TORONTO — Enterprise companies need to rethink the data they measure if they want to make the most of analytic applications, according to an analyst who predicted massive consolidation among vendors in the market.
Experts are using the term “”analytical applications”” to describe a sort of
grown-up version of business intelligence (BI) tools that help large organizations glean meaningful information from vast amounts of data. These in turn are the successors to decision support systems (DSS) that were prevalent in the retail sector for more than 10 years.
At a day-long series of educational sessions Thursday, however, Meta Group analyst Doug Lancy said many enterprises are too focused on using BI or analytic tools to track the wrong key performance indicators.
“”Measuring things like profitability was interesting, but we need to use analytics to give us data on things that are actionable,”” he said. “”We need to look at things that aren’t bottom line-oriented, like service and quality of work.””
Often, Lancy said, enterprises have implemented BI or analytic tools to satisfy the management curiosity rather than supplying information that turns into decisions. “”We need to get out of mode where we’re delivering analytics to eyeballs and move to a place where we’re delivering analytics to a process.””
While many organizations once built DSSes from the ground up, many enterprises now purchase a pre-packaged product that they then customized for specific business needs. There are many vendors in this space to choose from — J.D. Edwards and Informatica Corp. both spoke at the event — but some firms struggle to understand the parameters to implementing a solution.
Philippe Torres, vice-president of technology and R&D at Toronto-based SQLiaison, said his firm works with a variety of customers helping them design e-business intelligence strategies. Some of these organizations may be implementing a software tool to perform analytics on financial data today, he said, but may be interested in customer relationship management analytics tomorrow. “”When you’re buying versus building, you’re buying a certain amount of upgradeability as well,”” he said.
Torres offered a number of questions IT managers should consider before they purchase or deploy analytic applications: How many data sources need to be integrated into the final environment? Are the analytical application’s subject areas relevant to your business processes? Does the application’s data model provide enough coverage for required measures and dimensions? “”You will not find an enterprise analytical application that has 100 per cent coverage of the data model for your specific application,”” he said. “”That’s setting the bar too high.””
A lot of enterprise companies have already established data marts of data warehouses, and Lancy predicted that many of them would spend the next several years blending architectures as enterprise analytics extends outside departments to the entire enterprise. He outlined eight ways companies attempt this, from dismantling existing data marts and starting over from scratch (which he dubbed, “”Back to the Future,””) to temporarily synchronizing data marts within a large multinational organization on a periodic basis (which he called, “”a long distance relationship””). Other enterprises try to “”join at the hip”” by linking data marts with common fields, or try the “”oil and vinegar”” strategy of taking tables from multiple data marts and merging them into a single database.
Whatever the approach, Lancy said companies will be looking beyond the historic clustering, segmentation or population trend data they have typically run through analytic applications. “”We want more predictive data,”” he said. “”Not just this incidence-based record of what’s happened, but reports that will tell me what’s happening next and how I can act on that.””
The analytic applications market may be flourishing, but that doesn’t mean the industry is likely to see many more new entrants. “”If you attend DCI (a major trade show for the BI market), you’ll see one quarter of the vendors you saw there five years ago,”” Lancy said.
Torres recommended the audience check to see if there is a solution available from their enterprise resource planning vendor before choosing third-party products, but warned them that the tools should offer more than traditional reporting capability. Lancy agreed: he said a recent Meta Group survey showed that while a large percentage of enterprise companies have deployed some kind of reporting tool, they don’t plan to purchase any more, preferring to spend money on additional functionality.