By Ross Dennis
Editor’s note: Recently, we asked our members if they were using data mining and what questions they had about it. Ross Dennis, an independent consultant, sent us this commentary on data mining. He wonders if data mining has shown its value. Tell us what you think by joining the discussion.
Fifteen years ago, we were doing "data mining" but called it exploratory data analysis (EDA).
Now EDA is back, renamed "data mining."
Regardless of the name change, the concept still begs the same question from any CXO (chief executive officer, chief information officer, or chief financial officer): Where is the return on the investment?
When we were doing it, EDA was new, exciting, experimental, and "gee whiz." Funding for EDA projects was provided more out of curiosity and on the assumption that the process held great promise for business.
Today, funding for data mining efforts could be obtained in minutes if there were some conclusive ways of demonstrating a return on investment (ROI).
There are more than two dozen commercial data mining software tools readily available on the market today. Are they producing results that the companies can use to improve revenues or reduce operating costs?
Obviously, there is a need to review data for patterns that can be used to benefit the business, but CXOs who approve funds should view data mining as a deep-sea treasure expedition—a lot of resources committed with a small possibility of uncovering something.
So here is my question: How can data mining projects be presented, designed, and executed such that the return is made less risky and so that it can be demonstrated that there are concrete results that will improve your business by X amount for a Y amount of investment?
Do you agree with this assessment of data mining? Do you believe the questions Dennis raises are valid? Join the discussion and share your thoughts.