three years. In rapid-action catalog sales, even that may be too long a period. Some catalogers use quarters instead of years. Banks and insurance companies may compute lifetime value over five or even 10 years, because their retention rates are often quite high.
RFM The recency, frequency and monetary (RFM) approach is the only one of the four methods that accurately predicts future customer behavior. It works because past behavior is usually a good indication of future behavior. Of course, in any direct marketing program, the vast majority of customers-usually more than 95%-will not respond to a promotion. But RFM predicts with surprising precision who the other 5% are likely to be, recency being the most powerful variable.
Recent buyers respond much better than customers whose last purchase was in the distant past. Frequent buyers also respond better than less frequent buyers, and high spenders are better than those who spend little. Using these principles, marketers can group customers into as many as 125 different RFM “cells” based on their recency of purchase, frequency of purchase and total spending. Once this is done, their response pattern usually looks something like what’s shown in the chart above (“To Mail, or Not to Mail…”).
In this chart, customers are divided into 125 cells based on behavior. Their response to a promotion is shown. Fifty-three of the cells lost money when promoted. The remaining 72 cells were profitable. Knowing this behavior permits marketers to avoid mailing to cells that are unlikely to make a purchase.
Response rates can be increased over the results of mailing to the entire customer base. Besides responsiveness, RFM also often accurately predicts average order size. Successful cells, therefore, are profitable not only because they are more likely to buy, but because they are likely to buy more per order.
RFM can be used as a rough guide to profitability or lifetime value since frequency and monetary spending tend to be closely connected to profitability and lifetime value. RFM, however, is no substitute for the accuracy that accompanies these measures. RFM also helps forecast a campaign’s return on investment.
Which is Best? So, which of these approaches is the best measure of marketing success? Actually, they all have their uses.
ROI is a vital measurement of the success of any DM campaign. When we compare several campaigns, we can look at the response rate and the return on investment of each one. ROI is usually a more valid measure of the success of a campaign than the response rate, because it takes into account the money spent to acquire the responses. Every DM campaign should be measured by ROI.
Profitability is an excellent way to segment customers. It can be used to determine the worth of a particular customer at the present time. Lifetime value, meanwhile, is a good means to evaluate the success of proposed marketing strategies.
Which to use? I would go with all four. I would choose RFM to predict the success of all customer promotions, which I’d measure by ROI. I would use profitability to segment my customer base for the purpose of allocating marketing and service budgets. Finally, I would use lifetime value to evaluate the profitability of each marketing strategy before any serious money is spent.
