In the movie Moneyball the General Manager of the Oakland Athletics was faced with a dilemma: His star players were leaving, going to other richer organizations that could pay them more money. Using the old ways of doing business his team was bound to lose--they couldn't compete dollar for dollar with the richer clubs. The old way wasn't going to work for a team with a lesser amount of money.

Even those teams with money were paying millions of dollars per win--as much as 3 million--an incredibly stupid sum. In his book, Moneyball, Michael Lewis describes how Billy Beane, the General Manager of the Oakland A's, succeeded in finding a new understanding of the game of baseball. What Billy Beane found was that the reality of winning baseball games had very little to do with how his scouts modeled the game. His scouts had a faulty model of the game and refused to check or improve it.

Billy Beane found a new model for baseball based on rigorous mathematical analysis: a model where the numbers matched the game. That Billy Beane implemented a model that led the Oakland Athletics to the playoffs on the smallest budget in baseball is testimony to what an accurate model can do for a business.

Business Intelligence:

In our example statistics say that a baseball team that scores a certain amount of runs will make the playoffs.Statistics also show that if a player gets on base he has a better chance of scoring runs. Getting on base is directly correlated to walks for some players. Business Intelligence takes it a step further. Getting walks correlates with bat discipline and bat discipline turns out to be a key success factor for producing runs and parenthetically choosing the right players. Business Intelligence shows meaningful patterns within data; patterns beyond the capability of a spreadsheet or the experience of the General Manager reading it.

Advertising Lost:

Let's look at how Business Intelligence can improve advertising buy decisions. The average advertising budget for a car dealership is tens of thousands to hundreds of thousands of dollars per month. The hope is that a specific ad or series of ads can drive customers to a store is the hoped for result of all advertising. Data Mining, a powerful Business Intelligence tool, can look for key markers of a demographic and see within a certain confidence interval whether the target customers came to the store with enough frequency to declare the advertising a success for that particular weekend. If there is not enough variation from the norm--numbers above margin of error--then the advertising was not a success. Data mining can group customers through market segmentation; i.e., common characteristics of customers who buy that particular brand of car and determine their buying profile. This model can then be used on past advertising campaigns to validate its predictive power. Data Mining can determine effective timing of ads. If direct mail works best on Thursdays and Saturdays then print ads can be timed to Fridays and Sundays to complement and amplify their impact.

Customer Segmentation:

Who is your customer? What identifies that customer in a particular market place? Let's say that the customer in general uses the Internet to determine which brand of car they prefer. The customer then visits particular dealership websites to learn about, inventory, incentives, and prices. A dealership website click path can be isolated and correlated to a certain buyer within the market segment. If the website traffic indicates more interest in a particular model over and above what can be expected normally then advertising, inventory, and telemarketing efforts can be focused with greater effect. 

Intelligent Patterns:

Customer patterns and tendencies may not be intuitively obvious to even the most experienced General Manager. Business Intelligence and specifically Data Mining algorithms can find the meaningful patterns in the data finding previously unknown or counter-intuitive characteristics of the buying customer. Know the patterns, build the model, and make the right decisions.