When used properly, data can be the fuel that propels organizations towards growth and success. Most credit unions understand that the data that has been collected over the years can be one of their most valuable assets to improve member experiences, boost efficiency and revenue, and make faster, more accurate business decisions.
The problem we all face is having too much data that may not be properly compiled, cleansed, and organized. In other words, it isn’t information — it’s just data and raw numbers. So, how do you remove the data clutter and get better at data analytics that transforms unorganized data into insightful, actionable information?
These tips will help you and your team transform mountains of data into steps to improve your data analytics strategy:
Clearly Define Your Key Performance Indicators (KPIs)
Having clearly defined organizational KPIs gives your analytics efforts a guiding direction. There are thousands of different analytic opportunities your team could explore. But, if they don’t align with the organization’s strategic objectives (which the KPIs should directly measure), then is that effort really worth it?
If a KPI is to lower the average cost of funds, then your analytics initiatives shouldn’t focus on how to target more CD deposits but gaining greater engagement in core deposits. This strategic-to-operational alignment needs to exist in your analytics team’s efforts to maintain clarity and focus.
Many credit unions struggle to align their strategic goals with operational goals. This can be due to Board resistance to changing of traditional KPIs (like return on assets or membership growth), but this realignment is essential.
Define Business Objectives
The first step in the process of turning data into information is asking questions and defining your business objectives. Data for the sake of data is meaningless. Start by outlining some clearly defined business goals, use-cases and objectives around your data analytics and analysis. Figure out the questions you would like the data to answer. For example, how many members have obtained an auto loan or mortgage in the last 24 months, but do not have an active checking account? Why are so many loan applications falling off before being approved? These types of questions are multidimensional and provide greater opportunity for depth of analysis.
To get the best results out of the data, a question (or series of questions) needs to guide your starting point. Once those questions have been laid out, consider the question, “What am I going to do with this data or analysis once I get it back?” If the answer is “nothing” then it poses no tangible value to the organization.
You always need to consider the potential decisions you might take based on the outcomes of analysis. This forces data to become a decision-making tool instead of confirming a previous thought.
Avoid Confirmation Bias
As alluded to in the previous paragraph, many organizations use data today to add support to a belief they already hold. When viewed from a biased perspective, it is easy to manipulate an interpretation to match a previously held belief. Successfully embedding analytics in the decision-making process requires you to eliminate confirmation bias at the outset. Treat each analytics use case with the scientific method approach – establish a baseline hypothesis and objectively measure whether that hypothesis is true or false. Much organizational learning will come from this rigorous and honest approach to data.
Don’t Let Perfect Get in the Way of Good
Many organizations that are more immature in the analytics journey fixate on superficial data quality issues that arise. Things like bad social security numbers or inconsistent addresses drive most executives crazy. But if that data is only 90% or 95% accurate, will your decision change substantially with 100% accurate data? The answer is almost always, no.
Getting data to be perfectly clean is both impossible and expensive. It requires extensive cleanup effort, manual massaging of data, and may involve outside vendors to improve the data quality. My recommendation is to get data as clean as possible up to the inflection point where each additional unit of data quality starts to cost more than its benefit. Strike the right balance between data quality and data integrity without chasing some ideal perfection that will never be attained. Sometimes good enough is good enough.
Successful Analytics Programs are not Grassroots Efforts
Executive support is one of the leading reasons why some analytics succeed, and others fail. Support for using data as decision-making tool needs to spread from the top of the org chart down. Without this clear top-down buy-in, mid-level managers will not be required to provide data-driven justification for their recommendations. This then proliferates further into the organization and minimizes the importance of using data.
Many organizations believe that they can simply hire a few analysts and all of their data problems will be solved. The reality is that senior management needs to believe that data can be a powerful decision-making tool. If strategic decisions are not being driven based on some foundational data analysis, then your organization is merely guessing and hoping you made the right choice.
Successful analytics initiatives require executives to hold their teams and each other accountable for bringing data-driven discussions to the table instead of instinct-driven ideas.
The Right Technology
Inevitably, the right data platform needs to exist for success with analytics. Credit unions cannot continue to use Excel as their primary tool for data integration, data aggregation, and reporting.
Leverage data platforms like through experienced, time-tested vendors, who can integrate all critical data sources into a single solution. A single platform allows your organization to embed data definitions, data quality resolutions, and cross-system integrations into a single solution. This then becomes the definitive source of data for the credit union. No longer will you spend forty minutes of an hour-long meeting comparing data points and talking about “which report is right”.
Did you notice how technology only gets mentioned in the last section of an article on analytics? That’s because technology is not the hard part of analytics. While it is necessary for success, the previously mentioned cultural and change management components are far more likely to determine success than technology alone.
Getting our hands around all of our member data is no small task. But, by leveraging the tips discussed in this article, your credit union can start asking the right questions to begin leveraging data as a decision-making tool. This transition to treating data as an asset instead of a byproduct is the next step most credit unions need to take as they look towards the future.
Brewster Knowlton is the Owner and Principal Consultant of The Knowlton Group, a data and analytics consultancy with a focus on the financial industry. Brewster is an expert analytics developer and strategist whose driving mission is to help all organizations become data-driven.