 |
Step
1:
|
Identify a cohort of customers to be analyzed. Generally
speaking, all customers in the cohort should have been
acquired within the same time period.
|
| |
Step
2: |
Obtain
data regarding each customer's retention rate per period.
The retention rate for a given period is the probability
that the customer will repurchase in that particular period
given the current product offering.
|
| |
Step
3: |
Obtain
data for sales, gross margin, and marketing and customer
service costs.
|
| |
Step
4: |
Determine
how many periods into the future retention equity will
be computed.
|
| |
Step
5: |
Project
future retention rates for the customer using extrapolation.
|
| |
Step
6: |
Project
future sales, gross margin, and marketing and customer
services costs.
|
| |
Step
7: |
Obtain
from financial sources the discount rate to use in the
analysis.
|
| |
Step
8: |
Use
historical and projected retention rates to estimate the
number of customers who survive each period.
|
| |
Step
9: |
Compute
the profits for each period using the projected number
of customers and the relevant sales and cost data.
|
| |
Step
10: |
Discount
all future profits using the discount rate arrived at
in Step 7.
|
| |
Step
11: |
Sum
all discounted profits for past and projected periods.
This is the retention equity for the entire cohort.
|
| |
Step
12: |
Divide
the sum in Step 11 by the number of customers who began
in the cohort to find the retention equity of each customer.
|
| The
example of Firm B, illustrates these steps. |
|
Step
1:
|
Identify
a cohort of customers to be analyzed. Generally speaking,
all customers in the cohort should have been acquired
within the same time period. Firm B selected a cohort
of customers who were acquired through the same source
(that is, by the same acquisition method, such as direct
sales in combination with direct mail). There were 1,000
customers acquired in 1994. The company had five years
of historical data and the time period of analysis was
per calendar year.
|
|
Step
2: |
Obtain
data regarding each customer's retention rate (survival
rate) per period. Table
Retention-6 shows the historical retention rates for
seven years: 1995 through 2001.
|
|
Step
3: |
Obtain
data for sales, gross margin, and marketing and servicing
costs per period. Table
Retention-6 contains historical data on sales, marketing
and servicing costs, and gross margin percentages.
|
|
Step
4: |
Determine
how many periods into the future retention equity will
be computed. Firm B's management decided to project customer
data 10 years into the future. They felt that beyond that
point too many factors could change, rendering the numbers
arbitrary.
|
|
Step
5: |
Project
future retention rates for the customer using extrapolation.
A mathematical model is used to project the retention
rate into the future. Using a specific probability model
we find that the retention rate per period equals k(1e-bt),
where k and b are the model's parameters, and t is the
number of periods since the cohort group was acquired.
k is a parameter that reflects the maximum retention
level the firm would attain if it spent an infinite
amount on customer retention (i.e., the maximum retention
potential for the firm). The b parameter moderates the
impact that the amount of time elapsed since acquisition
has upon the probability of retention. e (the exponential,
or exp) is a constant value of 2.718282.
To see how this specific retention model works, suppose
the firm wants to calculate the retention rate 10 periods
after the customers began. The general form of the model
is k(1e-bt).
Based on managers' responses to questions, a decision
calculus model determines that b equals 0.6 and k equals
0.9. The actual value of the retention rate 10 periods
after acquisition is .9*(1e-bt)
= .8977. Figure
Retention-A shows the retention-rate pattern used
in this example.
|
|
Step
6: |
Project
future sales, gross margin, and marketing and customer
services costs per period. A simple extrapolation method
was used to project revenues and costs. A curve was fitted
to the historical numbers for each of these three types
of information, and projections were made from these curves.
In making the projections, Firm B took into account that
after customers mature, their future growth in terms of
revenue and profits is relatively low. Generally speaking,
if a firm develops new products to sell to mature customers,
growth rates can increase. In this scenario, however,
the firm had no such expectations.
|
|
Step
7: |
Obtain
from financial sources the discount rate to use in the
analysis. Firm B's finance department suggested a discount
rate of 20 percent, based on the firm's historical 12
percent cost of capital and on the high levels of uncertainty
associated with future revenue streams from customers.
|
|
Step
8: |
Use
historical and projected retention rates to estimate the
number of customers who survive each period. Using the
retention rates generated by the mathematical model described
in Step 5, management determined the survival rate. The
survival rate for any period is determined by multiplying
the survival rate from the previous period (i.e., the
probability that a customer survived the period just prior
to the current period) by the retention rate for the current
period. (For the first year analyzed-in this case 1995-the
retention rate and the survival rate will always be the
same.) The survival rate column in Table
Retention-7 indicates the probability that a customer
Firm B acquired in 1994 will remain a customer during
the given period. For example, according to the table,
7.9 percent of the customers Firm B acquired in 1994 will
remain customers in 2005. To determine the number of customers
in any time period, multiply the period survival rate
by the number of customers initially acquired.
|
|
Step
9: |
Compute
the profits for each period using the projected number
of customers and the relevant sales and cost data. In
Table Retention-7,
multiplying the sales per customer by the gross sales
margin, and then subtracting from that the marketing and
servicing costs, results in projected profit per customer.
Multiplying profit per customer by the number of customers
gives the total profit per period.
|
|
Step
10: |
Discount
all future profits using the discount rate arrived at
in Step 7. Once profit per period is determined, profits
are discounted by multiplying them by [1/(1+d)]t-1,
where d is the discount rate, t is the current period
and i is the initial period. The results appear in Table
Retention-7 in the discounted profit column.
|
|
Step
11: |
Sum
all discounted profits for past and projected periods.
Using the discounted profits per period arrived at in
Step10, Firm B's management summed the discounted profits
for all historical and future periods to obtain total
retention equity of all customers. This came to $635,033,
as shown in Table
Retention-7.
|
|
Step
12: |
Divide
the sum in Step 11 by the number of customers who began
in the cohort to find the retention equity per customer.
There were 1,000 customers who began in the cohort, and
so the retention equity of a customer comes to $635,033/1,000
= $635.
|
____________________________________________
* Robert C. Blattberg and John Deighton, "Manage Marketing
by the Customer Equity Test," Harvard Business Review,
July-August (1996): 136-144. |
