Customer Defection Rejection Part 1 : Attrition and Retention

This is a multi-part series, see: Part 1 – Part 2Part 3Part 4Part 5

Reducing costs is one way to get ROI from a Marketing and / or Service system; it’s also probably the simplest way.  But the more profitable, highest ROI way to get payout on these systems is to increase the value of customers by keeping them as active, engaged customers for a longer time.

Longer time?  That’s where the disagreements usually start.  People get bogged down in LifeTime Value and other arcane stuff and before you know it, once again, nothing will be done about customer retention this year.  

So let’s change that!

The first thing you have to do, and I mean have to do, is define a customer defection.  A lot of times this discussion gets bogged down in the minutia of LTV or inflexible attempts to “measure” defection.  This approach is not going to work very well until you have more data and experience working through these ideas by testing and proving the concepts.  Defection is often not an event, it’s a process.  If you’re just starting out, we need to use a simpler approach.

So the first thing I would put on the table is this: forget about triangulating “customer defection”, and talk about productivity.  I’m pretty sure you are familiar with all the discussions about marketing accountability.  Most C-level types get to where they are because they increase productivity – the profits generated per dollar spent – in one way or another.  If you want the marketing / analytics folks to have a seat at that table, you have to talk about marketing productivity.

To talk about marketing productivity, you need data.  So we’re going to create your first chart on customer productivity, which you can use to bring people to the table.  There are a couple of ways to do this; use the way easiest for the analysts or whichever way will be more likely understood by managers.  The attrition approach can be more difficult for analysts but is more easily understood by management because it is one data set over time.

Attrition Approach:  Go back 5 years (or 3, or 10, depending on how old your company is and what you have good data for) and find all customers who became new customers in that year.  Then find how many of those customers are still customers, defined by some kind of transaction this year.  If you think 2 years is a better measure of activity (highly seasonal businesses), use 2 years – new customer 5 years ago and the customer is active in the past 2 years.

What you will find is a major percentage of new customers from 5 years ago have not contacted you in the past 1 or 2 years – at least 50%, probably more like 60% – 80%.

Retention Approach:  Take your current active customers (some kind of a transaction this year or past 2 years) and bucket them by how many years they have been a customer; express these buckets as “Percent of Active Customers who were New customers X years ago”.  What you will find is the same idea as above, expressed in reverse: at least 50%, probably more like 60 – 80% of your active customers were New Customers in the past year or two.  The rest are spread out over the years back into time.

Either way, the conclusion is the same: the majority of revenue you are generating in the current time period is coming from customers who are newer customers.  “Old” customers generate very little in current revenue as a percentage of total current revenues.

For online only businesses, these numbers will skew even higher, into the 80% – 90% range.  For either approach above, the transaction defining “active” does not have to be a “purchase”, it could be a service interaction of some kind, or simply logging into a web site.  It should be an action that has the potential to generate revenue or decrease cost, because in the end, we are talking about increasing the  productivity of marketing dollars spent.

Now, given these numbers sitting right in front of you, you have to ask this question: does it make any sense at all to spend marketing budget equally against “all customers” when it is clear that customers become unresponsive over time when using this exact approach?  Wouldn’t it be more productive to reallocate marketing spend against the customers more likely to generate revenue – that is, customers with the more Recent transactions?

Going past revenue, if the average “old” customer generates little current revenue, what do per customer profits look like after you include the cost of marketing?  You don’t need to run LTV models to understand that you can’t possibly be making money with these customers.  The cost of marketing wipes out any profit you might have in them.  Right?

Now, don’t tell your boss these customers inactive for over a year or two have defected – that’s too controversial, even though you know it is true.  Tell the boss that in terms of bang for the buck, it makes sense to reallocate marketing spend from the non-responsive “customers” you have been burying with irrelevant marketing efforts towards the more responsive, active customers.

It’s really that simple.  This change will increase marketing productivity, and you can prove it.

If you measured back over 5 years and ended up at 20% of 5 year old customers still active in the current period, then your task is to increase this rolling 5 year retention rate to 21%, 22%, 23%, 24% and so on.  Or, if your business is short cycle, choose earlier periods, say 2 or 3 years, and approach it the same way.  The most productive time frame for retention marketing tends to be where you see abrupt drops in current activity.  If when looking at new customers back over 5 years, you find the percentage of currently active customers really drops off at 3 years ago, then use 3 years as the rolling benchmark.

Don’t over-think this or get lost in the minutia.  Use this simple, definitive data to track the performance of your efforts.  This must be done in order to measure how many customers you “save” with CRM or Retention Marketing, which leads to finding the value you created and calculating productivity.  

If it helps ease the internal political pain of “letting go” of these customers and declaring them “non-productive”, refer to these former customers as “new prospects”, for example:

“Hey Boss, look at the huge number of highly targeted new prospects we have – our non-productive “customers”.  We can still spend on them, but let’s change our marketing approach to them so it reflects their new status as prospects”.  Right?  That fixes the “if we tell the President 50% of our customers are not really customers they will reduce our budget” problem.  Shift the mindset, then shift the dollars; reallocate to active customers.

At the risk of repeating myself, this attrition or retention information is your required “base case” to get people to start thinking about marketing / customer productivity.  It provides the platform for you to call a meeting and get people interested in retention marketing.

The results of this analysis will drive a change in the way customer marketing is viewed, and making the appropriate change in tactics will increase the productivity of your marketing efforts right away.  I’ll detail some of those tactics for you in a couple of months – no hurry, you’ve got until Jan 2007 to get ready.

In the meantime, work on setting those “old customers” free by doing the attrition or defection reporting outlines above.  Heck, if you have the resources, break it out by quarters or even months – you know that will be the first question they ask when they see the annualized results – Hey, can I get this information broken out by quarter?

By doing this you will gain the most powerful knowledge an analyst or marketer can have – an understanding of the Customer LifeCycle.

Go to Part 2: What if the boss still doesn’t get it?  Simple follow-up analytics prove the point.


Get the book at

Find Out Specifically What is in the Book

Learn Customer Marketing Concepts and Metrics (site article list)

Download the first 9 chapters of the Drilling Down book: PDF