Good decision-makers are naturally suspicious of data quality. They rely on their experiences, particularly with data, to guide their decisions, and they have a healthy skepticism about the quality, origins, and meaning of data. That leads to thorough analysis and high standards. For these skeptics, it’s better to use previously known data and no new information than take a risk on the unknown. 

And those good decision-makers? Their quest for knowledge and accuracy typically elevates them into leadership positions. So when something goes wrong within a business, the leadership first investigates the data.

And that can lead to a few things:

• A lesser leader will blame the data. A sales exec looks at a number and says, “There’s no way my team performed so poorly this quarter. They’ve been working nonstop. This must be bad data.” They find fault with the data to justify the bad results, and then we have to prove the data.

• A better leader will look at the data and be troubled, but before they take action they’ll be concerned about the quality of data. They’ll be naturally skeptical of the data, but also of the performance driving the data. They do need to be convinced the numbers are right, but they also don’t assume they’re wrong. They expect proof to substantiate the data.

Data is guilty until proven innocent. Any numbers we provide are going to be suspect to leaders who will use them to make decisions because people are protective of the quality of their decisions. Data quality and leadership trust levels may not seem connected, but they’re inextricably linked. Here’s why ...

Leaders have to build trust

Study after study shows people are trusting less and less in this world. And with a tightening labor market, it’s getting harder to get and keep good people. Maintaining very hard-earned trust is critical for leaders. And breaking trust is a top way leaders can lose their talent. So how do leaders build and keep trust?

One study showed 71% of respondents trust others based on the consistency, predictability, and quality of work. How do they judge consistency, predictability, and quality of work? Evidence of past accomplishments and demonstrated capability. 

Leaders who make decisions based on bad data put their integrity on the line. Bad data puts a leader’s trustworthiness at risk. To build trust in the leader, we must build trust in the data. 

Building trust in the data

And to build trust in data, we need to truly know and understand the data we have. Much like a leader must demonstrate consistency, predictability, and quality of work, so must we prove those attributes of the data.

From a tools standpoint, the modern approach to do this is a data catalog: an inventory of all of your data, definitions for every data element and information about where the data comes from.

This data catalog makes what is being measured understandable to the business.

But that’s not enough. That just says what the data is supposed to be. It doesn’t say what the actual content of the data is in comparison to the business’ expectations of data quality. In order to understand the quality of the content of the data, you have to have a data quality system — something that measures quality for the data against a set of quality rules determined by the business. (Some data catalog software packages are sold with data profiling and quality management capabilities, however.)

For business leaders to understand and trust the data, they need to be a part of setting the quality metrics. And the quality system needs to alert the right people when there are deficiencies in the data beyond what is tolerated. Part of that process is to establish the thresholds of tolerance. What does good mean? What are the processes compared to that measure of good?

We could have a rule that we should have a federal tax ID for every business we work with. That’s a data quality rule that can be implemented. If there are more than a certain percentage of these that are missing, the system needs to let someone know so we can fix it.

You have to pile up all of those quality rules monitoring and running the data so when someone asks, “Is this data good enough?” you have an answer.

Because of the quality rules, our system tells us 3% of our accounts don’t have the federal tax ID numbers. We can give the leaders the numbers and they can decide if that accuracy is enough for them. Behind every business metric should be quality metrics that describe how good that business metric is. That’s what provides the trust.

A leader isn’t trusted because of how they do the mechanical functions of their work. They have to do those things. They’re trusted because of the values behind the functions. When we make those data quality rules, we are making statements about values for our company. We’re saying these are the things we value about quality, and then we can measure and meet those values.

By proving the data, we’re providing the trust our leaders — and we — need to demonstrate the data isn’t guilty.

BIO:

Aaron Fuller is the principal and owner of Superior Data Strategies LLC. Located in Lansing, Michigan, Superior Data Strategies focuses on data warehousing, dimensional data marts, operational data stores, service-oriented architecture and data integration in a variety of industries. Aaron’s role as principal involves expertly guiding clients toward reliable and valuable business solutions. He leads a team of highly skilled database professionals who are uniquely capable of planning and executing agile data projects.

Aaron is skilled in dozens of software, databases and standards and methodological programs. He is a faculty member at TDWI, educating the next generation of technology leaders and developers. Aaron holds a bachelor of arts in systems analysis and application development from Davenport University and is a Certified Business Professional at the mastery level.

Contact Aaron at: [email protected]