growth calculations considering aggregated or summarized data. In
other cases, it may leverage more advanced time series analysis.
Like query and reporting and OLAP, data mining solutions exist
in every industry—for example, predicting the likelihood of a
customer buying a particular product, switching to a competitor's
product, or defaulting on a loan; identifying false insurance claims;
contracting a certain type of leukemia.
The results of data mining can be fed back into the data warehouse,
data mart, or other data repositories to enrich query and reporting
and OLAP analysis. In query and reporting, once we predict which
customers are likely to purchase a portable DVD player, we can sort
customers by their likelihood to purchase that product or accept a
related offer, and then select the top N customers for a marketing
campaign. In OLAP, we can use data mining to identify which
dimensions are the most predictive for a particular measure. If there
are many, perhaps dozens, of dimensions to choose from, this can
guide selecting dimensions to include in a particular cube analysis.
In addition, the predictions from data mining models on the
individual data records can be fed back into the cube, perhaps as
additional measures that can be included in subsequent roll-ups. For
example, we could predict which offer a given customer is likely to
accept and then include a dimension that rolls up these offers into an
For business people and C-level 2 executives, the details of data
mining are likely to be a curiosity at best. What these individuals
desire are tangible results that they can use to make better business
decisions. From their perspective, knowing which customers are
likely to become next month's attrition 3 statistics is more important
than the technique used to get that information—as long as the
methodology used is sound and the tools trustworthy. At this level,
we can view data mining as a “black box,” as illustrated in Figure 1-3.
“C-level” refers to top corporate management, including the Chief Executive
Officer (CEO), Chief Marketing Officer (CMO), Chief Information Officer (CIO),
and so on.
Attrition occurs, for example, when a customer terminates service or stops pur-
chasing a product, an employee resigns, or student does not return the following
term. Attrition is often the term used in the industries financial services and higher
education. In the telecommunications industry, attrition is referred to as “churn.”