Those of us with a long history as business intelligence (BI) practitioners have pretty clear memories of all the days when we saw an overhyped technology promise to change the game by freeing business organizations of IT tyranny with a new class of products that made self-service reporting and analytics better, faster, and cheaper. We saw this with the arrival relational databases. Believe it or not, they were originally all about data access not transaction processing. We saw it again when Online Analytical Processing (OLAP) was available on top of Online Transaction Processing (OLTP). OLAP brought data access directly to our spreadsheets and PowerPoints where we really wanted it. In both cases, business organizations bought this technology and built organizations to use it thinking they could declare their independence from IT. It worked splendidly for a while. IT created extract files from their applications and celebrated getting out from under a backlog of reporting requests. Businesses felt empowered and responsive as they created reports, dashboards, and even derivative databases integrating internal and external data within their siloed subject areas.
Then reality set in.
All these new products required maintenance, documentation, training, version control, and general governance. “Shadow IT” organizations sprung up. They often became, in aggregate, far more expensive and just as cumbersome as what they replaced. Worse, the software vendors happily exploited this balkanization of larger organizations by selling redundant technology that had to be rationalized over time causing licenses to become unused and not transferable. Wouldn’t it be nice to buy a slightly used BI software license at a deep discount?
The fatal flaw in this arrangement is the proliferation of overlapping and inconsistent data presentations that we call multiple versions of the truth. These create mistrust and cause executives to go with their guts in lieu of their data.
Each of these technology advances, along with even faster hardware evolution, did have the impact of making decision support and analytics far more powerful even as the open source movement made it more accessible. This, in turn, created competitive advantage for those who learned to exploit it and made a strong analytics capability mandatory in today’s commercial climate.
One problem still remains. As we like to say, you can buy technology, but you can’t buy your data. Today’s analytics require integrated and governed data across finance, operations and marketing, online and offline, internal and external.
That brings us to the current generation of revolutionary BI tools like the latest data visualization technology that is all the rage right now. (I won’t name names, but think “T” and “Q”.) Just like the previous BI waves, they exploit technology advances very effectively with features like in-memory architectures, wonderful animated graphics for storytelling and dashboards, and even data integration through “blending” and Hadoop access. These products have been hugely successful in the marketplace and are forcing the bigger established players to emulate and/or acquire them. The buyers and advocates are usually not IT organizations, but business units who want to be empowered now.
What does this mean for business decision makers? Just like the technology waves that preceded them, these new visualization tools do not address the organizational and process requirements of a highly functional and sustainable BI capability. Data and tools must be governed and architected together to create effective decision support. Otherwise, you end up with unsupported applications producing powerful independent presentations of untrustworthy data.
We have seen this movie before and we know how it ends.
Mr. Robinson is currently a Business Intelligence and Analytics consultant with Booz Allen Hamilton. He has previously held practice and consulting leadership positions with Ernst & Young, Oracle, Cox Automotive (AutoTrader.com) and Home Depot.com