Micro versus Macro Analytical
Approach to CRM
First Published 3/28/01
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.
Users of CRM analytical systems are spending too much time and money on a
"micro" view of the customer while ignoring more profitable
"macro" analysis. The micro view, perhaps encouraged by the
"1-to-1" philosophy, is fine if you have the skills to
really understand and implement it. But managing micro-segments
or segments of one customer requires detailed and comprehensive
experience maximizing the profitability of macro customer segments
first; otherwise you set up a classic "can't see the forest for
the trees" scenario.
Many of the CRM-oriented micro segmenting approaches also lead to what I'd call "over-targeting."
This occurs when there is an attempt to subdivide customers into
groups so small they are no longer meaningful or practical for the
purpose of measuring and increasing customer value. Then there's
a backlash to this self-inflicted situation, usually "we're
drowning in data" or "we have too much data but not any
The micro approach skips over more fundamental macro or group analysis
which leads directly to higher ROI marketing. For example, what is the
average value of customers generated by banners versus newsletter ads?
The value of customers whose first purchase is over US $100 versus
customers whose first purchase is under US $100? These are
behavioral characteristics known to affect future customer value in
pronounced ways, and analysis of these attributes seems to be getting lost in pursuit
In the area of collaborative filtering, or product
recommendations based on past purchase history, who has not bought a
gift for someone in a product category they held no personal interest
in, only to be inundated with product offers related to the gift
As a retailer, do you want to "sell them anything" or
"sell them a specific thing"? By over-targeting people
like this, you risk losing a
sale by making offers so targeted they have no relevance. It
would be much better to recognize the person as someone with certain
overall purchase patterns, and sell them "anything they want to
buy" - a more
To continue this example, a customer who was a heavy cookbook buyer in
the past who stopped buying, then for some reason comes back to the
site, should not be met with "you bought cookbooks so here's a
cookbook offer." Perhaps the reason they haven't been back
is they own enough cookbooks, or bought them as a gift for
Recency of purchase is the single most powerful predictor of future
purchase; and Recency has nothing to do with "what" they
bought but "when" they bought it. This customer should
be met with "you haven't
been here for awhile, so here's $3 off anything you want to buy,"
not "here's another cookbook offer." A macro kind of
offer based on their overall behavior, not a micro offer targeted so
finely as to discourage an additional purchase from a "lost
customer" who decided to come back one more time.
The longer it has been since the customer made a purchase, the higher
the discount should be. This overcomes inertia and gets the
customer back on a buying track. The reverse is also true;
current active buyers should not generally be offered discounts until
they show signs of defecting. Very recent purchasers are highly
likely to make another purchase and discounting to these people is
just throwing money away. Ignoring these basic behavioral
patterns leads to significant customer value loss, as inappropriate
discount timing erodes margin.
In summary, people seem to have "skipped over" a lot of the
tried and true methods of customer value measurement and management in
favor of the latest "1-to-1" toys. The result is a ton
of barely usable data,
ineffective and expensive marketing, and a lack of awareness of what
really drives customer behavior and how to take advantage of it to
increase LifeTime Value.
Down book teaches you these tried and true methods for making
money with database marketing. Data mining not required.