FRONT PAGE STORY: Operational BI Drives DW Renaissance

May 2006

By Joe McKendrick

For Daune Kramer, senior data engineer at Manheim Auctions, inspiration on where to go next with his company’s massive and fast-growing data warehouse came from the general manager of a major league baseball team. What lessons could a sports franchise provide the world’s largest wholesale auto auction provider about business intelligence?
     For starters, formulating new and innovative analytic approaches for looking at all the information Manheim has been gathering--on not only the types of vehicles moving through its auctions, but also where they’re coming from, who’s buying them, and where and why particular vehicles are being bought. Kramer heard Billy Beane, general manager of major league baseball's Oakland A's, and the subject of Michael Lewis’ best-selling book, Moneyball, explain his non-traditional approach to operating a major league baseball team at a recent SPSS user conference. Beane built a winning team with one of the smallest budgets in the league by relying on statistical analysis for recruiting new players, rather than conventional baseball practices.
     “The A’s said, ‘Let’s throw out everything we know about baseball when we’re recruiting a player,’ ” Kramer explained to DBTA. “What’s important? Find statistical correlations between what’s successful in the game of baseball and a player--things that no one else is looking at.”

As forward-looking companies such as Manheim Auctions show, this is not your father’s data warehouse. Kramer has developed more than 65 cubes that look at a range of operational and historical data--from financials to vehicle condition--from a 27-million-record data warehouse. “The Oakland Athletics example inspired us to say, ‘We’re going to throw out everything we know,’ ” Kramer said. Decision-makers and partners are encouraged to use the data to drill deep and discover previously unseen patterns, Kramer explained. “We’re digging down and looking at all these aspects of the business that we keep in the data warehouse, looking to find things we don’t know yet, such as trends specific to a region of the country.”
     The company's data comes from many sources, including auction operations, financial lending, Web analytics, retail operations, remarketing solutions and human resources. Manheim is providing its trading partners--auto dealers and leasing companies--access to analysis cubes. Such capabilities enable Manheim to position itself as a valuable resource to members of its network of 200,000 dealers, who can look at demand levels through more detailed market analysis. “Dealers can keep track of our inventory, and analyze availability by region. We have most of the wholesale car market, and we know what’s getting bought where. For example, you may need more Ford Tauruses in your region. Our data warehouse can tell you that we know, in your region, people are buying a lot of Ford Tauruses, but that you don’t have enough on your lot to meet what we think your demand will be.”

Perfect Storm
This new generation of analytic data warehouses, such as that maintained by Manheim, represents a “perfect storm” of greater user expectations for right-time, analytical data, and increasingly standardized and inexpensive technology, said Claudia Imhoff, president of Intelligent Solutions, a consulting group, in an interview with DBTA.
     Historical and operational data has become an enterprise-wide resource, and companies of all sizes are taking a platform approach to business intelligence and analytics. The data warehouse supports not just historical analysis, but also right-time operational intelligence and predictive analysis as well. “Business intelligence is evolving to be far more than just looking at historical trends,” Karen Parrish, vice president of business intelligence for IBM, told DBTA. “In fact, the marriage between the historical trending information, and the transactional information that occurs minute by minute, hour by hour, sometimes second by second, is really what the world of business intelligence and the data warehouse is becoming.”
     In addition, “companies today are realizing the extreme value data warehouses hold for not only looking at historical data but, now, for predicting future outcomes,” said Martin DeBono, vice president of worldwide marketing for Insightful Corp. “By deploying predictive analytic solutions, organizations can now deliver to every decision- maker's desktop the ability to assess and prioritize alternative courses of action, and understand the potential risk associated with any decision made.”
     This all leads to an emerging data warehouse renaissance, said Imhoff. “Operational BI has ignited the whole market. We now have operational BI pushing closer and closer to a real-time --or right-time--business intelligence environment. We’re able now to integrate a lot of our business intelligence into the operations themselves, making it much more accessible and much more usable to just about everyone throughout the organization.” In addition, the rise of various compliance mandates--which require better information, if not a single view of the truth across enterprises--has helped fuel the data warehouse renaissance, she added.

Pervasive BI
Both business intelligence and data warehouse technology have been around for a while, and are pervasive at most data sites. A recent survey among members of the International Oracle Users Group (IOUG) conducted by Unisphere Research found that 63 percent now run some kind of business intelligence or analytics applications against their database environments. “We’re seeing a lot of interest in data warehouse business intelligence,” said Ari Kaplan, president of IOUG, in an interview with DBTA. “Until a couple of years ago, it was limited to C-level and top marketing executives, so only a handful of people were using business intelligence tools. Now, there’s been a proliferation of data, and a proliferation of the value of BI and predictive type tools. The number of end-users of BI and DW applications has grown tremendously.”
     Another survey, conducted among members of the Oracle Application Users Group (OAUG) by Unisphere Research, confirms that most enterprises (55 percent) either pull business performance analysis data off data warehouses or data marts. About the same amount, 58 percent, pull data from operational databases.

Temporary Quarters
For many companies, this enhanced ability to deliver operational analytics through a data warehouse environment has become a competitive weapon. Oakwood Worldwide, which provides temporary quarters to business travelers and insurance customers, has been able to deliver analytical reporting to its business partners using an Informatica data warehouse platform. “We’re using the technology platform to deliver the information in real time to our users, and extend that same information directly to our clients,” Vinnie Le, vice president of information systems for Oakwood, told DBTA.
     “One of our major services today is to be able to integrate and consume information in real time from clients’ systems, and share information in real time.” Information gathered and shared with partners includes number and size of units available, as well as pending vacancies at 200 locations worldwide. In the fast-moving relocation market, this information can change from hour to hour. “We needed to essentially automate the process of collaborating data and sharing data across systems in real-time mode, whether it’s instantaneous, hourly, daily, briefly, or monthly,” said Le. “Our next step now is to look at cost of sales, which is something that we didn’t have a way to get to the data. We’re able to tie all the pieces together, and look at profitability from the standpoint of how much money we spend on each client, by unit and by tenant, to see how we need to adjust our costs in order to increase our market.”
     Similarly, by opening its business intelligence environment to trading partners, Manheim has been better able to provide feedback on a two-way basis. This enhanced capability provides dealers and leasing companies “the ability to see how we’re doing with them, to either help us do a better job in supporting them, or make us aware of what’s important to them,” said Kramer.

Multiple Loads
Such unprecedented access requires changes within the data infrastructure, however. Moving toward operational business intelligence--which relies on extending access to a wider group of users, much of it on a near real-time basis--requires infrastructure changes, Imhoff cautions. “Once you go beyond the traditional analytics of long-term patterns and trends, and open business intelligence up to operations, your audience is going to expand dramatically. Now you’re talking about line-of-business managers, customer service reps, or anybody that has a need to understand what’s going on right now.”
     Additional challenges include the need for faster query performance than previously available to internal users, and more frequent loading of more extensive data. “The type and amount of data is going to change dramatically,” Imhoff relates. “Much more detailed data and much more data in terms of just volumes. Because we are now talking about multiple loads during a day; intraday loads as opposed to once a day or once a week. Also, very detailed information. It’s not summarized or aggregated data anymore, it’s really the nuts and bolts transactions.”
     IOUG’s Kaplan also advises companies to look at “what types of enterprise information they are looking for to help service customers or improve operations.” Then, start incrementally, he added. “The next step is identifying a few key initiatives to start, since you can get very tempted to go after everything all at once, and then you end up with nothing for anybody. Identify a few key initiatives to start.” Kaplan observed that analytical data warehouses often are used in two distinct ways--either for business queries, or for more sophisticated scientific or mathematical calculations.
     For many organizations, such investments are worth the benefits eventually delivered. For Oakwood, for example, extending its business intelligence platform has let the company increase employee efficiency, strengthen customer relationships, reduce risk and respond more effectively to new business opportunities, Le said. “We can establish how quickly and how efficiently we respond to an opportunity. At any given point, we can look at our inventory availability levels, and how we need to react to the change in demand. We can react much faster to changes based on daily events. With some 200 locations and so many properties worldwide, the ability to unlock data, manage inventory systems and provide a view into the profitability of our business has been huge.”

Back to Ari Kaplan's Home Page