New RFM: Using LifeTime Value for Ad
# 41: Jan 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
# Who we are, Where we are going
# Question - New RFM: LTV and Ad Budgeting?
# Question - New RFM: Modifying RFM?
# New RFM Metrics: Take 10 on Retention
Hi again folks, Jim Novo here.
There is not usually a lot of "meaty" content in the
December trade magazines, so for the January newsletter I replace the
"links" section with a few comments. Rather than make
all kinds of inane forecasts on what will be hot and not in 2004, I
instead like to update everyone on who you are and where this effort
will be going this year. We then have a couple very meaty
questions from fellow drillers on using LTV for budgeting and creating
modifiers to the basic RFM approach.
But first, a couple of job opportunities for those with web
Foster & Smith, Wisconsin
Karta, Los Angeles
If you have a job in web metrics / customer analysis you'd like to get in front of this newsletter audience, just send me a link.
Who we are, Where we're going
Globally, the majority of you are primarily interested in online
customer marketing, which should be no surprise given the medium we're
using. The interesting thing that happened this year is the
increasing interest in offline customer marketing, for example,
"Can we take what we are learning online and use it offline to
accomplish the same High ROI?"
This is a fascinating development (at least to me) because those of
you who have been with me since 1999 - 2000 know the Drilling Down
Project started as an effort to teach onliners how to "port
over" the ideas offline direct / database / relationship
marketing players are using to understand customers / business
processes and reduce costs while increasing profits (CRM, if you
want). Now the onliners are going back across the wall and
teaching the offline part of their company how to make it all happen -
I love it!
You generally comprise 3 groups:
1. People at the beginning of the learning curve, working to
get a handle on customer source tracking and the cost to acquire a
customer. These folks are primarily interested in web analytics
and conversion of visitors into buyers, leads, relationships, repeats,
2. People who have passed through this stage and are now
looking at the more complex issue of visitor / customer value over
time, for example, a customer I acquire for $10 may only be worth $5
in profits over the next 6 months, whereas a customer I acquire for
$50 may be worth $3000 in profits over the next 6 months. These
folks are primarily interested in the simple customer models outlined
on the web site and in my book.
3. Advanced folks, who already had offline experience in
customer value management and High ROI Customer Marketing before the
whole online thing happened, and the rest of the world started caring
about "CRM." These folks were likely doing the same
kinds of things CRM vendors now promise on a Unisys A Series the size
of a restaurant refrigerator and remember a hardware company named
Burroughs and disk packs (data storage) - hey buddy, can you spare a
byte for a good cause?
By the way, I'm in the last group, circa 80's.
Yes, this general topic of using information on customer behavior
to reduce costs while increasing profits has been chugging along for
that long. Did you think it was a new idea? It actually
started in the 50's, and back then, it was all done by hand. No
kidding. You gotta start somewhere, right?
The content format I've been using this past year seems to be the
most successful ever, that is, to not write a lot of long-winded
articles but instead answer more questions, and try to group the
questions in each issue by the 3 audience types above - web analytics,
simple customer models, and more advanced ideas proposed / championed
Of course, without the long-winded articles there wouldn't be a
library of content for me to link to and you to learn from, so that
phase had a purpose. But now, much of the core content is sent
directly by e-mail to you when you subscribe to the newsletter. So I intend to continue the Q & A format this year, with the
occasional article when I think an issue needs to be addressed. If you disagree with this approach, let me know by replying to this
e-mail with your comments and preferences.
Enough of the commentary, on to questions, which this month are
targeted to Group #3. Hopefully those of you in Groups 1 & 2
above can benefit by seeing what's "down the road."
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: LTV and Ad Budgeting
Q: You'll be pleased to know that your book has been
the catalyst for setting up a user forum on RFM
which now has around 30 actively contributing members. Not many,
but not bad considering the only way the word was spread is by word of
A: That's great!
Q: We operate a retail stores, a catalog, and web
site. I want to make sure I understand how LTV
tables help in drawing up your budget and marketing plan. Obviously in the context of my assignment I make a few assumptions,
because hardly any data is given. My take on the process is as
- Work out rough average customer spend
- Work out how much I want to increase my sales in FY 2004
- Work out how many customers I need to get me to sales budget
- Now split the overall number of customers by distribution channel
(distribute customer spend through each as I think appropriate)
- Work out an LTV table for Shop, Print Catalogue, and web site
- Take a percentage of the 'LTV at Net Present Value' figure for
2004 (say 20%) which will give me the amount of money I can afford to
spend on my marketing budget.
- Create a budget and campaign for shop, catalogue and website for
- Write analysis
Is this process correct?
A: Well, it's hard to define "correct" not
knowing the specific requirements of the task,
but I think you are generally on track. A couple of comments:
** "How much I want to increase my sales by in 2004" is a
retail concept you may find difficult to crunch into direct
methodology. Remember, direct is much more focused on profits
than sales. In retail, you follow a standard "percentage of
sales" model and as long as sales go up, profits are expected
to go up. Not so with direct. For one thing, in direct
actions you take today have value in the future. If you want,
you could say, "OK, I know a direct customer will spend $100
first year, $80 second year, and so forth, and crush direct into a
retail model, but to me this defeats the beauty of direct, because now
you are using a "periodic financial accounting" system
rather than a "customer accounting" system. More on
this idea here.
** In fact, I'm not sure what the value is of including "the
stores" is in the process, other than to determine what gross
revenue needs are. The stores involve a completely different
model than catalog and web, and I think down the line you are going to
get to a place where it is hard to "force" the stores into
the direct model. Now, if you are assuming that these stores
possess some way to behave like a direct operation - that is, they
always have customer-level knowledge of all customer transactions -
then maybe it works for you overall.
But in the stores, if you have walk-in traffic that transacts
anonymously, then you bust the direct model, you can't track LTV and
so on. So retail is much more like a "snapshot" idea
(activity today has value today) than a "movie" idea
(activity today has value over time). Offline retailing is
generally managed in the aggregate, not at the customer level. Direct is managed at the customer level.
These two models will
For example, you do newspaper ads for the store and you see a boost
in traffic. You have sales - product costs - ad cost = profit,
in the aggregate. There is no measurement at the customer level,
no LTV involved in this, no future stream. In direct, you know
who you attracted and can measure future profits. A campaign
that looks like a loser today can be a winner in 2 months. From
a management perspective, to try and "force" the retail
store manager to operate under the direct model would probably result in
sub-optimal performance from a manager who just doesn't "get
Q: My second question draws on your experience of the
catalogue market. I need to make an assumption on costs versus
profit. So, what I intend to do (presuming this is correct) is
to take a percentage of the 'LTV at Net Present Value' figure for 2004
(say 20%) which will give me the amount of money I can afford to spend
on my marketing budget.
Question is, what should this percentage cost be for the shop,
catalogue (print) and website distribution, bearing in mind that they
have relatively little experience of catalogue/web marketing and
already own three shops.
A: Yes, and do you see where this is leading? 20% NPV (Net Present Value) of LTV (LifeTime Value) is OK for direct,
3% of retail sales is what is typically spent in retail. How do
you reconcile this? You can't. This is what I mean by the
models colliding. If you went to a retail store manager and told
him he could spend 20% of the NPV of the LTV of the average customer
on ads, he would at least be a bit confused, at most think you're
Put another way, retail operates on a single transaction basis
(snapshot), direct takes into account transactions over time (movie).
Retail cannot afford to pay more than 3% of sales because they cannot
target and cannot track. In direct, you generally lose money on
the first purchase from the customer. Retail management could
not tolerate that approach, they try to make money on every sale.
It's simply a different operating model, and this difference is driven
by the difference in cost components. Retail is a high fixed,
low variable cost model; direct is low fixed, high variable cost
model. The seeds of "model collision" are planted in
this fundamental difference
between the two cost structures.
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: Modifying RFM
Q: I've been through your site and have found it
immensely useful, thanks a lot for the wonderful resource!
A: Well, thank you for the kind words, glad you found
Q: As head of Data analytics at a leading direct and
loyalty marketing agency in India, I've seen a fair bit of action on
the RFM front as well, and thought I should share a couple of
'additions' to RFM we use frequently. You might, at this moment,
be rolling your eyes in despair because there's a million guys who've
mailed you with this same suggestion, but I guess that's a chance I'll
A: Trust me - there isn't a million guys (or gals) who
are doing any of this stuff yet, the focus is still on getting
customers and not keeping them. Your odds just improved!
Q: Here's one - RFM(C):
We've found addition of 'Consistency' very useful. Two
customers with the same Recency and frequency score can have widely
different behaviours. One bought ten times in Jan and once in
November, the other bought once every month from Jan-Nov. Introducing
Consistency differentiates between these two species of
customers. Consistency measures could be Variance (monthly value
/ frequency), or simple ratios.
A: Yes, this is what I refer to as "rate of
change" or "acceleration / deceleration" and I have
actually had a few e-mail discussions on this one. It can be
seen as a "dx / dy" derivative idea if you take it further
down the road. This is an underlying theme in my book and
referred to several times though I don't go into it explicitly because
the book was not intended for an expert audience.
For example, if you track the RFM scores of an individual over
time, the more dramatic the change in score (a "rate" of
sorts), the more highly predictive it is of an acceleration (becoming
a better customer) or defection (ceasing the relationship). You
can also get to some of this by including a Latency screen, e.g. in
your example, these customers would have different Latency
Q: And this one - RFM(CS):
In retail situations where different product groups may have to be
treated differently, Saliency could become relevant. It's a
measure of the importance of that product group in the customers
basket. For instance, a customer with high RFM in Menswear where
Menswear constitutes only 2% of his / her sales, should be treated
differently from a customer with a lower 'M', but where Menswear
constitutes 60% of the sales.
A: Absolutely. As a matter of fact, there is a
specific discussion of this in the book. One of the great
untapped areas for direct retail is category migration, where
customers who start buying one kind of product need to be
"migrated" to another category or they defect and stop
buying completely. Some customers make the migration themselves,
and become long term, very profitable customers. Many do not
migrate. If you can "push" them to migrate with the
right kind of promotions delivered at the right time in the
LifeCycle, you get customers who stick around longer and achieve
much higher value. This kind of promotion strategy has the
highest ROI of any customer marketing I have ever seen because it
addresses fundamental LifeCycle issues. To work well, it has to
be tightly linked with good merchandising, something not very many
companies are good at (at least on the web). So it's not easy,
but very profitable.
As a matter of fact, this second issue ties to the first. Essentially, you want to hit the customer with the migration offer
when you see deceleration in the initial or primary category. For example, customer buys once a month like clockwork for 4 months in
a single category, then skips a month. This triggers the
migration offer. But you first have to understand the
"category pairs" to make this work - e.g. long-term
customers who end up buying fashion started in jewelry, long-term
customers who end up buying kitchen started in home fashion, etc.
You can "reverse lookup" these pairs by analyzing high value
customers over time.
Q: If at any point you want to know more about the
sort of work we're doing related to analytics for relationship /
loyalty programs here in India, mail me!
A: Consider yourself mailed!
New RFM Metrics: Take 10 on Retention
If you would like to know more about how to use the new RFM metrics to improve your profitability on the web, check out the free "Take 10 on Retention"
package I wrote. It includes a 10 minute presentation on the strategy and
reporting behind increasing web customer ROI using simple predictive
Here's the idea in a nutshell: when you make investments, you
expect the value of them to rise in the future. You have web
investment choices to make - ad design, media, building out content,
etc. Retention metrics tell you which of these investments are
the most likely to generate increased profits in the future.
Click here for the Take 10 on Retention
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