From the October 2011 issue of Treasury & Risk magazine

Where Did Things Go Wrong?

Predictive analytics tools let companies crunch the data to come up with trends and probabilities.

As the world of business speeds up, executives need more insightful analyses to guide their decisions and they need them quickly. With that sense of urgency comes a greater interest in predictive analytics, software that sorts through a range of information to come up with future probabilities. Robert Kugel, a senior vice president at Ventana Research, distinguishes predictive analytics from traditional forecasting. “The real value of predictive analytics is comparing actuals to predicted results, disaggregating why things are different sooner, and then being able—once you’ve spotted the discrepancy—to more accurately forecast how things will turn out next,” Kugel says. “It isn’t so much creating a forecast, which is almost certainly going to be wrong, but using why it’s wrong and how wrong it is to make better decisions on what to do next.”

He cites the example of a retailer finding a certain product isn’t moving as fast as expected and using that information to promote the product more aggressively or mark it down.

Although Ventana’s research shows just one in seven companies now use predictive analytics, Kugel says that adopting the software has become more “feasible” for companies.

Using predictive analytics used to mean hiring Ph.D.s, he says. “But that’s no longer the case. There’s some training involved, but larger companies have the resources.” And in-memory computing, in which data are on chips rather than on the network or in hard drives, has reduced the time it takes to get results, he says.

Kugel says the technology is more useful for large companies than small ones. “In small companies, somebody stands up in a cubicle and yells something when it looks like things aren’t going right. And someone else jumps up and says, ‘Here’s how we can fix it,’” he says. “The payoff is definitely greater for large companies.”

Companies that provide predictive analytics include SPSS, which was acquired by IBM in 2009, and SAS.

Predictive analytics “enables you to create models to better anticipate changes in the market,” says Michael Lock, a senior research analyst for business intelligence at Aberdeen Group, who notes “the increased importance people are putting on becoming more anticipatory.”

Companies are also dealing with an “explosion” in the amount of information they have, Lock says. “The challenge is maximizing the value of the large quantities of data companies have, drinking from the fire hose and transforming that glut of information into something that’s going to mean something to the business.”

 

Comments