Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, by Eric Siegel, John Wiley & Sons, RRP£18.99/$28
The jacket for Eric Siegel’s Predictive Analytics: the Power to Predict Who Will Click, Buy, Lie, or Die, contains a description of the “omnipresent science” of predictive analytics as affecting “everyone, every day. Although largely unseen, it drives millions of decisions, determining whom to call, mail, investigate, incarcerate”.
It is an ominous phrase, coming at the same time as Edward Snowden’s claims that the National Security Agency had access to information from the servers of internet companies such as Facebook, Google, Microsoft and Yahoo.
The data, the US whistleblower has said, are used to track foreign nationals suspected of terrorism or spying.
The claims have cast a shadow on the use of data by companies. The book’s author, a former Columbia University professor, is rather more gung-ho about the potency of using data, possibly unsurprising as he has also founded the Predictive Analytics World series of conferences.
The book shows such techniques have been applied by organisations to forecast, for example, your buying decisions, or the likelihood of your leaving a job, your health and date of death. The modelling methods vary but all use data, from commercial transactions, social network postings or human resources records. These are harnessed to anticipate behaviour, in order to drive and automate decisions by retailers, employers or healthcare providers.
Prof Siegel reveals insights into consumers’ habits, such as a pharmacy chain’s findings that if you buy nappies you are likely to buy beer, or an insurer’s discovery that a low credit rating means more car accidents.
He looks at the Flight Risk program, run by the PC maker HP, which predicted which staff were most likely to leave. In a pilot group, HP was able to reduce turnover from 20 to 15 per cent.
He also shows the limitations of predictive analytics – for example, by assuming too much. He cites the case of the Berkeley professor David Leinweber, who discovered that the annual closing price of the S&P 500 stock market index could have been predicted from 1983 to 1993 by the rate of butter production in Bangladesh.
“His analysis was designed to highlight a common misstep by exaggerating it,” writes Prof Siegel. “It’s easy to find correlations by searching through a large number of financial indicators across many countries, just by chance.”
The book is littered with lively examples, although the short chapters and subheadings lack focus.
Prof Siegel only briefly raises the privacy implications of data use, citing a rather flip reference to the comic character Spider-Man’s uncle: “With great power comes great responsibility.”
More seriously, he states that: “[This] is an important, blossoming science. Foretelling your future behaviour and revealing your intentions, it’s an extremely powerful tool – and one with significant potential for misuse.”
By Emma Jacobs
Originally published at The Financial Times
Response to this book review from author Eric Siegel: I would respond to the book critic’s main point by saying that the book in fact does something unprecedented among those written by technology practitioners: It addresses privacy and other civil liberty concerns up front – in Chapter 2, which is entirely devoted to the matter (not briefly or fliply, as Jacobs claims). To access a number of articles covering these ethical concerns, go to the “Press” page on the book’s website.
 |
 |
 |
