To print: Click here or Select File and then Print from your browser's menu

This story was printed from silicon.com, located at http://www.silicon.com/

Story URL: http://software.silicon.com/applications/0,39024653,39131403,00.htm

Sainsbury's scrutinises buying habits to woo customers
Case study: Supermarket improves data analysis to give its customers what they want

By David Braue

Published: Friday 24 June 2005

Sainsbury's, which has around 16 per cent of the UK's £110bn grocery market, has been steadily losing market share to competitors Tesco and Asda, a Wal-Mart-owned business that knocked Sainsbury's out of the number-two position in 2003 and hasn't looked back.

Since his installation last year, chief executive Justin King has pursued an aggressive plan to get Sainsbury's back on track. Conventional cost-cutting strategies, such as trimming prices and closing the least profitable of its approximately 500 outlets, have been complemented by a focus on areas such as revenue generation through store reconfiguration and targeted marketing.

Such changes require good intelligence about customer buying patterns but this had been one of the company's weak points in the past: poor analysis processes meant marketers could be working with data that was 14 months old.

Alex Fovargue, head of customer analytics marketing at Sainsbury's, recently told 300 attendees at the Teradata Universe conference in Melbourne: "You would typically ask the IT department to extract data, and they would come back after six weeks with something that wasn't complete.

"We were in this continual cycle of not getting data when we needed it, and having to write campaigns when we didn't know what was going on. We were very risk averse and were quite happy making whatever money we were making; there was no view of how much money could be made out of direct marketing. We were pretty much operating on gut feel."

Directed to improve Sainsbury's direct marketing (DM) techniques, Fovargue bolstered the customer data analysis team from five people to 16, and set them on the task of finding new ways to capitalise on the company's data warehouse - a five terabyte database containing months' worth of information about purchases made by the company's eight million customers. Transactional details are tied to specific customers through the company's Nectar loyalty programme, producing a goldmine of information about buying habits.

Better data analysis quickly showed Sainsbury's how ineffective its traditional mass-mailing approaches were - where large numbers of coupons were widely distributed in an attempt to get customers through its doors. Rather than buying more, many customers would cherry-pick the specials and go to its competitors for other items. This meant many advertising campaigns were running at a loss - a relative return on investment (ROI) of 0.7, according to Fovargue, when ranked on a scale where 1.0 represents breaking even.

Over the past three years, however, a concerted focus on timely data analysis and relevant marketing has helped Fovargue's analytical team design far more effective DM campaigns based on customers' actual purchasing habits. The ROI index has jumped to 1.6, boosting annual direct marketing revenues from £35m to more than £400m.

"Direct mailing allows us to be really highly targeted and to respond to those issues that are important to a small group of customers in a way that you couldn't do in a store environment," he explained.

Sainsbury's data warehouse is based on a Teradata data warehouse, business intelligence tools from MicroStrategy and SAS Institute, the Teradata TCRM customer relationship management suite, and Microsoft Access and Excel.

Taken together, these systems have provided the horsepower to let the team think far more laterally. Campaigns run using the data range from the conventional - promoting new in-store brands based on past purchases, for example - to the experimental, such as one campaign in which Nectar customers were sent birthday cards offering discounts on frequently-purchased items.

The retailer has also been able to add customer survey results into the mix alongside transactional data. This capability has proved exceptionally useful in shaping business improvement strategies by, for example, enabling stores to take note of customer complaints, to act on them and then entice customers back with specific, targeted coupons.

In another campaign designed to increase the value of customers' shopping baskets, Sainsbury's analysed purchases and identified the product category from which each customer purchased most frequently. A coupon for that category would then be sent, along with five other coupons for areas in which it was hoping to boost sales - to encourage customers to buy other types of products. The response rate was 26 per cent, which Fovargue called "a tremendous amount in retail".

"CRM has allowed us to operationalise all of our processes, and to produce standard reports which say how each campaign has done," he said. "You can look at real micro-levels in the data, and we wouldn't be able to do that if we didn't have the customer data linked into the data warehouse. "Now that we know how individual campaigns are performing, we can start rewarding behaviour and decide what is the correct mix of campaigns. We're trying to get people into the habit of shopping Sainsbury's."

David Braue writes for ZDNet Australia