Predictive analytics helps firms to get "about three per cent better at guessing" what a consumer would like to buy, according to the founder and chief of Applied Predictive Technologies (APT), Jim Manzi.
"Realistically, what predictive modelling does not say is: ‘Now I know for a fact that you want this'. What it means is I get about three per cent better at guessing - and that is worth a lot of money," he told Computing.
APT's customers include Boots, Asda, clothes firm Abercrombie & Fitch and Kraft Foods. Its software tracks customer habits in an attempt to help retailers, banks and other organisations to predict future outcomes and alter their strategies accordingly.
Last July, Computing looked into the possibility of supermarket Tesco changing health insurance prices for customers depending on their shopping habits.
This trend has also been highlighted by The New York Times, which reported that US retailer Target predicted a high school girls' pregnancy by analysing her shopping habits, and then with the use of predictive analytics, offered her a selection of products that are typically bought in early pregnancy.
The retailer reportedly sent out coupons for other products that she might have needed for her new baby. The girl's father, meanwhile, had not been aware - up until receiving the coupons - of his daughter's situation.
Manzi said the Target incident should not be used to give predictive analytics a bad name.
"What we're trying to do with predictive analytics is to get a little better at what you're already doing and so really our technologies could in theory be used to build experiments and from that build models to say that a customer ID is more likely to respond to the following offer than another customer ID," he said.
"But I think [in the case of Target and the pregnant girl] it's the mailing and not the predictive modelling that creates this problem," he added.
Last month, the Office of Fair Trading (OFT) launched an investigation into the extent to which businesses are using customer data to target consumers with personalised prices.
Manzi said that if society became uncomfortable with consumer-facing businesses using predictive analytics tools to tailor personalised deals then government could legislate against it.
"If there was a legal regulation that would forbid that, businesses wouldn't do it. My experience with large companies is that they will obey the law. I'm sure that there are laws and regulations that mean companies can't set out insurance rates in that way, for example," he said.
"All predictive modelling and analytics are doing is helping companies to get a little bit more accurate at predicting who wants what," he concluded.
By eliminating high entry costs for big data analysis, you can convert more raw data into valuable business insight.
A discussion of the "risk perception gap", its implications and how it can be closed