To use data analysis you need to acquire sufficient data that is relevant to the business need. Simply put, you need data on what works to promote your business, increase sales or reach other goals, and what stops it from growing.
Sam, director of Admission at a mid-size university, is in charge of retaining students — has plenty of data. His university regularly collects a great deal of data on all their students. Some of it is clearly relevant to understanding and improving future retention. It’s part of running the institution.
He just needs to pin down the factors that have the most influence in determining if freshman return for their sophomore year. Students may drop out because the work is too hard, or their grades haven’t measured up. Maybe their high school academics didn’t prepare them for the rigors of college, or they just aren’t smart enough. Perhaps it is a financial situation and they are unable to afford to return.
Sam has decided to measure these four factors:
- Net price (tuition and fees, after financial aid)
- Date of their application
- Program of study
- High School GPA
It was easy for Sam to figure this out, but that’s not always the case. Ingram Market Analytics can help you review your issues and identify the factors that need to get the answers you want.
With four key factors to consider, he can estimate how much data he needs:
Number of records needed must be greater than 60 x 2 Number of factors
With 4 key factors, the above formula estimates that he needs 960 records. Since Sam has thousands of records, the answer is YES, he has relevant and sufficient data for data mining.
Sam has it easy. Your customers may not share this much data with you. Generally, in considering your situation, it is important to have data that represents:
- Stimulus and response (what was done and what happened)
- Positive and negative outcomes
Simply put, you need data on both what works and what doesn’t. Both on products sold and products returned, people who ordered your service a second or third time, and those who never call again.
But before you do this analysis, let’s look at what you want to accomplish, that is the end results in terms of performance you need.