When We Talk About Data Quality, What Do We Mean?

Four Attributes of Quality Business Data

Recognizing quality data is not always easy; its value sometimes depends on who is using it and what their objectives are. To provide clarity around what businesses should be aiming for when creating a data strategy, here are four main aspects data should have to be generally characterized as “quality”:

Accuracy

This attribute seems straightforward; we are looking for data that is correct. For example, if we are mailing something to a client, the mailing address we use should guarantee delivery.

Keep in mind that “accuracy” itself can be situational — for example, among 10 telephone numbers that might be associated with a business, the one for investor relations is inappropriate if you are trying to reach the main switchboard for a local office. In other words, accuracy is sometimes dependent on context. 

In many cases, we talk about accuracy as if there’s some grand truth against which it can be measured, and that’s not always appropriate. The concept of accuracy is nuanced. It has to be taken in the context of the particular attribute that you’re seeking, and sometimes in the case of the ultimate usage of that data, in order to measure it.

Completeness

Completeness is determined by whether or not data sets capture all data points available for a given instance. For example, if two-thirds of a customer’s purchases are not recorded, then that data set will underestimate the customer’s value. This incomplete data will continue its disservice by inhibiting a company’s ability to identify all high-value customers.

Standardization

Having standardized data enables users to find meaningful ways to compare data sets. For example, establishing data standards for addresses allows companies to compare mailing information for customers around the world, even when comparing vastly different address layouts like Tokyo versus United States addresses. Standardization of your data format is necessary for inputting information, but it is especially important in identifying duplicate data points.

Authority

Data sources need to be authoritative, credible, and fit for purpose. Your data source must be reliable, an industry authority, and trusted by everyone using the data. Without the best input from authoritative data sources your business decisions will be faulty.

Why Data Quality Is Essential for Companies Today

Businesses everywhere are considering, testing, or adopting emerging technologies for competitive gain, organizational efficiencies, and tech investment time to value. A few years ago, enterprise leaders were preoccupied with cloud computing, blockchain, and the Internet of Things (IoT). At this writing, artificial intelligence (AI) is commanding the most focus and dedication of resources as leaders determine how it can best be deployed within their businesses. 

To be truly helpful, AI needs copious amounts of relevant data to base its assumptions and behaviors on. Good data is more likely to lead to good results, while poor data can cause hallucinations, biases, and other errors. And poor data is incredibly common: According to Deloitte, about a third of AI programs fail because of weak data management.  

While organizations have made strides in recognizing the importance of quality data, many have not yet made the connection between good data and AI initiatives. In Dun & Bradstreet’s 2024 B2B Data Report, almost 70% of survey respondents indicated their organization was likely to incorporate AI technologies in the next 2-3 years. But significantly fewer felt confident that their data quality efforts were strong enough to draw meaningful benefits and manage risks from AI use. 

For the business and data science professionals who are trying to perfect their organizations’ data and technology synergies, data quality becomes paramount. It’s critical that they treat data as a high-value asset, building strategies and platforms to ingest, reason with, distribute, and present new business insights across their organizations for countless use cases.

ROI gains become more likely when co-dependent data and technology have equal footing. Data quality therefore stands to be the differentiator for organizations who make it a business imperative.