• silicon.com
• Management
• CIO Insights

By Tim Ferguson, 15 May 2009 13:00

NEWS

With businesses gathering more data than ever before, they are using increasingly sophisticated tools to make the right decisions using that information.

At predictive analytics company SPSS' annual European conference in Prague this week customers discussed how they have used data mining and modelling to improve their businesses.

Italian car manufacturer Fiat has been using the SPSS PASW Modeler technology (formerly called Clementine) since 2005 to improve the way it interacts with customers.

Giovanni Lux, customer intelligence manager for Fiat, told silicon.com the company has been able to retain more customers and reduce churn, and said customer retention has improved by "six or seven per cent" since the company started to use the technology, with 54 per cent of Fiat owners replacing their cars now buying another Fiat Group car.

Lux added: "The point is we need to show that through the use of predictive analytics we sell more cars."

The technology from SPSS allows the company to collect data and create models of how customers are likely to respond to the brand, the actions of dealers and the online part of the business.

The data is collected through questionnaires completed by existing customers when they interact with the company, as well as those considering buying a Fiat. The company also buys in data - such as how often people change their cars - from other public sources.

Once it has this data, Fiat can then use data mining techniques to work out how it can target customers better and how dealers can improve the services they provide existing customers.

The company also employs web analysis to generate leads and capture new customers. Lux said: "The main focus is how to be effective in managing the data and the brands. We analyse all the data coming from the web."

Insurance group RSA has also been using SPSS technology to better understand its customers.

RSA analytics manager, Simon Dudley, said the company needed to do more with data analytics because customers can easily change insurers - around two-thirds of customers going to price comparison websites whenever they renew insurance policies. Customer satisfaction is therefore a key part of the success of the business.

Speaking in Prague, Dudley said: "We bought into the whole vision of being a predictive enterprise."

Much of the customer data RSA obtains is through email questionnaires and once obtained can be used to improve processes and the way the company deals with key interactions - or "moments of truth" - such as policy amendments, cancellations, renewals and claims.

Related data that can be used to give a guide to how these are being carried out includes the proportion of interactions which are resolved in a single phone call. The more calls required, the less satisfied the customer will be.

Another area measured is how long associated maintenance work arising from claims takes to be completed. By measuring this, RSA can identify which of its suppliers aren't performing as well as they could.

In the future, RSA plans to expand the technology into product research and to further improve customer engagement.

Meanwhile, UK-based DIY retail chain B&Q is using the SPSS technology to cut down stock loss which has been created by internal and external theft and process errors.

Using data mining to find out the most vulnerable product lines to theft or damage means action can be taken to better protect them while they are moved between locations or stored.

The technology can also be used to flag up process issues that can lead to other kinds of fraud, such as an operator making excessive refunds on a till.

Richard Davies, national investigations manager at B&Q, said: "The deterrent value of presenting the data [to staff] is really high."

In the future, the team plans to link the SPSS technology to the shop CCTV systems while creating a library of fraud transactions is currently taking place to improve predictive modelling work.