The Credit Union Path of Clean, Quality Data

Editor’s Note: This article is the continuation of an earlier article about how credit unions can prepare for the fourth industrial revolution – artificial Intelligence. In response to that article, readers asked how a credit union cleans up its data to prepare for the onset of AI and simply run a better credit union. This article helps credit union leaders get started on their path to data quality.

How does a credit union that is serious about improving its decision making begin? For most, data quality and hygiene are large problems – too large to solve with an all-encompassing strategy. Instead credit unions must resolve this problem incrementally, using principles and framework-like data clarity. Using this practical approach will put a credit union on the path to data governance that will allow it to make better decisions and serve its members far better than it can without the insights afforded by quality data.

The U.S. Army adopted six data principles in its newly-published “ruthless” data quality strategy. The principles the Army adopted are that data should be visible, accessible, understandable, trusted, interoperable and secure. It might be tempting to take these principles and apply them wholesale. However, a credit union should consider its data problems and select each principle carefully with a clear path back to solving the enterprise data problem.

Selecting these principles cannot be the work of a single individual, regardless of how competent or capable that person is. A group representing the major centers of gravity in the organization like physical/digital banking, operations and information technology should select the principles. Having a group comprised with these perspectives helps ensure that the principles are truly representative of the key parts of the credit union.

It is important to keep the data principles group large enough to make credible decisions, but not so large that scheduling becomes difficult. Typically, three to five members on a core team with the other stakeholders on an extended team is ideal. This group must be working under a charter authorized by the senior leaders to be successful.

With senior leadership support, a chartered team and selected principles, a credit union officially will take its first steps toward data governance. Now what? This is where the selected principles come in.  Depending on the principles selected, the data governance team can begin working with the organization to turn those principles into changes. Understanding that data governance is an incremental process that is never truly complete.

If the data governance team selected the three principles of visible, understandable and secure, the team would build a roadmap to support realizing these principles. This is where having senior leadership’s support is so important. If this team enacts changes to institute data governance, it will mean change throughout the organization and projects to support those changes. For best outcomes, the data governance roadmap should be part of the credit union’s overall strategic roadmap.

Applying these three principles is best done through making data products like reporting visible to the whole organization. Then, the team needs to train stakeholders in the organization on what the reports and its distilled data truly mean. The first iteration of these reports are likely the ones a credit union already has, but the key is to ensure more people in the credit union can see them. Visibility forces data to pass the reality test of, “Does this data make sense?”

The data governance team must collect data about the effectiveness of its efforts. For instance, they can ask, what data in these reports are correct and valuable? They must seek recommendations on how to improve the process, all while keeping data secure. There is a fine balance in data security that the data governance team can help navigate. Some organizations allow compliance to dictate data security, which can paralyze an organization. After all, the safest thing to do with data is not use it.

In this scenario, the team did not adopt “trusted” as its first principle on its incremental path to data governance. Why? Trusted data comes through accountability, visibility and reliability. Laying the foundation for trusted data only comes over time. That is why this notional team selected the principles it did.

The arduous journey of data governance takes its next step from defining principles to the technical execution on those principles. A credit union committed to implementing a data governance strategy expressed through action recognizes that the current state of reporting is not meeting the needs of the institution. To allow fact to become an integral part of decision-making, significant changes to the output and access of information demand equally significant changes to the input.

To meet the new demands, data must be accurate, deep, precise and democratized. Errors, inconsistency and gaps in the foundation of the datasets can result in erroneous analysis, poor outcomes and potential losses. The science of removing errors, standardizing and enriching your data is often referred to as the data clarity process.

Data accuracy is the first target of the data clarity process, and often the largest task. Accuracy means ensuring that all fields and values are as close to their true values as possible. Errors such as misspellings or missing dates have a negative impact on the accuracy of data set. Deep, in terms of data, implies that there is detail and context for all values. Making data deep often includes enriching a credit union’s data with external or third-party data such as demographic information.

Precision measures consistency within the data. Having repeatable results and a stable, documented process is the heart of any data governance initiative. Democratized data refers to giving data “to the people,” which compliments the principles of visible and accessible. This process of making data available should be purposefully managed alongside the security considerations mentioned before.

Much like the practices of choosing the initial data governance principles, the data clarity process should not be deployed wholesale. Cleaning data and making that data accessible is a monumental task. It is important to get quick wins and deploy access to the new-and-improved data in parts. Deciding which datasets to run through the data clarity process should be informed by the data governance team. Determining factors for ranking the priority of the datasets include an objective assessment of value.

Are you at a credit union that wants to implement AI or simply one that wants better data to aid in decision making? Both share a common path to serve members better, and it is through data quality. These are some of the first steps a credit union leader can take on the journey of data governance, which is key for data quality. This journey is long and difficult, but worthwhile, as it will aid in better decisions across the entire organization and in the end, better member experiences.

Ray K. Ragan is Assistant Vice President, Project Management for Vantage West CU. He can be reached at 520-617-4014 or raymond.ragan@vantagewest.org.

Timothy “Buck” Strasser is the Co-Founder of Clear Core and a Senior Performance Manager in CUNA Mutual Group’s analytics division.