How many of these names do you recognize: Duesenberg, Minerva, Packard, Isotta-Fraschini, Hispano-Suiza, Pierce-Arrow, Budd, Brewster, LeBaron, Murphy, Bohman, Schwartz, Peerless, Spohn, Fleetwood, Fisher, Darrin or Cantrell? Not many? No surprise. They were all automobile chassis, specialized component or body designers and manufacturers from the early 20th century. These are the names that dominated the young days of the automobile, back when each auto was custom made to order and painstakingly hand assembled from the disparate "best-of-breed" elements. In those days, autos were the exclusive domain of the very rich, for only they could afford the expensive craftsmen required to build and maintain them.
Time, progress and a headstrong gentleman named Henry Ford changed all of that. Once mass production techniques were applied to the construction of automobiles nearly everyone could afford independent, reliable transportation in the form of a Ford Model T. Of course, some sacrifices had to be made. There were no Cowl Phaeton models available, and you could have any color you liked as long as it was black.
Today, we stand on the cusp of the Model T business intelligence system era. Heretofore, data warehouse/business intelligence (BI) systems have been very expensive, custom built and handmade by highly skilled craftspeople. Moving forward, the trend will be to buy, rather than build, BI systems. They will be built on very low cost, ubiquitous platforms. It will truly be the democratization of business intelligence, just as Mr. Ford democratized independent transportation.
In the scariest manifestation of this phenomenon, your organization will be soon be swimming in the turnkey, low cost, business intelligence/analytical application offerings of multitudes of vendors. Unfortunately, you are not going to be the one making the purchase or architectural decisions related to these offerings. These systems are going to be priced comfortably within the signing authority of mid-level business managers, whose eyes will be aglow at the prospect of immediate business impact.
Another variation on this theme is the ERP-driven packaged solution. In this scenario, the business discovers that their brand new $30 million dollar ERP system provides little to no BI capability. However, the ERP vendor does have several very interesting and desirable analytical applications available that the business managers are lusting after. The only caveat is that you must buy the ERP vendor's data warehouse infrastructure in order to enable the analytical applications. Worse yet, you've got multiple brands of ERP systems, each with their own very powerful political constituencies, clamoring for their vendor's proprietary BI solutions. Of course, this throws your lovingly crafted, single-hub-and-spoke architecture vision out the window.
And herein lies the crux of the BI professional's dilemma. While certain vested interest vendors and the ivory tower types promote the hub-and-spoke as the "promised land" data warehouse architecture, those of us in the field at a typical enterprise are presented with multiple silo data warehouses and non-architected data mart systems. On the one hand, we've got the vendor/analyst camp, who only talk to and work with IT, proclaiming that the ber hub-and-spoke data warehouse is the only viable, money-saving architectural alternative. On the other hand, out in the day-to-day reality of life in the trenches, we've got the political reality of the business calling the shots.
What the theorists and idealists miss is that no data architect, IT manager or even CIO stands a chance of winning the "but architecture will save us $3 million" vs. "this turnkey application will bring the business $10 million in incremental revenue" argument championed by the VP of Sales or COO. In the last month, I've asked well over 500 BI professionals if they could cite even one example where an architecture argument prevailed over a business leader. The score so far: business 500, architecture 0.
The point is that most businesses today have multiple silo data warehouse systems and multitudes of non-architected data marts. Driven by the market forces mentioned earlier, this situation is not going to magically morph into the idealists' hub-and-spoke, monolithic, enterprise data warehouse. Instead, it promises to become even more diversified. To prepare for this future, and to start to show the business that you have a plan to achieve a single version of the truth (since your current silo data warehouse systems aren't very convincing), you should transition your existing systems to, and build your future systems in, a federated architecture.
Just as standards such as battery voltage, light bulb size and lubricating oil viscosity allowed multiple makes of autos to share components, a federated architecture allows you to support the inevitable proliferation of data warehouse and data mart systems in a productive, efficient manner. Next month we'll take a closer look at the federated future.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access