Thursday, June 28, 2007

The Commoditization of Business Intelligence? Part 1

A co-worker recently explained a previous consulting gig to me, an assignment with a well-known Asian restaurant chain. In an age of increasingly commoditized I.T. (real or perceived) , this prominent eatery's "decision support solution" consisted of storing rolled-up register tapes in a cardboard box in the stock room. These boxes were regularly shipped to corporate headquarters, where analysts would dutifully eyeball the tapes and manually crunch the numbers in a handheld calculator. Sound like a recipe (pun intended) for success to you? Me neither.

On the assignment, my friend developed software which tracked every item sold in the restaurant, saving the data to a centralized OLAP (online analytical processing) database. This afforded corporate analysts a real-time view of revenue and of diners' purchasing patterns. Sitting down with the company executives, my co-worker showed them how to run queries and reports against this ocean of data. A database query which took mere seconds to run revealed that - by a long shot - the most freqently ordered item at each franchise (aggregated across all locations) was........Volcano Sushi? Pad Thai? Mongolian Beef? No, not even close. The top seller by far was an obscure Japanese bottled beer. My friend immediately advised the execs to increase the brew's price 25 cents. They did, and the restaurant chain experienced record profit growth over the following 2 years.

Had they not implemented this technology, this sales detail would have remained lost amidst a bale of register tape, and the company's financial portrait over those subsequent two years likely would not have beeen as favorable.
Business Intelligence (BI), Data Wareousing, Data Marts, OLAP, Data Mining, Decision Support Services.........whatever you want to call it, it's the wave of the future for business informatics. The technology has been around for years, and (if one believes the advertisements and the pronouncements made at vendor-sponsored conferences), everyone who's anyone in corporate America is jumping on the bandwagon. One might even imagine that BI has become a commmodity, in the same way that the the Internet and database technology in general have become.

But the real-life experiences of those of us who work in database/software development reveal a different story. While many high-profile companies do indeed employ BI tools and methodologies to their full potential, countless more - including the majority of the companies that I and my colleagues have worked for - are mired in the dark ages of spreadsheets and ad-hoc queries against production databases.