Taxonomies' Impact on Data Reliability and Use

​Companies gather masses of data to drive their decision-making, but not all data are relevant or reliable. Much depends on how that data is being defined, gathered, entered and accessed. For better reliability, some companies are taking steps to create enterprise-level taxonomies—organizational structures and rules around how data is categorized.

What Are Taxonomies?

While taxonomies may seem like something new, the truth is that the concept has been around for some time—even in the HR space. In 1991, Philip Bobko and Craig Russell published an article on taxonomies in Human Resource Management Review.

And, in fact, taxonomies have been around for as long as people had the need to organize information. Likely the most well-known would be the Linnaean taxonomy, which classifies organisms based on domain, kingdom, phylum, class, order, family, genus and species.

"A taxonomy is the process of defining and naming a set of classifications or attributes that are associated with a given key or a given object," said Christine Reges, director of solutions consulting at Claravine, a data integrity software company in Provo, Utah. The taxonomy for a job posting, for example, could be the department that job is in, who the hiring manager is, the compensation range and the technical skills needed. "All of those attributes that roll up to that job posting would be metadata attributes defined within the taxonomy."

Max Blumberg, who is the founder of the Blumberg Partnership consultancy, a research associate at the University of Southern California, and a visiting professor at the University of Leeds in the U.K., said: "When you have a whole lot of data of any kind, you need to organize it in some way. Otherwise, you'll be overwhelmed." Taxonomy, he said, "is just a way of organizing stuff."

Taxonomies represent "a single source of truth—it fulfills the role of what we call a data dictionary," Blumberg continued. A good taxonomy "is all about delivering information as smoothly, as quickly, as usefully and as effectively as possible." 

The Importance of Taxonomies

While terms related to HR management may seem clear-cut and easily, perhaps universally, understood, there is often variation in the use of these terms not only between organizations, but also within organizations. Global organizations face additional challenges.

Blumberg provided a simple example of the use of taxonomies in HR and how issues related to validity and reliability can develop. Suppose a global organization wants to do a comparison of the performance ratings of staff members across its locations. They gather the ratings scores for various positions. But, in doing so, they fail to address variations in scales—for instance, Australia is using a scale of 1-10 and the U.S. a scale of 1-6. Reporting an "average performance score" with this variation is going to be misleading—in fact, inaccurate.

What's required, he said, is a data dictionary where everything is defined for the organization. Not surprisingly, it can take some time to create a data dictionary. Blumberg pointed to a well-known bank he worked with whose board wanted to gain clarity over its headcount globally—it took four months to get there, he said.

That upfront work is critical but can take a lot of time and effort, Reges agreed. "I would say it is probably the most tedious and time-consuming part of the process," she said. But, she added, once that work is done, "it's kind of smooth sailing from there."

How Taxonomies Work

The terms used for the data fields where people are entering information—or where computers may be drawing that information from other systems—are very important and must be consistently understood and consistently applied to ensure data are reliable.

Jon Walden, chief technology officer for the Americas at Blue Prism Software, based in Chicago, explained that "every organization out there has its own language both in reference to the terms used and the processes to accomplish work." Some, he said, have stricter taxonomies due either to their maturity or to regulation.

The digital landscape that most companies now operate in is driving awareness of the need for, and the development of, taxonomies. Many organizations are in the beginning stages of doing this, Walden said.

"The driving factor between both the grammatical and the process taxonomies is for consistency," he said. "If terms and processes are consistent, they can then be more easily automated."

Taxonomies also ensure that data are being labeled appropriately and consistently across the organization so that data outputs will be consistent and can be relied upon—that the data measure what everyone understands is being measured.

Taxonomies and the HR Function

Taxonomies have become increasingly important to HR as the profession has become more sophisticated in understanding the value of data and using data to make people decisions. The field of "people analytics" is burgeoning.

Walden said that his company has worked with several organizations to "define and automate full functionality from hire to retire." HR processes, he noted, "are critical when you consider the need for accuracy and consistency."

While the word "taxonomy" may not be widely used in organizations, and many may not have formal taxonomies, they are using taxonomies whether they know it or not. The term simply refers to how they classify and organize information, as Blumberg pointed out.

Instead, most companies talk about databases—databases where they're storing information. The data is entered into each field based on individuals' understanding of what is being collected. Unfortunately, that common understanding may not exist. When it doesn't, any reports generated from that data will be suspect.

HR Applications for Taxonomies

From an HR standpoint, creating a great taxonomy requires "sitting down with stakeholders and users and really drilling them about what and how they use data about people to execute their business function," Blumberg said. "It's a question of understanding how the data will be used.

For example, if the business is interested in the employee experience, you will want to gather all the possible fields, or pieces of data, that relate to the employee experience. The same with employee engagement, or the hiring and onboarding experience—whatever it may be. The starting point is really the endpoint—what is the information that you need to make business decisions related to people or human resources?"

Reges shared examples of two clients she's worked with that are using HR-related taxonomies—one to track and measure employee engagement with internal communications and another to measure the effectiveness of their recruiting efforts. Tracking employee engagement with communications might look at e-mail open rates or clicks on website links. Recruitment metrics might include the sources of job applicants—e.g., Glassdoor, LinkedIn or internal referral—and which applicants became full-time employees. Then an analysis could be done to determine which sources brought in quality leads that converted from applicants to full-time employees.

Consequently, identifying taxonomies and creating the data warehouses or data dictionaries (the sources of truth) behind those taxonomies will bring together "all of the internal stakeholders who are going to be the end users of the data," Reges said. "Don't be deterred if it takes multiple conversations, because it does."

Lin Grensing-Pophal is a freelance writer in Chippewa Falls, Wis.