How To Measure Future Customer Value and Manage it with E-mail

First published:  Business Intelligence Toolbox,  2/12/01
"How To Measure Future Customer Value and Manage it with High ROI E-mail"  

The following article is from the advanced topics section; you might want to take the tutorial Comparing the Potential Value of Customer Groups before reading it.  If you would rather see a general description of the Drilling Down method and specific benefits first, go to the home page.

Folks, here's an honest to goodness, classic customer retention marketing program, suitable for automation, complete with metrics and testing methodology.  This is the first program I developed for interactive customer retention at Home Shopping Network, averaging a 135% net ROI at 60 days, month after month.  

This is the exactly the kind of work you can do "pre-CRM" to determine whether CRM techniques will increase the value of your customers and by how much.  Going through the following process will also identify any special needs you may have to consider when choosing a CRM package.

It's an example of the kind of information found in my book, similar in format, using examples and a step by step approach.  This version is not as detailed as typical explanations in the book (attention spans are short on the  Web).

Customer Retention
and Valuation Concepts

Have you ever heard somebody refer to his or her customer list as a "file"? If you have, you were probably listening to someone who has been around the catalog block a few times.   Before computers (huh?), catalog companies used to keep all their customer information on 
3 x 5 cards.

They’d rifle through this deck of cards to select customers for each mailing, and when a customer placed an order, they would write it on the customer’s card.  These file cards as a group became known as "the customer file," and even after everything became computerized, the name stuck.

Who cares? It happens that while going through these cards by hand, and writing down orders, the catalog folks began to see patterns emerge.  There was an exchange taking place, and the data was speaking.  What the data said to them, and what they heard, were 3 things:

1.  Customers who purchased recently were more likely to buy again versus customers who had not purchased in a while

2.  Customers who purchased frequently were more likely to buy again versus customers who had made just one or two purchases

3.  Customers who had spent the most money in total were more likely to buy again.  The most valuable customers tended to continue to become even more valuable.

So the catalog folks tested this concept, the idea past purchase behavior could predict future results.  First, they ranked all their customers on these 3 attributes, sorting their customer records so that customers who had bought most Recently, most Frequently, and had spent the most Money were at the top.  These customers were labeled "best."   Customers who had not purchased for a while, had made few purchases, and had spent little money were at the bottom of the list, labeled "worst."

Then they mailed their catalogs to all the customers, just like they usually do, and tracked how the group of people who ranked highest in the 3 categories above (best) responded to their mailings, and compared this response to the group of people who ranked lowest (worst).  They found a huge difference in response and sales between best and worst customers.  Repeating this test over and over, they found it worked every time!

The group who ranked "best" in the 3 categories above always had higher response rates than the group who ranked "worst."  It worked so well they cut back on mailing to people who ranked worst, and spent the money saved on mailing more often to the group who ranked best.  And their sales exploded, while their costs remained the same or went down.  They were increasing their marketing efficiency and effectiveness by targeting to the most responsive, highest value customers.

The Recency, Frequency, Monetary value (RFM) model works everywhere, in virtually every high activity business.  And it works for just about any kind of "action-oriented" behavior you are trying to get a customer to repeat, whether it’s purchases, visits, sign-ups, surveys, games or anything else.  I’m going to use purchases and visits as examples.

A customer who has visited your site Recently (R) and Frequently (F) and created a lot of Monetary Value (M) through purchases is much more likely to visit and buy again.  And, a high Recency / Frequency / Monetary Value (RFM) customer who stops visiting is a customer who is finding alternatives to your site.  It makes sense, doesn’t it? 

Customers who have not visited or purchased in a while are less interested in you than customers who have done one of these things recently.  Put Recency, Frequency, and Monetary Value together and you have a pretty good indicator of interest in your site at the customer level.  This is valuable information for a business to have.

Assuming the behavior being ranked (purchase, visit) using RFM has economic value, the higher the RFM score, the more profitable the customer is to the business now and in the future.  High RFM customers are most likely to continue to purchase and visit, AND they are most likely to respond to marketing promotions.  The  opposite is true for low RFM customers; they are the least likely to purchase or visit again AND the least likely to respond to marketing promotions they receive.

For these reasons, RFM is closely related to another customer direct marketing concept: LifeTime Value (LTV).  LTV is the expected net profit a customer will contribute to your business as long as the customer remains a customer.  Because of the linkage to LTV, RFM techniques can be used as a proxy for the future profitability of a business.

High RFM customers represent future business potential, because the customers are willing and interested in doing business with you, and have high LTV.  Low RFM customers represent dwindling business opportunity, low LTV, and are a flag something needs to be done with those customers to increase their value.

RFM scoring of individual customers is a catalog and TV shopping technique used to select which customers can most profitably be promoted to.  There is a more simplistic application of RFM web sites can use to easily track the quality of overall customer retention, without going through the effort of RFM scoring individual customers.  This tracking can be used to measure customer retention and trigger  profitable customer retention promotions.  The basic technique creates a platform for learning the key customer behavior metrics needed to manage customer retention, and provides a foundation for building a more comprehensive effort down the road.

Measurement:  How do I know when I have a customer retention problem?

Here's a real life story I have seen repeated over and over.  Many companies judge their best customers by looking at just Frequency of activity, either purchases or page views.  They set a threshold, like 25 purchases or 100 page views, and then count the number of customers who have achieved this goal.  As long as this number of customers keeps growing, they think the business is on track and doing fine.

Then someone with experience in database marketing does an analysis, and the company finds out that 60% of these "best customers" haven't purchased or visited in over 12 months!  So they desperately try to e-mail these people offers and get them to come back, but get truly lousy response rates.  The customer relationship is already over, and the company has lost a ton of their best customers because they have no formalized, proactive customer retention program.  These defected best customers are a  troubling sign for the future value of the business and more or likely to follow.

Customer tracking by Frequency is a rear-view mirror, because it doesn't take into account the future potential of a customer to contribute to revenues.  You have to track customers by Recency to predict future value, because Recency is the strongest indicator of future customer activity.

If you have been tracking the loyalty of your customers as a group using the Drilling Down visual method, the equivalent of the above scenario would be seeing the Frequency Hurdle Rates rising while the Recency Hurdle Rates are falling, a classic sign of failing customer retention.

You can act to slow or prevent some of this customer attrition by implementing a basic customer retention measurement and management program using e-mail.  The following program uses customer Recency to categorize customers and create a framework for profitability measurement and automation of the retention program.

1.  Choose the activity you wish to measure and manage the future value of: purchases, page views, downloads, click-throughs, whatever "action metric" is important to you and the business model.

2.  Choose a time metric to define customer Recency in this activity.  Blocks of 30 days are pretty standard and also tie in with other monthly reporting and operational  cycles when databases might be updated.  Over-achievers with significant database horsepower might use weekly data, especially if you are looking at visits and you are a time-driven site, for example, a site focused on news.  You want the freshest data you can have to use these techniques; fresher data = better results.

3.  Identify customers who have engaged in the activity you are measuring and determine the date these customers most Recently engaged in the activity.  If you are using 30 day blocks of time, you would identify customers last engaging in the activity in the past 30 days, in the past 31 - 60 days, in the past 61 - 90 days, and so forth.  Go out at least 6 months in 30 day blocks.  After 6 months, you can have a count for "everybody else" who has not engaged in the activity for over 6 months.  

4.  Print your report.  This is a report of the number of customers who last (most Recently) engaged in the specific activity a certain number of days ago.  A customer is represented only once in any of the 30 day blocks below; remember, we are looking only at the most Recent date the customer engaged in the activity.  An example using purchase Recency could look like this:

Table 1 - Customer Recency of Purchase

Read: "5,786 customers last purchased within the past 30 days, 4,356 customers last purchased 31 - 60 days ago, 3,872 customers purchased 61 - 90 days ago..."

or it might look like this:

Table 2 - Customer Recency of Purchase

Read: "1,198 customers last purchased within the past 30 days, 2,577 customers last purchased 31 - 60 days ago, 3,872 customers purchased 61 - 90 days ago..."

Which of the two tables above represents the business with the most future potential?  Which table represents the business where the most customers are likely to continue engaging in the activity being profiled?

If you guessed Table 1, you're right.  Both these tables represent businesses with a total of 24,141 customers, but there are many more Recent customers in the Table 1 business then there are in the Table 2 business.  Since the more Recent a customer is, the more likely they are to repeat an activity, the business in Table 1 can expect more business out of their current customers in the future than the business in Table 2.  The business is Table 1 has much better customer retention, and the customers on average have higher future value.  Real world visual examples of visitor Recency comparable to the ones above can be found here.

OK, now what?  Well, if you do this exercise every month, you can compare trends in the 30 day customer Recency blocks and watch the customer Lifecycle play out before your eyes.  In a healthy business, the number of customers in the most Recent block should grow faster than the numbers in the other blocks.  If the number of customers in the most Recent block is shrinking while the numbers in the other blocks are rising, you've got a customer retention (future value) problem, and need to take action.  Note: You don't want to see growth in the 180+ block at all, but it's inevitable, and the longer you are in business, the larger this number will grow.  You should be most concerned with managing (reducing) growth in the blocks from 60 days to 180 days, where you can still take effective, profitable action.

Management: How do I do something about customer retention problems?

Customer Retention management involves trying to drive as many customers as you can into the most recent customer block as profitably as possible.  

Think of it this way.  You want customers to remain active and Recent with you so they are generating revenues.  Some customers will do this without any special marketing attention from you.  Others will need an incentive.  High ROI customer retention programs focus on only the customers who need an incentive.  By approaching retention this way, you avoid spending precious marketing dollars where they are not needed and can increase them where they are needed.  In other words, if you allocate the budget away from customers with a low likelihood to defect (the most recent customers), you can put more money per customer to work against customers who are more likely to defect (more distant customers).

Generally, the sweet spot for customer retention activity, the point at which spending money retaining a customer generates the highest ROI, is somewhere in the middle of our chart above.  Spending money on very Recent customers or very distant customers is not usually profitable.  How do you find the sweet spot for your business?

Check out Table 3 below:

Table 3 - Response Rate by Customer Recency

If you e-mail the exact same offer (say, $3 off anything on the site, or a free download, or promotion of new content) to all the customers in this table, you will get a response rate grid similar to the one above.  What's important to understand here is not the actual numbers, because they will vary depending on the offer and media used.  What is important to understand is the relative differences in the response rates.  You can expect the most Recent customers to have a very high response rate, and the response rate to drop sharply as customers get less Recent.  The most Recent customers will generally be 8 to 40 times more responsive than the least Recent customers.

So how do you turn all this into something you can use?  You create a customer retention test, measure the results, and turn the test into a monthly customer retention e-mail promotion.  It's a bit complex to set up the first time test, but once you complete the test promotion, you will either have your systems set up to measure the results every time, or you can just run a test periodically to make sure your results are still on track.  This process can be totally automated.

 Here's what you do:

1.  Select an equal percentage of customers from each of the 30 day blocks on our Recency grid above.  A good number for a test like this is a 10% random sample of the customers in each block.  It's very important the sample is truly random.  This sample will receive your promotional e-mail; these customers are called the test group.

2.  Make sure you can identify every customer (not just the ones selected for the promotion) by their Recency block before you do the test.  Either tag their record somehow or make sure you can determine when they last engaged in the activity being promoted before the date the test e-mail is sent.  Customers not receiving the promotion (90% if you used 10% for the test group) are called the control group.

3.  E-mail the exact same offer to each 10% of the block group as a single promotion, making sure all other potential variables are equal.  For example, don't send the e-mail to different Recency blocks on different days of the week.  Tabulate response by Recency block, including total sales and cost of goods sold.  If you can't get the actual cost of goods sold, use the average for your business.

Pure content businesses would look to potential ad sales generated for the "sales" metric in the following formulas.  If you perceive you have "no costs" to doing a promotion, there is no need to do the following ROI analysis.  May I humbly submit you might consider offering something of value (cost to you) if you're serious about getting defected customers to return to your site, for example, great new content you have to pay someone to write.  The beauty of this method is, after the test,  you are only making an offer to those customers you are likely to lose and most likely to get back, so you can afford to spend more per customer on any promotion efforts.

4. Use the following formula to look for your sweet spot.  You want to do this calculation for each 30 day Recency block in the promotion

Start:  Sales Generated by Test Group
minus: Cost of Goods Sold to Test Group       minus: Cost of E-mail Promotion to Test
           Group 
minus: Cost of Discount (or other incentive,
           or special content) to Test Group
Equals: Promotion Profit by 30 day customer
            Recency Block
Divide by: Number of customers in Test Group
Equals: Promotion Profit per Customer by
            Recency Block

5.  Now, calculate the profit per customer in each Recency block during the time period of the promotion who did not receive the promotion.  Note: this could be done using a 10% random sample of these people, if you wish.  Given a choice, I'd use the whole group.

Start: Sales Generated by Customers NOT
          in Promotion (Control Group)
minus: Cost of Product Sold
Equals: Profit by 30 day customer Recency
            Block
Divide by:  The number of customers in Control Group by Recency Block
Equals: Profit per Customer by Recency Block

6.  Compare the profit per customer by Recency block of customers in the Test Group versus customers in the Control Group.  Subtract the profit per customer in the Control Group from the profit per customer in the Test Group.  This difference represents the profit due to your promotional efforts, the profit existing because you spent money on one group of customers versus the other group of customers where you spent no money and these customers did not receive any promotion.

It should look something like this:

Table 4 - Profit per Customer by Recency Block

The most profitable 30 day block (61 - 90 days for this example) in the promotion is your sweet spot.  This is where you should focus this customer retention promotion.  Each month, select the customers who have "rolled over" into this block and e-mail your promotional offer to them. 

For example, if the 61 - 90 day Recency block is the most profitable for you, each month, select all customers who have not engaged in the activity you profiled (purchases, page views, downloads, click-throughs, whatever "action metric" you are tracking) for 61 - 90 days, and send them your promotion.  

People who respond and engage in the activity become "30 day customers" again, and are now much more likely to continue in the behavior after the promotion.  Using this promotion over time, you will begin to shift your whole customer base to a more recent status.  Or said another way, your business will start to look more like the one in Table 1 rather than the business in Table 2.  You will be increasing the future value of your customers and making money at the same time!

Once you do the work of the initial test, this simple retention promotion becomes a very easy to execute, regular program that can serve as the base for a building a customer retention effort.  Further testing of offers, discount levels, and so on can be used to optimize the profitability of the promotion.  From there, automation of this promotion through the CRM engine or other internal processes creates a highly efficient and effective "lights out" promotion which automatically minds some of your customer retention problems for you.

Note how linear the profits are by Recency in the Control Group.  This is why Recency in the customer base is so important; the most Recent customers are generally the most profitable customers.  If you are wondering why the most Recent customers were the least profitable when promoted to, you need to understand the concept of subsidy cost.  Subsidy cost is the primary reason why the net profit difference between the test and control groups (Test - Control in the table above) is similar to a Bell Curve concept, rising then rolling over and falling.  Subsidy cost can be measured and best (usually most Recent) customer programs designed to avoid subsidy costs.  There is an entire chapter devoted to this subject, with mathematical models describing it, in the Drilling Down book.