The UK's major retailers are investing millions of pounds in technology to track what their customers purchase, where they purchase it and when. Such information is a vital commodity in the fierce battle for market share. Tracking exactly what you put in your trolley each week is big business.
Leading the way are the big three retailers: Safeway, Sainsbury and Tesco. All have implemented data warehousing technology to enable them to collate and analyse details of shopping purchases throughout the country. A data warehouse is a vast store of information which has been specifically designed to allow analysis and extraction of trends.
Information is gathered via loyalty cards - an increasingly common device in the retail sector. Such cards identify the shopper and enable their purchases to be recorded every time they visit. Customers are rewarded for using their cards through points giving discounts and other incentives.
"There will be a record of the card, all the items bought, how you paid, how much change you were given, the time of day, the cashier, and which checkout you used," says John Whitworth, Safeway's systems development project manager. "Basically, every piece of information about your time at the checkout is stored." Over time, information gathered in this way creates a clear picture of customer buying habits.
The amount of information involved is staggering. Safeway installed a data warehouse for its loyalty card scheme, ABC, about 18 months ago. During that time, the company has built a collection of around 7.5 billion records from its six million cardholders. Safeway expects this to grow to between 10 and 12 billion records after two years.
Tesco's Clubcard system, launched in February 1995, has more than five million active members and is building a similar data store for the retailer.
Information collected at Safeway checkouts is stored in computers in each of the company's 400 stores, says Whitworth. Every night, this information is downloaded to the company's headquarters in Hayes, Middlesex, where it is stored in a data warehouse on a dedicated IBM S/390 mainframe computer. Once stored, sophisticated tools are used to analyse the data and extract information on buying habits and trends. Despite the huge amount of information involved, queries can be answered quite quickly.
"For example, to list all the customers who have bought a product on more than one occasion over the past year would take about an hour," says Whitworth. "This would involve checking five to six billion records."
A more common query may involve looking at a four-week period, which would be completed in less than five minutes. Such speeds are possible because the data warehouse's computer is equipped with five processors working in parallel. This means different parts of the data warehouse can be accessed simultaneously, thus saving time.
It is this capability to analyse vast quantities of data which has led to major retailers embracing the technology so enthusiastically. The individual transactions alone are of little interest, but given enough of them, the trends they can show are invaluable.
Product associations are one area which has particularly excited retailers. Finding links between seemingly unrelated items - beer and nappies, for example - might allow a store to increase sales by moving the items closer together or offering discounts when purchased as a pair. This type of correlation would not have been possible without a data warehouse.
"Our store cards tell us the trading patterns and the life cycles of our shoppers," says Mike Fisher, Safeway's database marketing controller. "What we are trying to do is develop a relationship with the customer and build a sense of loyalty."
A better relationship equals better sales and retailers are constantly looking to their data warehouses to suggest new ways this can be achieved. Sales data is often compared with information from other sources, such as market research companies, to better understand what makes shoppers tick. Customers are categorised into life stages - single people, married couples, families and retired people - and their theoretical buying habits are compared with those seen in store.
"If the average shopper for a given life stage buys a total of X products each week and she is only buying Y from us, we can identify what she's not buying and encourage her to buy from us," says Fisher. This is marketing at a micro level.
Using computer technology to keep such close tabs on shoppers' purchasing habits has prompted questions about privacy.
"People ought to be very aware of what information is being stored about them," says Harriet Hall, spokeswoman for the National Consumer Council. "It's easy to say that it doesn't matter because it's only your groceries, but it's like having your till receipt linked to your name and address. If you are concerned about your privacy, this is a way that privacy is being leaked into the outside world."
This view is not shared by the retailers. "I don't believe that's an issue," says Fisher. "The feedback we get from customers is generally positive. The reason we hold data on people is so that we can make the store more applicable to them and give them what they want. It's of benefit to the customer that we understand how they shop."
There's every indication that data warehousing techniques will remain a major tool for the retail sector. "Once you get into it, it's very difficult to get out," Fisher says. "The company will become used to having this level of information about what people do. Having better information about your market and your customers enables you to make better decisions about the future - it will always justify itself."
Despite their obvious benefits, it is difficult to judge whether retailers are getting value for money from their data warehouses. None is willing to confirm how much has been invested, nor whether the sales gains achieved have justified the amount spent.
Thomas DePasquale, vice-president of data warehousing specialist Platinum Technology, remains sceptical. "I would challenge whether retailers are getting a payback from their loyalty card schemes," he says. "I think there tends to be a herd mentality in that if one installs such a scheme then the others feel as though they have to follow." He said data warehousing technology only worked when the company using it had a clear idea of what it was trying to achieve from the outset.
But one thing remains clear. The ability to probe into the shopping trolleys of customers and analyse their purchases will remain an attractive ability for major retailers. The data warehouse is here to stay