Information Age: News, analysis & insight for IT & business leaders

There is something particularly disorientating about walking into a cavernous DIY superstore. There are typically few, if any, staff ready to direct customers to the materials they are looking for, and the couple of dozen overhead notice boards can only go so far in pinpointing specific products. That is especially true when the outlet is catering for professional tradespeople as well as serious home improvement enthusiasts.

Until recently, Wickes, one of the UK's largest DIY retailers, used little more than gut feel to plan the layout of its stores. Not only did that induce frustration in its customers as they wandered the aisles, but it meant that the company was not putting popular goods in prominent places and grouping goods together that might trigger related purchases. The idea behind Wickes has always been simplicity - it sells own-brand materials, at lower costs than most of its rivals, to DIY devotees and the construction industry.

The no-nonsense attitude appeals to both groups. So when the company investigated the prospect of increasing its revenues by actively planning its store layouts, it did so with some caution, unwilling to upset its winning formula.

Dave Cooper, space planning manager at Wickes, says he opted to group the materials in the company's 172 nationwide stores throughout the UK into clusters that would reflect the way customers actually shop in the stores and thereby providing insight into how to reallocate selling space more appropriately.

To understand those buying patterns, whilst maximising the return of the space available, Wickes turned to software from business intelligence specialist SAS, implementing SAS 9.1.

This relates the layout of goods to basket analysis and demographic data, as well as overall store sales statistics. "[The project] will ultimately provide increased revenue for individual stores and the group as a whole," Cooper believes.

Profiling stores by sales patterns, rather than by region or size, gives each cluster of stores a 'personality', and allows Cooper to tailor the shopping experience to suit customers. So the layout of the Tottenham store might be different, say, from a store in Bracknell where demand patterns differ. "We're only really scratching the surface at the moment," says Cooper. "The system is built to do [product] clustering at the moment, but we can tune it to do other things."

Access to the system is provided to brand merchandising employees who can export SAS information as web pages or Acrobat files for wider distribution. Moreover, familiarity with Microsoft Access or Excel makes the software easy to grasp, says Cooper.

Using the clustering concept to better understand the macro buying patterns has spurred Cooper on to get to grips with the micro picture - the way products on the shelves are distributed. This data will eventually be combined with detailed basket analysis and influence product placement. By analysing buying habits, Cooper predicts Wickes will be able to trigger additional customers' purchases - and make it easier for them to navigate the stores.