Magazine Article | June 1, 2003

Multiple Apps, One Database

Source: Innovative Retail Technologies

$700 million retail technology giant Ultimate Electronics is tearing down its data silos so applications can tap a single, relational mega-database.

Integrated Solutions For Retailers, June 2003

The less retail technology our customers encounter in our stores, the better. You wouldn't expect to hear those words from a retail CIO or director of technology development, would you? Not today, not in this check yourself out, swipe your own card, don't ask me - ask the kiosk retail environment, right? And you certainly wouldn't expect to hear those words from Alan Kessock, CIO, or Steve Connell, director of technology development, at high-end consumer technology retailer Ultimate Electronics (Thornton, CO). After all, these guys sell technology. But Ultimate Electronics isn't like Circuit City or Best Buy. The goal here is not to expedite the shopping experience and get the customer out the door. No, the goal here is to upsell. The goal is to extol the virtues of the $15,000 plasma screen - to add a warranty here or a complementary product there. These goals can't be achieved by technology alone. But in order to reach their goals, Ultimate customer service associates - both those on the sales floor and those taking customer service calls - must be equipped with retail technology that facilitates service of special customer needs.

How Does The Tech Store Buy Tech?
There came a point in 2000 when Ultimate realized a need for POS software that would drive detailed order management functionality to its store-level associates. But, purchasing decisions are not made lightly at Ultimate. A cross-functional purchasing implementation team exists at the company, consisting of IT and operations employees and senior executives. This team defines what systems are needed, how they should be designed, and how they will be proofed and put in place. In the early stages of systems analysis, Ultimate is careful to make sure the project is not handled entirely by the IT department. "We focus on the operational need, and at this point the IT department is simply a facilitator," says Connell.

Before buying any retail technology, the purchasing and implementation team at Ultimate asks, 'Is it buildable? Is it deployable? Is it maintainable?' If the answer is yes to all three, Ultimate looks at the overall cost to implement. Purchasing decisions are always made with margins in mind. "We are always trying to drive cost out of the business, so when we evaluate a new technology, we ask ourselves, 'where are we experiencing cost? How can a particular technology come in and help us leverage against those costs? How will it affect the customer?'" Connell explains that before a technology implementation, VP and C-level employees in the departments affected by the project are charged with determining exactly where their departmental budgets will be positively or negatively impacted. They are then held accountable by budgeting for any predicted savings prior to the implementation, and then required to measure and report on post-project performance.

"When we adopt a project, we obviously want to make sure we get a return, so we operate on a very stringent budget model," says Connell. Those budgets incorporate line items ranging from manpower to expenses and capitalized activity. With strict attention to budget, Ultimate can see the slightest improvements in efficiency and assess whether or not they can be attributed to a technology implementation.

Detailed budgeting and savings forecasting prior to a project has proven a wise approach and saved the company millions in unnecessary technology spending. When it recently revamped its WMS (warehouse management system), Ultimate toyed with the idea of building a brand-new addition onto its DC (distribution center). But the company found that new WMS software alone would help it use the existing floor space more efficiently, putting off the need to build. Similarly, as the company's growth rate expanded to 12 or more new stores per year, the technology team considered investing in project management and costing software. "The analysis showed that in fact it would not drive cost out of the process, so that project did not go forward," says Connell.

Not Just POS, But Order Management
But late in 2000, the decision to upgrade its POS/order management software was based on definite needs, so the project did move forward. With growth averaging 20% per year, there was a sense at Ultimate that with its current system, what the company terms a "hard stop of scalability" was just around the corner. So in 2000, the company began searching for a POS system that could drive the order management functionality Ultimate needed to all of its 58 stores using a thin client architecture. At that point, the retailer found that such a product didn't exist in as mature a state as it had hoped. Like many retail technology projects, this one was driven by a desire for improved customer service. Connell explains that his CSRs (customer service representatives) - which in Ultimate's case include POS operators, employees on the sales floor, and those taking customer calls - needed to be able to communicate things like purchase history, complementary product information, and service and warranty options to Ultimate customers. In order to do this, CSRs needed access to several kinds of data, and they needed that access on both traditional POS terminals and dedicated customer service stations. "With this kind of data available to any CSR at any terminal, we can market specific products to our customers," says Connell. "We can make suggestions that complement past purchases. We can sell them post-sale services, like warranties and service contracts. And we can do it regardless of mode or point of contact," he says. But making this kind of information accessible to CSRs is database intensive. Ultimate had a 14-year-old flat file architecture, which would make querying and reporting a daunting task.

Ultimate has since been working with retail software provider Tomax (Salt Lake City) to develop a single, relational database from which to pull data across thin client store terminals. The solution is based on Tomax' Oracle-based retail data model, which focuses on three areas critical to retailers - inventory management, labor management, and customer service. The philosophy behind the data model is that all the information necessary for retail operations should be stored in a single database and made accessible to multiple applications. Oracle calls this data structure "single-instance." The goal is to reduce integration complexity, eliminate data silos, and eliminate the overhead related to maintaining and supporting multiple databases.

Relational Database In Action
With stress testing of in its final stages, Connell reports on his company's results. Through a standard, single screen interface at any store terminal, CSRs at Ultimate can see all the warranty coverage a customer has on past purchases. With a service department that services every product the chain sells, this kind of data is important and is accessed frequently. The old way, which entailed gathering data on customers and their purchases at the time of the sale and filing the pertinent information electronically in its own database, was cumbersome enough. Compiling a whole view of customer purchase, service, and warranty information required hitting several databases and toggling between screen views, which was extremely time-consuming. The new system categorizes purchases and all information related to them in a single database, allowing CSRs to respond more quickly. "Customers are satisfied when things work, so we need to keep them working as long as possible. We can do that more easily with the new system," he says.

But the ease of data access at Ultimate has improved the retailer's inventory and sales visibility as well. "In addition to improved customer service, we can also chart changes in our sales and changes in consumer preferences. That was very difficult with the old system because we couldn't pull reports in a timely manner," says Connell. With the relational database of, Ultimate can access data and generate reports in real time. With its flat file system, the best it could do was to summarize old data and translate that into projected sales numbers. "The close out, reporting, and posting lag time with the old system might put us two or three days behind. With, we've also moved our payment settlement time up," says Connell. This is advantageous in a company that processes the majority of its transactions on credit. "When you come in to buy a $10,000 TV, you probably aren't paying with cash. You're probably putting it on some type of credit account. We've improved our settlement times with the new system from about five days to less than 24 hours," he says.

Ultimate also expects to decrease new employee training time by 66%. "It's simply more intuitive and easy to run. It doesn't require as much keyboard function knowledge because it's more mouse driven," says Connell. With real-time product inventory data at the fingertips of CSRs, Ultimate is expecting a 75% reduction in the manpower it takes to move merchandise between locations, both DC-to-store and store-to-store. "Replenishment is being received in about 25% of the time it used to take us. We use UPCs (universal product codes) and we're able to scan at the store using one of four methods to determine what the product is - either our internal product code for it, the vendor product code, or one of two UPCs. Tomax' makes this possible," says Connell.

Departmental Data Integration
With POS, decision support, and budgeting systems live on the Oracle database, Ultimate has now begun to integrate data from its HRMS (human resource management system). "The beauty of the relational database is that it presents one version of the truth. People are now making decisions based on one version of knowledge instead of guesses and estimates," says Connell. "Before, we couldn't have that one true, integrated vision of all these systems." Ultimate has learned the case for a relational database on a customer service level is a strong one. CSRs are equipped with just enough technology and data to nurture the sale, while the only technology the customer interacts with is that which he's purchasing.