CANTON, Mass. -- The server room at Casual Male's headquarters here looks almost empty.
Getting to the few machines in the back corner of the room is like walking across an empty "football field," according to Dennis Hernreich, chief operating officer of the retailer, which specializes in clothes for big and tall men. That's because the company upgraded its infrastructure about five years ago, getting rid of its huge old mainframes in favor of modernized -- and space-efficient -- systems. The new systems now run Casual Male's specialized retail applications and support multi-channel sales through a Web site, a major catalog business, and 520 retail stores throughout the U.S., U.K. and Canada. But under the hood of these servers, something is missing. There's no data warehouse, no sophisticated business intelligence (BI) software -- and no dedicated IT team coding reports. That's because Hernreich has what he calls a "secret weapon" -- hosted, on-demand BI.
Increasingly, companies like Casual Male are turning to on-demand software for their BI needs, drawn by low start-up costs and minimal IT investment.
Evaluating business intelligence options
Back in 2003, Casual Male discovered that the reporting and analysis capabilities of eCometry, the new application it had implemented to facilitate its retail and direct sales business, were inadequate. Casual Male needed a better way to plan inventory, understand customer buying behavior, and track the results of catalog mailings. The company looked at traditional BI software but balked at the price and implementation requirements.
"We wanted better information quickly," Hernreich said. "We didn't want to spend $1 million and we didn't have the people to build [a new BI system] because they were all engaged on other projects."
So when Waltham, Mass.-based Oco Inc. claimed it could develop, host and deploy an on-demand BI system for Casual Male in only six weeks, Hernreich figured that even if the claim was as much as 25% off, he would still have an operational BI system in less than two months. So the retailer took the bait, signing on with Oco.
First, the company wanted better reporting capabilities for its direct catalog and Web sales. So Oco set up the platform -- a hosted warehouse that integrates data from eCometry and a few other minor source systems via a nightly batch upload process. The warehouse includes data quality checks, which append incomplete fields or flag data discrepancies. With more than 40,000 active stock-keeping units (SKUs) -- or kinds of merchandise -- in its system, this was no small task. On top of the data warehouse, Oco layered its BI reporting and key performance indicator (KPI) module, which Casual Male accesses over the Web. As promised, the system was up and running in about six weeks, Hernreich said.
On-demand business intelligence in action
At first, the merchandisers and planners, who coordinate planning, inventory and buying for retail stores, were leery. Some thought they'd prefer the old way of doing things, through manual analysis. But after seeing the reports, they use the system "all day long," Hernreich said, relying on it to operate the business.
Executives also access the reports -- Hernreich logs in every morning. Before, he had to pore through 150 reports that didn't quite have what he needed. Now, Oco has helped the company develop 23 reports that have exactly the data needed to run the business.
Once the initial implementation was in place, Oco went on to develop a feature that enables authorized users to create and send out purchase orders through the system while they have current inventory, buying patterns and other relevant data at their fingertips. The order is routed to the vendor and back to Casual Male's finance systems.
Last year, Oco revamped the existing "spreadsheets and guessing" planning process with a forecasting module and created a task management system -- to better automate communications to retail store managers. It's all had a measurable financial impact, improving profit margins by 3.5% over two years.
"By having this information at our fingertips, our business began to improve," Hernreich said. "Our gross margins improved too."
Buying on-demand business intelligence
Success stories like these have not gone unnoticed, and the on-demand BI market is heating up, according to John Hagerty, vice president of research with Boston-based AMR Research. While vendors have different approaches and specialties, he said, companies such as Bellevue, Wash.-based SeaTab Software Inc. and San Mateo, Calif.-based LucidEra Inc. are also offering on-demand BI services. Last year, the Cary, N.C.-based SAS Institute released five on-demand BI modules. In addition, Ottawa-based Cognos Inc. recently acquired Celequest, a small BI vendor that offers dashboards via a Software-as-a-Service (SaaS) model, and Paris-based Business Objects SA purchased NSite, an SaaS platform provider -- further proof that big BI vendors see benefits to the new deployment style. In fact, Hagerty said, a 2006 AMR spending survey found that 17% to 22% of companies would prefer to buy their BI or performance management technology in an on-demand model.
"People see the bottlenecks in IT, the lack of trained resources to put a strong BI infrastructure in place, and just the general cost and time to do it," Hagerty said. "So a lot of companies are wondering how they can short-circuit that whole process."
There are a few key considerations for those evaluating on-demand BI. Many companies, especially larger ones, aren't comfortable letting data outside of the firewall. There are also technical integration requirements, since useful BI all starts with current and accurate information, Hagerty noted. And because many companies look to these types of solutions to solve tactical business problems, they need to be sure that an on-demand system will fit in with the business, culture, process and existing infrastructure.