International Journal on Criminology Volume 4, Number 2, Winter 2016 | Page 75
International Journal on Criminology
just Silicon Valley but also the military, our universities, the world of publishing and
even Hollywood!
Let us start with some of the books that have been published on the subject.
Predictive Analytics for Dummies is fairly typical, the authors introducing us to the
holy grail that is predictive technology with great enthusiasm and cut-and-dried
assertions. But what is actually between the covers? The book, which is full of
business school marketing banalities, is based on nothing more than extrapolation.
Neither the exhaustive table of contents (which runs to seven pages) nor the index
(which is twice as long) features a single reference to the idea of time or temporality—
which are clearly crucial concepts in the field of forecasting, whether it is predicting,
anticipating, or making assumptions.
Here is another book that is typical: Predictive Analytics, a work by a gimmick
merchant who never even defines the term “prediction”. Observing behaviors or
habits, using your savvy, optimizing, extrapolating, and estimating probabilities—is
that all forecasting or predicting?
Take an example: when a store sees a customer buying a maternity dress, do
you really need a super computer to ask her if she also wants a baby bottle and diapers?
Likewise, when a reader orders a detective novel on a website, do you need to be an
internet genius to suggest similar titles? Is not it the same thing as the commonplace:
“Customers who bought this item also bought…” that we see on commercial websites?
But that is how this rather crude arrangement works, endeavoring to make sense of
data that is diffuse, chaotic, and collected in bulk. It is a system that has everything to
do with optimization, or common-sense marketing, but nothing to do with predicting.
A case in point: the speed and power of computers is (in this instance) undeniably
superior to that of the human brain, and software sorts e-commerce customers into
four categories:
A—People who buy a product but ignore the advertising (put them on the
back burner).
B—People who only buy a product when there is no advertising (on the back
burner).
C—People who surf the internet without buying anything on-line (on the back
burner).
D—People who do not buy anything without advertising but do buy if they
see it:these “definite receptive customers” are put in a specific database then
bombarded with advertising.
The repeated use of the word “predictive” in Predictive Analytics will exert a
hypnotic effect on the reader. But nothing in the book really helps to predict anything.
And seeing as book series bearing this title are all the rage, what we have here is more
74