New RFM: in Academia & Wireless
# 45: 5/2004
Drilling Down - Turning Customer
Data into Profits with a Spreadsheet
Customer Valuation, Retention, Loyalty, Defection
Get the Drilling Down Book!
In This Issue:
# Topics Overview
# Best Customer Retention Articles
# Question - New RFM: Student Retention?
# Question - New RFM: Wireless Retention?
# New RFM Metrics: eMetrics Summit
Hi again folks, Jim Novo here.
You've probably heard the hot new buzz phrase "behavioral
targeting" thrown around a lot in the online display advertising
space lately. Are you happy folks are so excited about joining
the party we've been throwing right here for the past 4 years
now? I'll have more to say on this topic in the future,
including why I think a lot of what is being done is **not**
behavioral targeting, but I'll leave that off the table for
Instead, we've got two great article links, one an example of what
**is** classic behavioral targeting in online display advertising, as
behavior trumps demographics yet again. The second great article
delves a bit into why people have a hard time turning data mining
results into actionable programs. Plus, two great questions from fellow Drillers: a rare question on
student retention programs (a personal favorite of mine because they
are so darn challenging) and the most common question I get -
retention in wireless services. Rare or common, we try to take 'em when they have to do with
customer retention programs.
So let's do some Drillin'!
Best Customer Retention Articles
Note to web
site visitors: These links may have expired by the time you read
can get these "must read" links e-mailed to
every 2 weeks before they expire by subscribing to the newsletter.
Woman Wants Shopping Online
April 16, 2004 Internet Retailer
The title is a bit deceptive, but the results remarkable. People who
clicked on women's clothes were more likely to buy them than people who
identified themselves as women. When you are looking to generate behavior,
use behavior to target; it beats demographics
every single time. After all, do you really care who the buyer is,
as long as they buy?
Beyond the Black Box
May 10, 2004 Target Marketing
This article addresses a very common issue. People work very hard on
capturing data and turning it into information, but then they don't know how to act on
it. This is particularly true with data mining, when you might not be sure
why the segmentation is relevant. Think about what your models means in
terms of the customer and design actionable programs that make
use of this knowledge.
If you are a consultant, agency, or software developer with clients
needing action-oriented customer intelligence or High ROI Customer
Marketing program designs, click
Questions from Fellow Drillers
If you don't know what RFM is or how it can be used to drive customer profitability in just about any business,
New RFM: Predicting Student Churn
Q: It is clear that retention of students is a complex
issue. The students' satisfaction with the university will be
partly determined by their experience during their first semester with
the university. I have identified that each service encounter
will contribute to the overall impression that the student has of the
university. Some encounters are 'moments of truth' and will have
a major impact on the student's perceptions of the university.
You just got very lucky, I happen to have first-hand experience on
this topic, which is very rare, as not many educational institutions
are thinking this way. You should be congratulated for making
this connection, though it will be a difficult battle dealing with the
university administration on making changes, in my experience...
Q: I would much appreciate if you could advice
me on the retention strategy and what approach the university should
take to retention. Also, any ideas on management of moments of
truth, particularly what enhances and detracts from the customers'
encounters with the university.
A: Please consider this old business maxim: Only attempt
to control what you can actually control; otherwise you will end up
not having an affect on anything. It very well may be that the various "touchpoints" exist
and can be defined, but can you reasonably control any of them?
Which ones, and how will you control them? This is where you
need to focus your efforts.
In my experience, a university is
not the kind of place where you can undertake a "customer service
education program" with employees and expect compliance at all
the touchpoints. So what you have to do is pick the major points
of influence where you know you can exert some control and seek to
prove your case with facts and testing.
The best way to do this is with (no surprise here) data. What
you need to do is take a population, e.g. "first semester
students in January 2003" and then segment them into two groups -
those who continued and those who did not. Then compare - what
is it about the "not enrolled" group that differs from the
"still enrolled" group?
I can tell you from experience some likely differences between the
segments. Please be aware these are based on US experience and
may not apply in your culture, but they provide good examples:
1. The courses taken in first semester - this is very highly
predictive of continuing enrollment.
2. Within a single course, the professors teaching it.
For example, you may find students taking Calculus are very likely to
drop out after first semester, and will think "Calculus" is
the problem. But when you look at Calculus by Professor, you
find the drop out rate is concentrated around one or two professors,
not the subject.
3. Financial aid amounts and qualifications
4. Involvement in non-academic pursuits - clubs, sports,
student government, etc.
In the US, this last segment is also highly predictive of future
donations to the institution, which is an important funding concern -
in the language of CRM, the "Lifetime Value" of the student
is positively correlated to participation in non-academic university
Are any of the above "controllable"? In the US, if
you value your job, you don't even bring up any of the results of #2.
This is a losing battle, the politics are thick and the issues not
always clear. #3 is likewise surrounded (in the US) with a lot
of uncontrollable and often legal and marketing issues.
But on #1 and #4, you may be able to actually "do
something". You can change marketing materials and
"first semester required courses" to promote specific
courses which tend to generate long-term students.
are enrolled, you can boost the promotion of non-academic pursuits by
having educational and recruitment events. A further refinement
of this strategy would be to single out students enrolled in courses
showing a high "first semester defection rate" and put extra
effort into getting them into non-academic pursuits.
The amount of effort required to address #2 & #3 is enormous
because there are huge institutional barriers to changes in these
areas. Addressing #1 & #4 can be a more "under the
radar" effort and not nearly as costly, either personally or
Once you know the most critical "touchpoints"
that you also **can have some control over**, make your changes and
measure your results. When you can prove first semester
retention rate increases, you will get a seat at the table for going
further and asking for control of other important touchpoints.
If you are in SEO and the client isn't converting the additional
visitors you generate, you can help them make it happen - click here.
New RFM: Wireless Segmentation
Q: I have taken up a new assignment in this new
financial year in my company in India (Cellular / Mobile
connections). I would like some direction from you; also I have
book to my management.
A: Well, thanks for the plug on the book and I'll give
the "direction" a try...
Q: Objective: To create loyal customers who become
1. To drive customer loyalty to ensure 80% of the customers recommend
brand to others.
2. Customer Behaviour Profiling: Create an action oriented customer
profile, use profiles to create marketing & service programs to
retain & increase value of customers.
3. Predictive Marketing / Promotions: To predict the likelihood of
future events based on customer models & to predict the
profitability of a promotion to encourage customers to do what we want
them to do & achieve the
highest ROI (Return on investment). Predict when a customer is
about to defect / leave us.
4. Life Time Value: To find what a customer is worth in the future
and based on this, find how much you can spend on retaining them
& still make a profit.
Please reply on how to start this activity?
A: Yikes! That's a pretty long list of
"areas to be covered", you are going to be very busy!
Some of it sounds pretty familiar too, like I've read it on my web
site...you might want to get that
book after all...
The creation of retention programs always starts with customer
segmentation, you have to understand the behavior you have before you
can create programs to modify behavior.
That probably means starting with #4, LTV. You want to look
at LTV by segment. Get records of defected customers, put them in a spreadsheet or
database, and determine:
* Average length of time as subscriber
* Average spending over that time
* % of this spend that is considered "profit",
used as a proxy for LTV.
In the beginning, you can use a company "profit" average
for LTV until you get more sophisticated. In communications, the
number often used is EBITDA Margin; ask your finance people what you
should use to determine % of spending that is LTV.
Once finding the average, it is time to segment by different
dimensions and determine the same 3 variables above for all the
different customer segments. For example:
Spending quintiles - highest 20%, high 20%, middle 20%, low 20%,
lowest 20% of
defected customers; what is the average length of subscription,
Product / service bundling - identify different levels of service /
tiers / add on services of defected customers; what is the average length of subscription,
Source of customer - which ads / offers / selling methods
originally attracted the defected customer; what is average length of
subscription, spend, LTV?
Geography - using transmitter locations or other natural boundaries
dividing the defected customers; what is the average length of subscription,
Hardware - group defectors by type of phone or terminal or other
hardware; what is average length of subscription, spend, LTV?
Contract details - if contracts vary widely as to their basic
nature and terms, group defectors by contract type; what is the average length of subscription,
After running these studies you should have enough data to
logically and critically construct your other 3 initiatives in
New RFM Metrics: eMetrics Summit
Jim Sterne's eMetrics Summit is in Santa Barbara June 2 - 4 and
be there speaking on (guess) customer retention metrics and
management. Or, translated into the language of the newly reborn
online display advertising market, the hot new discipline called
"behavioral targeting". Join the party!
The official title of the piece is "LifeTime Value Without
LifeTime" and explores how to predict which customers will be
valuable than others in the future. Why would you want to know
Because future value should drive all you customer investment
decisions, from how much you can afford to acquire a customer to when
to stop wasting money marketing to a customer. If you attend the
conference, be sure to grab me and say "Hi." More info
eMetrics Summit 2004 Santa Barbara:
2004 Santa Barbara
That's it for this month's edition of the Drilling Down Newsletter. If you like the newsletter, please forward it to a friend - why don't you do this now while you are thinking of it? Subscription instructions are at the top and bottom of the newsletter for their convenience when subscribing.
Any comments on the newsletter (it's too long, too short, topic suggestions, etc.) please send them
right along to me, along with any other questions on customer Valuation,
Retention, Loyalty, and Defection right here.
'Til next time, keep Drilling Down!
- Jim Novo
What would you like to
the book with Free customer scoring software at:
Out Specifically What is in the Book
Marketing Models and Metrics (site article