Business intelligence strategies for enterprise CIOsExecution: BI systems implementation strategies, best practices <<previous|next>>
Business intelligence (BI) for the midmarket
Operational BI entirely new landscape of analysis
By Michael Ybarra, Contributor
Things were a bit blurry at 1-800 Contacts Inc. Managers at the online contact lens retailer didn't have a very clear vision of how well customer service reps were doing their jobs until well after the fact, when they ran manual reports on weekly numbers and posted the spreadsheets on the wall.
"We didn't have a real comprehensive view of the call center," says Jim Hill, the data warehouse manager at the Draper, Utah-based company. "An agent wouldn't know there was an issue until the following week."
Contacts needed a better view of the business, so the company deployed a new business intelligence (BI) product that allowed call center agents to see their performance metrics -- closing ratios, average order size, calls in queue -- in real time.
"Each call center agent has the ability to see how they're performing," Hill says.
Business intelligence is nothing new for many midsized companies, but what's changing is the way firms are integrating the tools into their daily, minute-by-minute activities. This newest flavor is operational BI -- as opposed to strategic or tactical, such as launching a new product line or opening a new store.
"Operational business intelligence supports the day-to-day operations to generate revenue and increase profits," says Wayne Eckerson, research director at The Data Warehousing Institute (TDWI) in Hingham, Mass., who has surveyed companies on how they use BI. "Historically, that's not been the province of business intelligence; traditionally, BI has been used by analysts to do after-the-fact analysis. Operational BI opens up a whole new landscape and drives greater value from the investment."
A survey by TDWI found that 53% of 423 companies are deploying BI in an operational capacity. Of those, 35% say they are still exploring the concept; 27% are under development and 11% have recently deployed.
Another company having some success with operational BI is Eastern Mountain Sports Inc., an outdoor retailer based in Peterborough, N.H. A couple of years ago, the company deployed WebFocus, a BI product from Information Builders Inc., to create dashboards layered atop its midtier AS/400 platform in order to better manage inventory.
Then last year the company added a payroll dashboard that allows managers at its 80 stores to track staffing and sales in almost real time and make adjustments to put more workers on the floor at peak hours.
"We put up a set of graphics and management can see very quickly where staffing may be out of whack," CIO Jeff Neville says. "You can see hour by hour. We could have updated every 15 minutes. We chose a 24-hour cycle. We wanted floor managers on the floor, not slaving over reports. I didn't want them playing a video game. As part of their daily routine, they can review trends and make plans." Leveraging its earlier BI deployment, EMS was able to go live in six weeks with the new dashboard.
"With that information in place we could add different dashboards for very small money," Neville says. "For a very small investment, we're getting a very big payoff."
There are technical challenges. Operational BI requires a fundamentally different design and architecture; more frequent, intraday updates; and low-latency data.
Operational BI also requires a new way of doing business for some IT managers. The data warehouse can no longer be updated in a batch process during off hours. That means new architecture for data acquisition and delivery mechanisms and high-availability processes.
Queries must return in real time and reports must update dynamically, which requires capturing large volumes of data without degrading the performance of existing processes. That puts a premium on building resilient systems with good backup and recovery.
"When we created our data warehouse we processed records early in the day," Hill says. "We had to layer on an additional intraday load every 15 minutes. We had to write two sets of ETL [extract, transform, Load] processes and another layer of monitoring for SLAs [service-level agreements]. For the most part, the tool set is working well. It's a matter of taking the concepts and building a smaller version to refresh."
For a very small investment, we're getting a very big payoff.
Jeff Neville, CIO, Eastern Mountain Sports Inc.
Nevertheless, Hill says the biggest challenge wasn't technical but training: keeping the high-turnover workers up to date on the business logic used for incentives.
"The ultimate objective is to make the presentation of information so clear that it doesn't take a lot of understanding," Hill says. "It was a challenge because of the complexity of the incentives. We compare everyone working the same minutes, compute their scores so it's fair across the busy and quiet times of the day. We've achieved a good level of understanding of how they're performing today, but that link of how they're performing to earn their score isn't there. We're trying to fix that. All of the performance measures will stay the same -- closing ratio, average order size -- but we're looking at how to make the accumulation of scores much easier."
"We saw a lift of $50,000 a month in revenue; average order and closing size improved, while maintaining quality scores," Hill says. "We were able to increase closing ratio and average order size without sacrificing the customer experience."
Still, it's also possible to oversell BI and Hill says CIOs need to partner with the business to determine the system's optimal requirements.
"Do they really need the data real time, and what is the business benefit?" Hill says. "I don't think all data needs to be real time; there's a cost to it. You need the systems, the bandwidth. It will require an investment in hardware to hand the volume. You don't want to provide the data faster than the business will consume the data.
"Fifteen minutes is plenty fast. We don't want the agents to spend all their time checking their numbers."
Michael Ybarra is a former senior writer at CIO Decisions magazine. He is also the author of Washington Gone Crazy. Write to him at firstname.lastname@example.org.
13 May 2008
Disclaimer: Our Tips Exchange is a forum for you to share technical advice and expertise with your peers and to learn from other enterprise IT professionals. TechTarget provides the infrastructure to facilitate this sharing of information. However, we cannot guarantee the accuracy or validity of the material submitted. You agree that your use of the Ask The Expert services and your reliance on any questions, answers, information or other materials received through this Web site is at your own risk.