• Home
• About
• Writers
• Landscape
• Wizards Index
• Fraud
• Wanna Be Like Larry?

Training Wheels for Business Intelligence?

Business intelligence is not like riding a bicycle. In fact, business intelligence requires quite a bit of statistical and mathematical sophistication. Some pundits and marketers believe that visualization will make the outputs of business intelligence systems “actionable”. I don’t agree. There’s another faction in business intelligence who see search as the solution to the brutal costs and complexities of business intelligence. I am on the fence about this “solution” for three reasons. First, if the underlying data are lousy, the outputs are lousy and the user is often none the wiser. Second, the notion of “search” is an interface spin. The user types a query and the system transforms the query into something the system can understand. What if the transformation goes off the tracks? The user is often none the wiser. Third, the notion of visualization combined with search is a typical marketing play: take two undefined notions which sound really good and glue them together. The result is an even more slippery term which, of course, no one defines with mathematical or financial precision.

Now read Channel Web’s “Visualization, Search, Among Emerging Trends in BI”, and you will see how the trade press creates a sense of purpose, movement, and innovation without providing any substance. The source of the article is none other than azure chip consultancy, the Gartner Group. I wrote about the firm’s assertion that no one can “copy” its information. I know at least one reason: I find quite a few of the firm’s assertions off the tracks upon this goose’s railroad runs.

Here’s the key passage in the Channel Web write up for me:

Schlegel identified seven emerging trends that will be key drivers for BI implementations, perhaps even down to the consumer level, in the future. The trends are: interactive visualization, in-memory analytics, BI integrated search, Software-as-a-Service, SOA/mash-ups, predictive modeling and social networking software. "A lot of technologies we’ll talk about to help build BI systems don’t even exist today, but some are right around the corner," he said. "Business intelligence can break out of the corporate world. Usually it’s consumer technology moving into the corporate world. I think it could be the other way around."

“Intelligence”, in my opinion, is an art or practice supported by human and machine-centric systems. Business intelligence remains a niche business because the vendors who market business intelligence systems rely on structured data, statistical routines taught in second and third year stats classes, and anchored in the programming tools from SAS and SPSS (now a unit of IBM). By the way, IBM now owns Cognos and SPSS, which seems to be a market share play, not a technology play in my opinion.)

The end of enterprise libraries caused a vacuum in some organization’s information access. The “regular” business intelligence unit focused on structured data and generating reports that look pretty much like the green bar reports I obtained from stats routines in the mid 1960s. To say that business intelligence methods are anchored in tradition is a bit of an understatement.

The surge in end user access to information on the Internet has thrown a curve to the business intelligence establishment. In response, SAS, for example, licensed the Inxight tools to process information and then purchased Teragram to obtain more of the “unstructured text goodness” that was lacking in traditional SAS installations. New vendors such as Attivio and Clarabridge have exploited this gap in the traditional Business Objects (now part of SAP and owner of Inxight), Cognos, SAS, and SPSS product offerings. I am not sure how successful these “crossover” companies will be. Clarabridge seems to have an edge because its technology plays well with MicroStrategy’s Version 9 system. Attivio is in more of a “go it alone” mode.

With Google’s Fusion Tables and WolframAlpha’s “search” service, there is increasing pressure on business intelligence vendors to:

1. Cut prices
2. Improve return on investment
3. Handle transformation and meta metatagging of unstructured information
4. Deliver better for fee outputs that the math folks from Google and Wolfram do for free.

My hunch is that the Gartner position reflects the traditional world of business intelligence and is designed to sell consulting services, maybe a conference or two.

Much can be done to enhance the usability of business intelligence. I think that in certain situations, visualization tools can clarify certain types of data. The notion of a search interface is a more complicated challenge. My research suggests that Google’s research into converting a query into a useful query that works across fact based information is light years ahead of what’s referenced in the trade publications and most consultants’ descriptions of next generation business intelligence.

When structured and unstructured content are processed in a meaningful way, new types of queries become possible. The outputs of these new types of queries deliver useful business intelligence. My view is that much of business intelligence is going to be disrupted when Google makes available some of its innovations.

In the meantime, the comfortable world of business intelligence will cruise along with incremental improvements until the Google disruption, if it takes place, reworks the landscape. Odds are 70 – 30 for Google to surprise the business intelligence world in the next six to nine months. Fusion Tables are baby steps.

Stephen Arnold, September 17, 2009

Comments

Copyright © 2008 Beyond Search · ArnoldIT.com · Disclaimer · Revolution News theme by Brian Gardner · Log in