Page 375 -

Chapter 9 Business Intelligence Systems
374
new loyalties to new brands. Thus, when people start their first professional job, get married, have
a baby, or retire, retailers want to know. Before BI, stores would watch the local newspapers for
graduation, marriage, and baby announcements and send ads in response. That is a slow, labor-
intensive, and expensive process.
Target wanted to get ahead of the newspapers and in 2002 began a project to use purchasing
patterns to determine that someone was pregnant. By applying business intelligence techniques
to its sales data, Target was able to identify a purchasing pattern of lotions, vitamins, and other
products that reliably predicts pregnancy. When Target observed that purchasing pattern, it sent
ads for diapers and other baby-related products to those customers.
Its program worked—too well for one teenager who had told no one she was pregnant. When
she began receiving ads for baby items, her father complained to the manager of the local Target
store, who apologized. It was the father’s turn to apologize when he learned that his daughter was,
indeed, pregnant. 4

BI for Entertainment

Amazon, Netflix, Pandora, Spotify, and other media-delivery organizations generate billions of bytes
of data on consumer media preferences. Using that data, Amazon has begun to produce its own
video and TV, basing plots and characters and selecting actors on the results of its BI analysis. 5
Netflix decided to buy House of Cards, starring Kevin Spacey, based on its analysis of custom-
ers’ viewing patterns. Similarly, Spotify processes data on customers’ listening habits to determine
locations where particular bands’ songs are heard most often. Using that data, it then recom-
mends the best cities for popular bands and other musical groups to perform. 6
A popular adage among marketing professionals is that “buyers are liars,” meaning they’ll
say they want one thing but purchase something else. That characteristic reduces the efficacy of
marketing focus groups. BI produced from data on watching, listening, and rental habits, how-
ever, determines what people actually want, not what they say. Will this enable data miners like
Amazon to become the new Hollywood? We will see.

Just-in-Time Medical Reporting
Practice Fusion Inc., a medical record startup, provides injection notification services to doctors
during exams. As the doctor enters data, the software analyzes patient’s records, and if injections
are needed, it recommends that the doctor prescribe them as the exam progresses. It seems to work,
too. During a 4-month study period, patients whose doctors were using the recommendation service
prescribed 73 percent more vaccinations than those in a control group who did not use the service. 7
The service is free to doctors. Practice Fusion is paid by Merck, a pharmaceutical company. While
Practice Fusions software recommends many products that are not sold by Merck, it also recom-
mends many that are. Hence, the service stands on the blurry edge of medical ethics. Should a drug
company provide software that recommends its products to doctors “for free”? If the injections are
truly needed, who could object? On the other hand, how unbiased is the Practice Fusion software?
Setting the ethical issues aside, Practice Fusion provides an excellent example of data mining
and reporting in real time. During your career, there likely will be many examples of middle- of-
the-sales-call sales assistance.
Given these examples, we next consider the process used to create business intelligence.

Q9-2 What Are the Three Primary Activities

in the BI Process?

Figure 9-3 shows the three primary activities in the BI process: acquire data, perform analysis,
and publish results. These activities directly correspond to the BI elements in Figure 9-1. Data
acquisition is the process of obtaining, cleaning, organizing, relating, and cataloging source