July 11, 2007

Data Mining vs. Mining Data In Multichannel Retailing

If there's one thing I've learned during my first four months as a small business owner, it is that multichannel retail executives are hesitant about "data mining", but are very enthusiastic about "mining data".

Assume you have a catalog/online CEO who wants to understand how website customers behave between a first visit and a twentieth visit.

A statistician might provide a series of reports and analyses that thoroughly explains the process from soup to nuts, publishing exciting findings along with complex statistical information to support his findings.

The multichannel CEO nods politely, even offers verbal kudos, then leaves the room feeling like she still doesn't understand how her customers behave.

The person who "mines data" identifies the handful of key findings that every CEO must know, then puts the findings into a context, a "story", that the CEO can use to create actionable strategies that drive sales and profit. There is something that the CEO can "do" with the findings, and it is easy for the CEO to "know" what the next steps are.

Mathematically, this type of work is much less satisfying. Professionally, this kind of work can be more gratifying.

Data mining software and data mining experts are generally plentiful and affordable in multichannel retailing.

Folks who "mine data" are generally in short supply, and are desperately needed by multichannel executives.