It is said that “information (data) is power,” but this is only partially correct. What power is there to be gained from information you don’t trust? We argue that power actually equals information (data) + trust. Business leaders must be able to trust and rely upon information (and by extension, data) to enhance decision-making and ultimately derive value from data. Both parts of the data + trust equation must be taken into consideration to achieve data value realization. The crucial requirement for building trust and laying the foundation for business advantage is high data quality.
In our previous articles we have emphasized that data monetization requires an understanding of data as the key resource (see What are Data-Driven Business Models? and Mobilizing Capabilities of Data-Driven Culture). We further highlighted that data quality and reliability are decisive for the value proposition. In this article we provide a deeper view on data quality and its importance for business.
What is Data Quality?
Historically, there has been a variety of definitions of data quality. While most definitions focus solely on the dimensions that should be considered—such as accuracy, completeness, timeliness, consistency, integrity, reliability, uniqueness, and accessibility—others take a broader view on data quality, and describe it as the fitness of data for an intended purpose. A good definition should accommodate both aspects. First, as the purpose of data use is determined by the business user (data consumer), the requirements that the data set must meet may vary. Therefore, data quality very much depends on business context and business needs.
For example, while a data set may meet the requirements for a business use case related to shipping, because it contains required location information in the form of a postal address, it may not be appropriate for use cases requiring more precise information in the form of customer geographic position. Second, to be fit for its intended purpose, the data must be without errors, and data quality attributes must be defined in relation to quality dimensions for that data. Therefore, data quality dimensions—for example data accuracy, data completeness, and data consistency—can be directly related to data quality attributes. These data quality dimensions should be linked to business requirements.
Data quality is data’s fitness for an intended purpose, which is assessed on its possession of certain data quality attributes set alongside data quality dimensions.
Various businesses and institutions have developed methodologies for the assessment of data quality attributes alongside quality dimensions. For example, UnitedHealth Group’s Optum healthcare services subsidiary created the Data Quality Assessment Framework (DQAF), a method for assessing its data quality. The DQAF provides guidelines for measuring data quality dimensions that include completeness, timeliness, validity, consistency, and integrity. Optum has even shared details about the framework so that other organizations may use it.
The International Monetary Fund (IMF), similarly, has its own assessment methodology, also known as the Data Quality Assessment Framework. The focus of the IMF’s assessment tool is the accuracy, reliability, consistency, and other data quality dimensions of the statistical data that member countries submit to the IMF.
Business Impact of Bad Data
The impact of bad data on business revenue is immense. A widely read MIT Sloan Management Review article from 2017 asserted that most companies lose up to 25% of their revenue due to the cost of bad data. All parts of a business suffer when there exists inaccurate, incomplete, redundant or duplicate data. Costs pile up as companies seek to accommodate bad data by correcting errors and looking for confirmation in other sources. They then have to deal with the inevitable mistakes that follow.
Some examples of the economic damage that data quality problems can wreak include added expenses due to products being shipped to the wrong customer address, lost sales opportunities because of erroneous or incomplete customer records, fines for improper financial or regulatory compliance (e.g., GDPR, BCBS 239, CCAR, HIPAA), and a lack of timeliness in goods and services going to market because businesses are acting on delayed business data.
From Data to Business Outcomes
It is not difficult to convince business stakeholders that bad data is connected to bad decision-making and ultimately leads to bad business outcomes (garbage in, garbage out). For data monetization initiatives, poor data quality is recognized as a potential obstacle and threat. Business leaders understand that business practices do not exist in isolation and that successful business outcomes depend on having a data infrastructure informed by policies and practices that lead to the required level of data quality. However, issues can arise with regard to who has ultimate responsibility for data quality and who will fund quality assurance activities.
Who is Responsible for Data Quality?
Clarifying who is to fund the necessary activities for quality assurance and who ultimately “governs” the data across an enterprise can be a difficult task. Such a responsibility requires taking ownership of critical assessments of the prevailing policies and practices for data governance. A large number of quality issues arise from a lack of a cross-business-unit view on data. Therefore, the ideal is that everyone in an organization is responsible for data quality.
Other data quality challenges arise when there is unstructured and semi-structured data, AI and machine learning, and issues around data privacy and protection laws.
There is No Perfect Data
Of course, no data will ever be absolutely perfect. However, an acceptable margin for error must be clearly identified by a business, and the potential impacts of accepted errors on the transformational impact of the resulting analytical models assessed.
Contact us to find out how Assefy can accelerate your data governance journey.
