Sounds like the opening of an ad for an emerging new drug. Do you suffer with Regression Analysis?
“Ask your doctor or pharmacist about twice weekly Dataphyl.”
Let’s be honest, there’s a bunch of math involved in data analysis. And while we can all do some cool stuff in Excel and use formulas to calculate KPIs to keep our credit unions on track… the moment someone says “Coefficient” many of us respond with: “I’m out!”
But if you want to examine the relationship between two or more variables to determine whether a change to one affects another then Regression is your tool. A powerful statistical model, Regression Analysis shows the influence of one or more independent variables (price) on a dependent variable (sales).
The cool thing is that the ability to do a meaningful regression analysis of smaller datasets can easily be done “in-house” using Excel. Getting to the bottom of causation through objective data analysis can have a very material impact on a credit union’s performance.
Take employee attrition as an example. HR best practices have likely led you to conduct exit interviews. You may even be applying some sort of subjective review to “spot trends.” But have you done an actual analysis as to why people are leaving? I am willing to bet you have most of the following data available:
As credit unions have limited numbers of employees to manage, simply combine your data into a single spreadsheet with individual employees in the “rows” and the attributes above making up the “columns.” Your dependent variable will be Attrition. Whether an employee has resigned. The rest of the columns will be the independent variables you are assessing as contributing to the decision to leave. When the Regression is run, Excel will show you the relationships between these variables with just a couple of mouse clicks.
From the result, the number you will be looking to maximize is called the Coefficient of Determination.
Oh no, there’s that word again! OK, so let’s just call it R-Squared (R2).
R2 measures the “amount” of relationship between two variables. The bigger the number, the greater the relationship. So, in analyzing our attrition problem if the variable “TrainingTimesLastYear = 0” had an R2 = 0.8347 this might be worth further investigation.
A key to data efficacy is knowing what is a Data DIY project and what analysis requires professionals. The attrition example is a do-it-yourself exercise that every credit union can and should be doing. Doing a small Regression Analysis in Excel requires watching a YouTube video, not a doctorate in applied mathematics. However, to effectively predict a member’s likelihood to buy a particular product or service based up their transactional behavior… that will likely require advanced-level help.
The important thing to remember is that even though the magnitude of the analysis may change, the method doesn’t. Understanding the approach and the key element (R2) makes you more capable at guiding the best results. A “bigger” R2 indicates a stronger relationship between variables.
BTW in this example: 2x + 3y = 42
2 and 3 are the Coefficients. 🙂