Not so fast for enterprise-wide analytics
- 14 June, 2004 09:49
Software is entering the Age of Analytics. Buzzwords fly fast and furious. Predictive analytics. Enterprise reporting. Analytic applications. BPM, CPM, EPM. Metrics and dashboards. Three key performance indicators, two turtledoves, a scorecard and a few decision trees.
And if there aren't enough for you already, it's time to get ready for the next blazing software industry buzzword: enterprise analytics, a catcall term describing analytic technology deployed across a whole enterprise.
Enterprises want to whittle down their list of analytics vendors. Business intelligence (BI) vendors are racing to offer soup-to-nuts enterprise analytics product suites, application software vendors are trying to beat them to the punch and Oracle claims to have occupied the high ground for years. If 'enterprise analytics solutions' catch on, just about everybody will benefit.
Unfortunately, it won't be quite that easy. Very few organizations have ever deployed true enterprise-wide analytic technology. Indeed, even the BI vendors themselves generally haven't rolled it out. And there are good reasons for this.
The major problems inhibiting enterprise adoption aren't in the technology itself, which on the whole is delightfully real and affordable. BI and other analytic technologies are much cheaper to buy and install than OLTP application suites. Recent BI-oriented advances in relational database technology make even huge data warehouses scalable. True, data integration and quality issues are often messy when you build or expand a data warehouse, and dealing with them isn't fun or cheap.
Actually, the problem may be even worse than it has been for transaction-processing applications. Controlling and changing how your managers work is a lot harder than transforming the activities of your data entry operators, shipping clerks and other low-level transaction processors. If your CEO really hopes to create an enterprise-wide culture of sober, rational, unbiased number-driven decision makers, that project could take decades.
There's another side to the problem: many people just don't feel comfortable dealing with numbers. That's usually not a problem for MBAs or for engineers and other technically literate sorts.
Dashboards may be excessively hyped but the reality is pretty cool too. Most of the benefits from analytic technologies over the years have come from shoving reports in people's faces until they discover useful facts about their business.