Using Good Data to Make Good Decisions

Every decision is a risk and every risk is a decision. This is a mantra often recited in military science. While it might be obvious why the military holds this mantra, it also applies to credit unions.

Historically, financial institutions tend to be risk averse in both their long-term strategic planning and unforeseen, daily operational challenges. This risk aversion is sensible, as all credit unions are beholden to key metrics such as loan to share or net worth to asset ratios. Outpace these metrics and your credit union can be in real trouble. However, a credit union leader can use modern data-driven decision making to take the uncertainty out of risk.

Modern data-driven decision making finds its roots in probability theory with expected value. This theory helps give decision makers a way to quantify a risk, whether it’s good or bad. Simply put, expected value is the value of a risk multiplied by its probability. So, if you have a $1 million risk that has a 10% chance of happening, the expected value of that risk is $100,000. This is extremely helpful when leaders consider competing courses of action with multiple risks; in fact it gave rise to an entirely new form of strategy called value-based strategy.

A common critique of expected value is that a risk is intangible or unknowable. This statement is incorrect. Risk may be difficult to quantify, but it is not completely unknowable. Whatever risk your credit union is facing, another organization has faced a similar risk and either succeeded or failed. It takes some creativity to use historic use cases, but it can be done. The preferred method involves using recent and relevant internal data in quantifying risk. To accomplish this, it is best to use a technique called bracketing, in which formal or informal risk managers find the lowest and highest likelihood of a risk. Bracketing gives a leader a much clearer picture of the full spectrum of risk.

This modern approach to data-driven decision making is frequently met with a statement like, “We’ve always done great when I followed my gut.” This may be true, at least in the short run, but when followed over the course of years, it will prove to be wrong. That bold and seemingly counter-intuitive statement is true due to the statistical phenomenon known as regression to the mean. The phenomenon says that given enough chances, a performance will eventually find its natural place among the average – after all, that is what the average is. The regression may be delayed by talent, tools and luck, but eventually it will find its place in the peer group mean.

Knowing this, a credit union leader may step back and ask him or herself, “How many of my team members are making decisions by following their gut?” If the answer is most or all, then the credit union is being exposed to many forms of risk that its leaders may not even be aware of. Eventually, one or more of the outcomes from those gut-based decisions will fall below the mean. Some of those below-the-mean performances can be far below the mean. Using data-driven decision making with expected value will normalize the outcomes and place them closer to the mean for consistent results. This is why computer models always out-perform humans when tasked with making the same decision over time – consistency is the key.

When applying probability theory and expected value, it is critical to use data. Data is what allows a credit union to transform decisions that were previously made according to a gut feeling to something that is repeatable and visible. Once leaders use data in their decision making, the models and data can be compared with each iteration and leaders can assess the success or failure of a strategy, policy or tactic.

A credit union leader in pursuit of data-driven decision making will most likely be confronted by a monumental challenge in accessing the data. Even when the data is made available, its integrity will be poor and often contradictory. This is typical, but for a credit union leader seeking long-lasting strategic outcomes, the resulting frustrations will not dissuade them. Rather, they will serve as the rallying call for the organization to adopt a data culture and improve data quality.

Adopting a data culture is no small feat – especially when data challenges long-held beliefs or contradicts tribal knowledge of things that “always” or “never” occur. It is something that must be done across the organization and reinforced time and time again. A strong data culture is reinforced in the day-to-day activities of frontline team members but ultimately lives and dies by the example of decisive, data-driven leaders. Demands on the credit union’s data providers must be high and data governance must be taken seriously – with the dedicated resources in place to match. This is all for the purpose of enabling decision-making using data. Good decisions and good data are inextricably linked.

Former Secretary of State and retired General Colin Powell famously introduced the 40-70 rule in decision making. In this rule, he said that when faced with making a decision, he prefers to gather between 40% and 70% of the data and information. If he has less than 40%, the decision will be a gamble. If he has above 70%, the decision likely will be made too late in the game and come at the expense of the initiative. This serves as a good guideline for balancing the needs of the credit union and decision quality.

Clean data is key for making quality decisions. Modern data-driven decision making makes it quite easy for credit union leaders to consistently make the right calls and avoid falling prey to unknown risk exposures. Data-driven decision making allows leaders to normalize the outcomes of strategic and tactical decisions against the mean. This mitigates the risk that would place an outcome far below the mean.

Having the clean, quality data required to make decisions will likely be a source of frustration for leaders, but it is not impossible to overcome. In the end, by adopting data-driven decision making, a leader can expect his or her credit union to out-perform other financial institutions that do not use this repeatable and visible approach to decision making.

Ray K. Ragan, Co-Founder Clear Core, Tucson, Ariz.

Timothy “Buck” Strasser, Founder Clear Core, Tucson, Ariz.