GOLDEN MEANS: CREATE THE ANGLES, GAIN THE ADVANTAGE
by Inderpal Bhandari, executive editor at large

In the last few columns, I described Virtual Gold's conference, "The Evolution of Data Mining: Technical Strategies to Beat your Competition by Year 2000" and then expanded on some of the messages from the conference. I will return to that theme later. In this column, I would like to dwell on a breaking story in the National Basketball (NBA) playoffs that sheds light on a fundamental issue in the business of data mining: What happens when all competitors make use of data mining? Does that level the playing field? Or is there still a competitive edge to be had and, if so, how does one get that edge?

Nine of the sixteen teams that made it to the recently concluded first round of the NBA playoffs make use of the IBM Advanced Scout data mining program extensively. Consequently, there were several matchups where both opponents were using data mining. On surface, such a situation appears to offer no advantage to either party. Since both teams make use of the same data mining program, this is about as level a playing field as you can get. However, the playoff experience has made me realize that the war has just begun.

But not the cold war. In the past, I had often mused about how two opposing coaches who were both using Advanced Scout to hone game strategy would react to the knowledge that both sides had the data mining weapon in their arsenal. Perhaps my thinking then is best captured by the following anecdote about the cold war.

A high-ranking officer went to his superior and informed him that the other side knew about their top secret Operation X. The superior smiled indulgently. "I know that," he said. "Yes," replied the subordinate, "but they also know that we know that they know." "I know that too", smirked the boss. "They also know that we know that they know that we know that they know," said the underling, in one final desperate attempt to impress. "They know that?" said his incredulous boss. "Don't just stand there, get me the hot line immediately."

I knew that NBA coaches would have to play out all the angles as part of their game plan given that the other side knew about the same patterns that they did. That would obviously take them to a deeper level of preparation and competitiveness. And, in the end, the better coach would benefit more because he would come up with better answers in the chess game of move and counter move. But what was most interesting was that the coaching staffs went beyond that. They began to think of clever ways to differentiate their use of Advanced Scout so as to retain a competitive advantage over the other side.

Let me explain. The Minnesota Timberwolves, a young and upcoming team, were matched up against the mighty Seattle Supersonics. The Supersonics were heavily favored, yet the Timberwolves nearly pulled off a stunning upset. They went up 2 games to 1 in the best-of-five series, before narrowly losing 3 games to 2.

I found out during the middle of that series that the Timberwolves were using Advanced Scout at half time, the only team in the entire league to start doing so. For reasons having to do with the logistics of data collection, Advanced Scout is designed for post-game use. The NBA collects the game data, uploads it into a central server at the end of every game, cleans it up and then it can be downloaded directly into Advanced Scout. Every team that uses Advanced Scout, including Seattle and Minnesota, was taught to follow this procedure. Thus, the principal use of data mining was for post-game analysis to be used in the game plan for the next game.

But during the playoffs, Minnesota innovated. At half time, they walked over to the operators of the courtside computers, got the data on a diskette and then read it into Advanced Scout. I am sure they knew there was some risk of errors in the data and perhaps they worked around that by restricting Advanced Scout to simpler analyses than they would use for post game. But consider the advantage: What Seattle knew at the end of the game, Minnesota knew at half time!

And, that brings me full circle. What should one expect to see when all parties are making use of data mining in business? The development of innovative and creative methods that distinguish one's use of the technology to confer a competitive edge. The chess game has just begun.

Interested in creative uses of data mining to impact your business? Contact us at http://www.virtualgold.com