Since its first store opened in 1933, grocery retailer Haggen, Inc. has been consistently heralded by customers and industry analysts alike for its innovations. From its position as the first store in the Northwest region of the U.S. to offer a self-service meat counter to becoming the first grocery chain in the U.S. to house an in-store Starbucks Coffee, company officials never stop asking themselves "what if?" and "why not?" when working to improve customers' experiences and planning the organization's future.

In 2006, Haggen decided to implement a data warehouse from Teradata to help answer these questions and secure the company's position as an innovative industry leader. The results have exceeded expectations and "changed the language" of the entire company for the better, says Haggen Vice President and CIO Harrison Lewis.

"I chose Teradata because I look at the whole area of business intelligence and data warehousing from the standpoint that it is not a destination, it's a journey," says Haggen Vice President and CIO Harrison Lewis.

Fresh food, stale data
When Lewis joined the Haggen team in September 2005, he was already known in the retail world as a process innovator and results-achieving visionary. Though Haggen was one of the smallest companies Lewis had ever worked for, he says its leaders had big goals and expansive ideas about the strategic benefits of data integration that belied the organization's size.

"One of the first things I was asked to do was to find a business intelligence [BI] and data warehouse solution," Lewis recalls. The decision to move to an enterprise data warehouse (EDW) and BI model was based not on hard return-on-investment (ROI) calculations but on a "shared vision" from company leaders that having timely, integrated information was incalculably valuable to the organization's future.

Before looking at specific tools or solutions, Lewis cataloged the company's data needs and tallied its informational strengths and weaknesses. To its benefit, Haggen had always intrinsically valued data and kept large amounts of transaction and systems data on backup tapes and other storage devices. However, officials didn't know how to best leverage that stockpile of data to their advantage. Additionally, the company's few information systems were siloed and their reporting options minimal.

Equally troubling, Lewis says, was that access to operational data often trailed seven to 10 days behind actuals. For a grocery store chain that emphasized its fresh and prepared foods, relying on stale data diminished daily operations capabilities, stagnated innovation and hindered managers' ability to "retool and recalibrate" their offerings for optimal outcomes.

Lewis says he quickly realized that an EDW could help meet the company's information needs and create countless new opportunities for growth and improvement. "I looked at the organization, and I said our needs are as significant, or even more, than many larger companies out there," he says. "So it was appropriate for us to look to the best in class to deliver on our business requirements."

A well-balanced solution
After an exhaustive search, Lewis selected Teradata to create a retail-specific solution ideally suited to the company and its specialized needs. "I chose Teradata because I look at the whole area of business intelligence and data warehousing from the standpoint that it is not a destination, it's a journey," says Lewis. "And as the organization grows and improves its understanding, we need to have an infrastructure in place that is flexible enough to accommodate that growth.

"I also chose Teradata because they were interested in our business, interested in what we're about, and they saw that there was a value proposition for them beyond just making a sale."

The road to integration
Lewis admits the implementation process was not without its bumps but says the seven-month project was the smoothest he has ever been involved with.

The biggest challenge came when business stakeholders challenged IT leaders' plan to load only six months of historical data into the data warehouse for the initial rollout, arguing they needed a full year's data to make necessary comparisons. IT leaders agreed to load the additional data even though the decision added two months to the project schedule. In the end, Lewis says, "it was the right thing to do."

Another snag came when company leaders found instances in which they didn't have the data they needed to appropriately calculate costs between stores or to compare costs over time. To combat the problem, developers created a custom algorithm to assign a "last weighted average cost" to items and processes.

enlarge

Though the organization had doubters and naysayers, company leaders eventually won over nearly everybody by involving them in the key performance indicators (KPI) development process and letting them see firsthand how new and better information benefited their jobs and the entire organization, Lewis says. He attributes much of the venture's success to the Teradata Professional Services consultants who helped guide the process to completion.

"They had a very critical role," Lewis says. "They did a great job in terms of understanding our strategy, what we're trying to create, and then architecting the database in a way that allowed us to accommodate the granular data and really insisting on this from a very early stage."

Fact-based foundation
Since the initial system rollout in April 2007, Haggen has moved from an organization based extensively on theories to one based on facts, Lewis says. For example, store managers can now quickly and accurately determine what additional food items are most often added to the shopping cart when a customer is buying a turkey, a head of lettuce or any other item. Managers can then use that information to make appropriate product placement and marketing decisions.

"I think the POS [point of sale] transaction data is the real jewel for our organization, and it goes beyond looking at what products are selling," says Lewis. "We have visibility on so many things: What specific check lane is a customer going through? What's the time of the day? What kind of tender method is being used? The data is so rich and so unique to the retailer, and there is so much to be harvested from that data."

In the future, Haggen officials plan to harness EDW information to implement a price optimization system, and IT staff members will continue to make system enhancements. Additionally, Lewis says, Haggen employees are constantly finding new and interesting ways to use the data.

"I know the results have exceeded my experience and my expectations," Lewis says. "And I think they have exceeded the entire organization's expectations as well." T

Behind the solution: Haggen, Inc. Database: Teradata Database V2R6.1 Users: 125 (25 concurrent) Data model: 3rd Normal Form Operating system: Microsoft Windows Storage: 3.9TB Teradata Utilities: Teradata Tools and Utilities 8.1: FastLoad, MultiLoad, and Teradata Manager; Teradata Utility Pack: Microsoft .NET Data Provider and OLE DB Provider for Teradata, Teradata Administrator, Teradata Basic Query Language (BETQ), Teradata Call Level Interface (CLI), Teradata MultiTool, and Teradata SQL Assistant Tools/applications: Products from GoldenGate Software, Manthan Systems and Microsoft

Lori Janies writes about business and technology for various publications.

Teradata Magazine-December 2008