Though recent economic news about retail sales has been upbeat, retailing continues to be a fiercely competitive environment. With nearly 22 square feet of retail space for every man, woman and child in the country, the gains of one retailer usually come at the loss of another. Indeed there have been nearly 150,000 retail bankruptcies since 1990. As a result, retailers who have always been intensely focused on improving return on investment for assets such as real estate and inventory are now turning to their information assets for competitive advantage. A well-designed data warehouse or data mart with effective decision support/query tools can answer questions for retailers that were previously unanswerable. Retailers are tapping into their transaction data to better plan sales promotions, improve merchandise assortments by store, improve vendor tracking, streamline budgeting and find new store locations. Some retailers are tying their risk management applications into the item and transactional data to reduce employee theft and shopper fraud. Successful data warehouse efforts have resulted in increased sales, improved productivity, decreased inventories and, lately, enhanced customer loyalty.

Strategic customer marketing is now at the forefront of most retail data warehouse strategies. The old retail axiom of having the right product at the right place at the right time is now being amended to include "for the right customer." Recognizing that customer equity is one of the only ways to sustain true competitive advantage, retailers are intensifying their efforts to identify, satisfy, retain and maximize the value of their best customers.

For the retail data warehouse, customer loyalty programs mean linking customer information to transaction data to glean knowledge about customer purchase histories, shopping preferences, motivations and triggers and leveraging that knowledge throughout the organization to make integrated customer-centric business decisions. Retailers can then not only determine whom to sell to but whom to listen to, whose patterns to track and whose behavior is predictive.

At a micro-marketing level those particulars might include, for example, customers who have purchased a suit within the last two weeks and who, therefore, might respond to a special promotion on ties or dress shirts. Or, a store's (or even a department's) largest spenders could be sent invitations for a special preview of a new merchandise shipment or early admission to a sale. It can show store managers that customers who purchase lingerie are bypassing the cosmetics counter, an indication that relocation of that department should be considered. Or maybe those shoppers need better inducements (such as a coupon) to help them make the connection next time they shop.

While data warehousing holds many benefits for the retail environment, it also has its unique challenges. Data preparation continues to be difficult. Populating the retail data warehouse starts in the store at the point-of-sale registers ­ not the most conducive environment for detailed data collection. And though item-level information and pricing are an accepted part of the retail transaction, harried consumers have very little patience for giving personal information such as an address or telephone number. As a result, retailers have to focus on other data collection techniques. Collecting information manually from checks is time-consuming but at least gets the customer on the mailing list for ZIP code analysis. Department stores have traditionally used their in-house credit cards as data collection devices. Now service providers exist for helping retailers backfill customer demographic information from third-party credit cards. Loyalty cards, however, with an incentive program are proving the best method for collecting customer information once and then motivating the customer to use it again and again with each new purchase transaction, thereby automatically attaching the customer information with the sale.

The sheer volume of transactional information is also a challenge for any retail data warehouse initiative. It's not unusual for a mid-size, 100-store specialty hard-lines retailer, such as an arts and crafts chain, to have upward of 100,000 SKUs and process nearly 30,000 sales a day. Department stores have closer to 1,000,000 SKUs to track on sales transactions. Add variables such as seasonal sales, special promotions, multi-sourced inventory and a commissioned sales staff, and you've got some big numbers to deal with over time.

Retailers are learning that the best approach for taking on a data warehouse project is the "crawl, walk, run" approach. In other words, don't build the ultimate data warehouse from the start. Instead, begin with a few business initiatives such as inventory reduction or margin improvement and add others slowly over time.

This may mean starting with a prototype solution from a known data warehouse solution provider. This will help you develop your data source and architecture, as well as your data cleansing operations. Using an off-the-shelf solution will help the corporate culture change from a report-driven process to a query-based process.

Don't make it an IT-only project. Implement data warehousing as a process with a tangible and measurable business solution in mind. Start with some "low-hanging fruit" initiatives with fast sizable paybacks to gain support for the rest of the data warehousing project.

Design for flexibility. Once smart users get answers to their initial queries more questions will be raised.

Invest in scalable hardware. Information-hungry merchants will find more and more creative uses for the data warehouse and will want more and more data available. Therefore, cost effective modular systems that can scale up to very large enterprise computing platforms and accommodate multiple generations of processors are critical to handling long-range requirements. For example, being able to add processing elements or additional storage to the system rather than perform an expensive and business disruptive "box swap" will save money in the long run. With analysts predicting that over 90 percent of all enterprise installations will include both UNIX and Windows NT by the Year 2000, the system's ability to provide high levels of interoperability between these operating systems and to integrate newer applications within the same system will also be key.

With the need for retailers to compete in the world marketplace today and their dependency on computer systems to analyze and run their operations, high availability and near-continuous operation as well as surviving catastrophic failure is critical. A computing architecture that allows enterprises to construct various levels of infrastructure availability is paramount to staying competitive. This includes high-availability system features built directly into individual servers, operating systems and storage subsystems. For even higher levels of system availability, clustering technology which allows multiple systems to share information in a near-continuous environment, should be considered. Replicating information across a company's campus environment can ensure a major form of information protection.

Don't underestimate the need for data preparation. Inventory, sales and vendor statistics are usually the easiest data to get a handle on. Getting accurate customer information will be time-consuming ­ but the payoff is large.

Improved product distribution, inventory flow, promotion planning and customer loyalty are vital to retail success. Though transaction management systems have always been the core technology of the retail enterprise, if handled properly, data warehousing technology will open up new opportunities for a stronger competitive advantage.

Building a Data Warehouse: One Customer at a Time

Premier retailers have a long and proud history of serving their customers and know that understanding customers' needs and preferences is the first step in moving toward 1-to-1 customer relationship management.

The vast quantity of customer data, from every dimension of the retail experience, has driven database designers and technology professionals toward using summarization or aggregation as the starting point for building data warehouses. But in today's marketplace, this approach won't allow you to deliver the individualized attention customers crave and will put you at a competitive disadvantage right from the start.

If you start with summarized data, you can only obtain summary results. And with summary results you will never be able to analyze and execute individualized marketing programs.

Analytical application and infrastructure software now exists to build customer-centric data warehouses and data marts to enable analysis of detailed transaction data for each individual customer.

By beginning with transaction-level detail, tied to loyalty card and credit data sources, you can perform detailed analysis and begin marketing on a 1-to-1 basis. Exact transaction-level analysis is the first step in attaining the level of customer loyalty, profitability and satisfaction that you and your management desire.


John Thompson is the vice president of marketing at RTMS, a leading provider of CRM solutions for 1-to-1 marketing. Thompson has over 15 years of experience spanning all major technology management functions for software organizations. His technology expertise includes knowledge discovery, decision support, data warehousing and database systems. He can be reached at (414) 650-8228 or jthompson@rtms.com.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access