Retailers are seeking to ignite new levels of ingenuity among category managers and other decision makers by providing them with analytical tools linked to massive databases called data warehouses.
While companies are now beginning to explore ways to use data warehouses to enhance micromarketing and micromerchandising initiatives, the thrust is primarily in category management.
Through simple and intuitive graphic-user interfaces, category managers can query these databases to get a quick snapshot of a category's performance overall or pose a
series of questions to gain insight into the finer details, such as promotional lift at the store level and intracategory cannibalization.
At Price Chopper Supermarkets, Schenectady, N.Y., for example, category managers are urged to dig deeply into the company's data warehouses and ask the important and not-so-obvious questions.
"I want to bring the curiosity of these people to the fullest," said Wayne Barton, director of category management at Price Chopper.
"The best question is really a series of questions driven by the data," Barton said. "You can ask, 'What are my sales?' and it shows you. Then you can ask, 'Why is Zone 1 up?' and it shows you, then, 'Why isn't Zone 3 up?'
"Now you can say, 'Compare Zone 1 with Zone 3,' a question you weren't thinking of when you started," but which proves to be key to understanding a category's performance.
The ability to pose such questions in a free-form manner is enabling category managers to sharpen their decision-making skills and implement strategies based on precise and timely data.
At Schnuck Markets, St. Louis, an effective data warehouse can provide a type of "artificial intelligence," said Bob Drury, vice president, management information systems.
"A data warehouse can assist in answering questions you've not asked yet, and can find patterns in the information that are helpful to the user," Drury said.
"More and more that kind of 'artificial intelligence' has to be factored into the design because we can provide enormous amounts of information, but unless it's actionable, we haven't accomplished very much."
Schnuck has been giving category managers access to point-of-sale and other information through a database application introduced 10 years ago. Though the system works well, Drury said it will soon be replaced by a newer data warehouse loaded with daily POS data.
"We need to have a more microview of the world, so we know the differences between a Tuesday and a Friday, or the day before the Fourth of July," Drury said.
At Giant Food, Landover, Md., which just completed loading two years' worth of item-movement history into its data warehouse, the system is used regularly by nine category managers in grocery, bakery, nonfoods and health and beauty care.
John Clutts, Giant's director of efficient consumer response, said category managers can drill into the data warehouse to better understand the impact of promotions, for example.
"If I promote a given product, not only can I find out what the lift on that product is, but what is the cannibalization across the category as a whole, or the segment, or the subcategory," Clutts said.
"Our next major phase is to load market data into the system," he said, noting Giant plans to do so by year-end or the first quarter of 1997.
"Our intent is to have seven geographies of market data that we can marry with our internal data, so we can develop a viewpoint of how Giant is doing -- internally and vs. the market. We'll be able to look at shares and at growth trends across a market," Clutts added.
Another strategic business area retailers are beginning to evaluate for data-warehousing applications is customer relationship marketing, said Walter Deacon, partner in the Management Horizons consulting division of Price Waterhouse LLC, New York.
"Understanding the buying habits of each customer and being able to tailor coupons or product offerings to a particular consumer is an excellent use of data-warehousing technology," Deacon noted.
Deacon said the price-performance ratio of the large database technology continues to improve, and will fuel more data- warehousing projects. For the time being, however, loading the data warehouse continues to be the most difficult challenge.
"A lot of legacy systems used to populate a data warehouse often contain data that's been massaged [and] summarized, and over time it becomes less useful and less accurate," he said.
"The data cleansing needed to disentangle all those changes over time has been a very significant job in every data-warehouse project. A lot of it is 'pick and shovel' work," Deacon said.
No chain appreciates that characterization more than Giant Food.
"Like many people, we underestimated the amount of work involved in validating the data" going into the warehouse, Clutts said. "We were pulling the information from a whole variety of disparate sources and you don't want to just load it. You want to make sure it's correct."
Clutts said Giant paid careful attention verifying the integrity of deal data. "We found that deals in certain categories would often be applied differently so we had to develop new rule sets" particularly for large, complex product categories such as soft drinks.
Industry observers believe the greatest opportunities to exploit data-warehousing technology lie in its potential to drive operational systems, and the more ambitious companies are beginning to recognize the value of such a "closed loop" approach.
"It gets to the point when you realize the quality of data is good and you say, 'I can use this to forecast' and once you do forecasting and make this data available to suppliers, you can do replenishment based on that," said Jerry Singh, managing partner, The Partnering Group, Cincinnati.
"There's a myth out there that data warehousing is for decision support only," Singh added.
Deacon agreed: "Instead of just getting hard-copy reports out of the warehouse and using them manually, you see the loop getting closed and a lot of that decision-support data driving the transactions," Deacon said.
