Forget big data: Let’s get small data right first
It seems everywhere you turn nowadays, big data is being touted as the cure-all for what ails just about every industry. While big data does hold great potential for credit unions down the road, the current costs and difficulty to implement the collection, sorting and formulation of actionable insights is still quite high. So before you even think about big data, let’s take a look at “small data” first.
Small data in contrast to its much-hyped Big Data cousin is information a credit union can conveniently extract, store and analyze using their current hardware and software technology (i.e. server, desktop/laptop, MCIF, Google Analytics, etc.). This data is often just sitting on credit union servers waiting to be analyzed. An article published by Forbes last year addressed the differences between big and small data and concluded that making small data actionable is far more realistic and cost-effective for most organizations at this point.
The good news is that small data can be found in many areas of your credit union. One example I ran across recently: monitoring when members delete all of their bill payments, as this can be a clear sign they are moving accounts elsewhere. Instead of ignoring these members, you could proactively have your staff reach out and check in on them. This data won’t help you save every membership (if the member is indeed leaving), but you might be able to save some accounts and at least gather feedback from your members either way.
Another example is systematically storing pre-employment data, such as interview questions, assessments and demographic data for newly hired employees. This data can be analyzed over time to determine the types of employees you are retaining, see what similarities exist among employees who leave, and can also help you hone your training and onboarding processes as well.
A third example would be emailing members on their anniversary date of joining the credit union with a special customized offer. If the member has never had a loan with your organization, you could offer them a discount, good for 60 days. Another offer could be thanking them for their business and giving a gift card to their favorite restaurant (found by mining transaction data) for completing a survey.
While the examples above are just three of many, the key is to begin monitoring and organizing various data sets proactively. This process does not require data scientists or fancy software and can be integrated into a broader data management strategy or be instituted on a smaller scale for relatively low or no cost.
No matter where you begin, an important element is to maintain the quality of your data. Don’t keep adding data if it isn’t directly applicable to your analysis. Too often, managers and IT staff get overwhelmed with the idea that you need to gather larger and larger volumes of information to make it actionable and that simply is not the case. By focusing on small data, insights can be gained quickly and then scaled up over time.
