Drilling Down Newsletter # 18 - March 2002 -
Guide to Web Analytics, LTV
Drilling Down - Turning Customer
Data into Profits with a Spreadsheet
Customer Valuation, Retention,
Get the Drilling Down Book!
Now also available online through
Amazon and Barnes & Noble
Drilling Down Newsletter # 18 -
In This Issue:
# Topics Overview
# New Report on Using WebTrends
# Best of the Best: Customer Marketing Links
# Tracking the Customer LifeCycle: Examples
# Questions from Fellow Drillers
Hi again folks, Jim Novo here. This month, we have a new behavior analysis product you might find helpful,
and some "expiring links" to "must read" articles. Also up is Part 3 of the Beauty Salon Example
(Maximize), and a Real World question from a practitioner who wants to
prove to management they have to spend less to make more money.
Let's do some Drillin'!
New Report on Using WebTrends
Those of you who have been with me a while know I have been quite involved with the discipline of server log reporting using WebTrends.
I was convinced locked away in those logs are many of the answers needed to improve the productivity and profitability of web sites.
After several months of testing my log analysis ideas, I'm pleased to announce I've picked at least part of the web
The original Excel spreadsheet I developed to create actionable server log metrics you can track over time has morphed into a report -
The Guide To Web Analytics: How to Understand and Use Your Web Trends to Maximize Results
This report has garnered a lot of attention, and you are sure to be hearing more about it in the near future from some very well known people in the industry.
If you'd like to get a preview of what it's all about, you can still download the metrics calculator spreadsheet at the core of the report
The report itself goes into great detail on how to use log analysis and the calculator to create actionable metrics.
It also includes advice on setting up WebTrends to give you the "real picture" of activity on your site, and describes how to use these metrics to create "accountability" for changes implemented to a web site.
To see what others are saying about it and get the report, click
Best Customer Retention Articles
This section flags "must read" articles moving into the paid DM News archives before the next newsletter is
delivered. If you don't read these articles by the date listed, you will have to pay $25 to DM News
to read them from the archives. The URL's are too long for the newsletter, so these links take you to a page with
more info on what is in the article and a direct link to the article.
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.
Only one "must read" DM News article this cycle (Expires 3/19); and
this one is a little different. After receiving a ton of e-mail from you folks on this particular article, and agreeing the claim in the title was a substantial twisting of the facts, I wrote an entire article explaining what to watch out for when interpreting case study metrics.
Check both the original article and
my review out. And...
Due the lack of "must read" DM news articles, I also provide a link to a fantastic article written for the McKinsey Quarterly, which tells how emerging markets are "doing CRM" without all the fancy software - and are quite successful at it.
They suggest - hang on to your hat - that quite possibly, this approach might work in developed countries.
Really? Wish I had said that - 3 years ago. To check
out this article (free registration required), click
Tracking the Customer LifeCycle:
Real World Examples
Note: If you are new to our group and want to know more about the following ongoing discussion, the background theory is
The "Beauty Salon Example" (3 parts total) starts right here.
A Tale of Two Hair Salons
Part 3: Maximize
Recall the events of Part 2
(Manage) - the owner of Salon B used a
simple spreadsheet to find the average number of days between visits of best customers was 40 days.
The owner sits down each week and mails a "where have you been?" discount postcard to each best customer who has not made an
appointment in the past 40 days. This program has been quite profitable; appointments from these customers
have paid for the postcard mailing many times over.
Despite this success, two things bothered the owner of Salon B.
The first was some customers who responded said, "Thanks for sending me the discount, I was going to make an appointment in the next couple weeks and I'm happy to get 15% off."
The second was the 75% of Salon B's best customers never responded to the mailing.
And as far as we know, the owner of Salon A is still wondering when best customer Mary Lou is going to make another appointment, and still does not know how many "Mary Lou type" customers there are.
The owner of Salon B thinks:
Some customers who respond say they were going to make an appointment in the near future. Clearly, I have sent the a discount to these customers to soon, and I am giving up margin.
But many best customers don't even respond, which probably means I am sending the postcard too late.
How can I figure out how to Maximize my results?
Well, fellow Driller, have you got an idea? You know Customer
Retention is all about process:
Action - Reaction - Feedback -
The owner of Salon B has taken an action, and there has been a Reaction.
How should the owner go about Analyzing the Feedback?
The owner of Salon B then has an idea:
What about this group of customers who said "they would have scheduled anyway without the
postcard." Are they similar in any way?
If there is a common reaction to the postcard among these customers, perhaps there is a commonality in the behavior or backgrounds of the customers.
If I can find the key linking these customers together, perhaps I can understand why this is happening.
The owner of salon B goes back to the CRM software (a paper appointment book and the customer spreadsheet).
The owner has entered "response date" in a spreadsheet column for each customer who responded to the postcard and any comments.
The owner sorts the customers by the responders and looks at those customers who said "would have scheduled anyway without a
For each customer who responded and said this, the owner looks the customer up in the appointment book to find
"Long hair cuts!!!!," the owner exclaims. "They all have long hair
cuts!," which the owner immediately realizes is the problem
with the discount postcard mailing program.
The owner thinks:
Best customers with long hair styles can come in
much less often than every 40 days, even through the average of all best customers is a cut every 40 days.
So customers with long hair cuts are getting the postcard too early - they're not really
"defected," and schedule a planned appointment with a discount I did not have to offer.
They should get a postcard possibly at 60 days, or even 90 days or
longer after their last appointment.
Since the owner has a lot of customers with long cuts, most are getting the postcard too early for the cut.
This explains the low overall response rate.
Best customers with short cuts however, are probably getting the postcard too late.
By the time I get them in the mail and they reach the customers with short cuts, it could be too late, they may have already gone elsewhere for their
short hair cut.
The owner of Salon B resolves to recalculate the average days between appointments separately for best customers with long cuts and best customers with short cuts.
Next month, we'll check back in with the owners of Salon A and Salon B
and see what they cook up. We still don't know what happened to Mary Lou, the "tardy best customer" from Salon A.
And what will the owner of Salon B do with customers that have "medium
cuts"? Only the data knows...
Go to Part 4 of the Beauty
I can teach you and your staff the basics of high ROI customer marketing using your business model and
customer data, and without using a lot of fancy software. Not ready for the expense and resource drain of CRM?
Get CRM benefits using existing resources by scheduling
Questions from Fellow Drillers
Q: I'm a "long time listener, first time caller," and a big fan of your site and your approach to data-driven marketing.
I also have two copies of your book - one was not enough.
A: Well, thanks for your kind words. I love the talk radio reference, that is so funny.
Never though about it like that, but makes perfect sense! Glad to know I'm actually helping people
with the book too.
Q: I have a question relating to some work I am doing now with our best customers that other users of your site may have.
I work for a medium sized mail order company selling skincare products (high margin) via space ads and direct mail. Our best customer "Gold Club" has about 8000 members at the moment, although members are being promoted and demoted all the time.
According to my initial analysis, if a member does not purchase a product for more than 60 days, the chances are that they are defecting. I would like to attempt to bring them back with an offer, and leave those that don't reply for at least 6 months for a deeply discounted "kickstart" offer (although the logistics of sending out very small mailings are a
A: This is a common and logical approach, particularly for "renewable
You don't say what the product is, but if it is "typical" skincare product, it has a sales cycle very tightly tied to product use.
In this case, Latency usually makes more sense to use than Recency as the primary trigger for a campaign.
If you have people on different "supply amounts" (some get 30 day, some 60 day, some 90 day) then it can get confusing and be ineffective to just say "60 day Latency gets
30 day should get 30 day mail, 60 day should get 60 day mail, etc. in the optimum scenario.
You have to, obviously, understand what the costs are to do that and optimize around the schedule with the most bang for the
Q: I am having some problems convincing management that these former best customers are a lost cause,
they don't like the idea of "losing contact with our customer base" and want to prolong the contact longer to see if we can bring some of these customers back.
In fact, when I first looked at the data around 2000 people hadn't purchased for at least 6 months, and were
receiving mail from us every month at $1 US a pop.
A: This is a very typical attitude in small mail order companies, particularly high margin ones.
But what they don't realize is there is tremendous leverage in there, absolutely tremendous.
The question for the owner is something like this: would you give up 10% of annual topline sales for a 30% improvement in annual profits?
That should get their eyes sufficiently bugged out to listen a bit more.
Q: Is there any sense in the "losing contact" hypothesis or is it best to let customers go without undue struggle and prevent them from getting to the defection stage in the first place? What about a short "grace period" that includes customers a bit longer than necessary?
There may be a situation in which a certain RF group is unprofitable for a DM, but the few people that do repurchase buy again, thus coming into the black again (keeping in mind the high margins.)
Any insight you have would be greatly appreciated.
A: Look, most mail order companies blast mail to people up to 2 years after the last purchase.
I think that's silly and wasteful, and you are on the right track.
They have to learn to "let go." The grace period idea is a bit unclear to me, but in general, there is some "optimized" grace period, yes.
For example, a customer on a 60-day supply cycle might most profitably be approached 10 days before it runs out, 5 days before it runs out, or 5 days after it runs out, or 10 days after it runs out. When your are talking
profit (as opposed to response), there is no way to know without a
test of the concept
My suggestion would be do what I call a 30-60-90. This idea is very similar to the "RF Grid" from the Drilling Down book in nature:
1. Segment by supply cycle, if there is one. Don't mix in "90 day supply" people with "60 day supply people" -
if it is efficient to mail them separately. If not, well, this will still work, but won't be as profitable.
2. Take a 10% random sample of the population by Latency - 30 days before supply runs out (if applicable), 30 days after, 60 days after, 90 days after, 90-120 day after, 120-180 days after, etc if applicable.
Execute the mailing to the segments.
3. There should be a 30 day (or larger, depends) window where it is the most profitable to mail based on supply run-out date.
It should be very clear how much money is wasted "chasing" dead customers in say, the 120-180 day Latency window.
Use this as ammo with management. If they still don't want to let go, say you'll reallocate the money to really work on the "windows" (for example, even though 120-180 day is not profitable, 120-150
only might be) and then you'll do a bi-annual or annual postcard to "the
dead," to keep in touch.
That way you efficiently "scrape" the dead for the few you can get back for very little cost, and focus the majority of the resources where the money is.
Hope the above helped. Thanks again for the kind words, and good luck with it.
Any questions, let me know!
If you are a consultant, agency, or software developer with clients
needing action-oriented customer modeling or High ROI Customer Marketing
program designs, click
That's it for this month's Drilling Down Newsletter. Any comments on
it (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
'Til next time, keep Drilling Down!
- Jim Novo
Copyright 2002, The Drilling Down Project by Jim Novo. All
rights reserved. You are free to use material from this
newsletter in whole or in part as long as you include complete
attribution, including live web site link and/or e-mail link. Please
tell me where & when the material will appear.