There is lots of buzz about big data and data analytics, but all the data in the world does no good unless it’s being used. Credit unions are behind retail and online companies in using data to make informed decisions. For example, if your forms, such as a HELOC application, does not have the fields for name, address, etc. already filled in for your members, that indicates that you are probably not using your data. You should already know this information about your members. Save them the hassle and give them the option of updating if necessary.
There’s so much more data than was available in the past that can be collected and used for purposes that can benefit the member, and there are better tools than previously available to aggregate the data to help decision makers. These two factors are starting a wave of bringing data analytics to credit unions. More importantly, it’s a matter of survival. Credit unions must take advantage of these opportunities to guide their sales initiatives. For example, the data can help you to decide who to target, like a new member who joined to access an auto loan should be offered your CU-branded credit card to maintain a sticky relationship. Or who not to target for a specific product, like a member who already has your credit card but accessed a new loan, should not be sent another offer for the same credit card. Rich data helps you to determine who your target is for specific products and services and this helps to enhance your member experience
One of the stumbling blocks when using data for decisions is to get the data into a format that is normalized and searchable. Unfortunately, most credit unions get reports from their processors, loan origination systems, core platforms, rewards program and mortgage provider in human-readable form, but not machine readable. The data must be uploaded into a database that can be used for data analytics or inputted into another system that analyzes it. Having a person type this data into a spreadsheet for analysis is time-consuming and prone to errors, and the volume of data makes this a near-impossible task.
continue reading »