Strategy & Leadership

Data quality: the foundation of effective data governance

December 12, 2018


December 12, 2018

Michael Hoffmann


Michael was an editor for The Economist Intelligence Unit’s thought leadership division in the Americas. He was previously an equity research analyst covering cybersecurity, data networking, cloud computing and IT infrastructure. He has also worked on several grants from the National Institutes of Health to research topics including HIV/AIDS, non-communicable diseases and international human rights law. His work has been published in several peer-reviewed journals including AIDS Care, Global Public Health and the Journal of the International AIDS Society. He received his bachelor’s degree in International Relations from Brown University. He has worked in Latin America and is fluent in Spanish and Portuguese. 

Even as many companies are recognising the potential of data as a strategic asset, data quality remains a critical component of data governance and efforts to unlock business value from data.

A survey of more than 500 business executives in North America and Europe, conducted by The Economist Intelligence Unit and sponsored by Collibra, finds that the objective of data governance programmes is accuracy above all. Improving data quality ranks as the most important benefit of data governance for 38% of survey respondents, equalled only by the related goal of data security. Fifty-one percent of respondents say data accuracy is an important metric of success for their data-governance programmes, the highest percentage for any metric.
Accurate, complete information is essential to every objective of data governance, from regulatory compliance to business strategy. New privacy rules such as the EU’s General Data Privacy Regulation (GDPR) require companies to verify consumer information collected and used in business operations. Inaccurate data—such as errors in credit reports or procurements—can expose companies to lawsuits by aggrieved consumers or disrupt global supply chains. 
Indeed, data-related errors and omissions undermine any effort to tap the strategic potential of data. The primary benefit of effective data governance is insight into matters such as customer needs, market trends or operational efficiency. But conclusions based on poor data are not useful for business strategy. Instead, they mislead companies into chasing shadows at great cost in time and treasure.
Yet companies looking to transform their data into a source of value also recognise that data quality requires more than minimising inaccuracies. High-quality datasets are rich in detail and include contextual information that enables company leaders to make informed, forwardlookingbusiness decisions. Actionable insights flow from datasets that reveal not only ongoing  trends but also their underlying drivers.
Moreover, inaccurate data cannot be resolved simply by adding new technologies to business operations. In fact, having accurate data is a critical first step to maximising the value of emerging technologies. For instance, “people are waking up to artificial intelligence and machine learning in order to make better decisions, but those algorithms are highly dependent on accurate data. If companies want to realise the promise of decision automation, data governance helps provide a level of accuracy that is foundational for those technologies,” says Greg Arnold, senior director of data science and data governance at Level 3 Communications, now part of CenturyLink.
The role of data governance
A strong data-governance programme ensures the reliability and integrity of information by identifying trustworthy data sources and establishing verification standards and processes. Data governance can help determine what information decision-makers need and creating systems and processes to gather these data appropriately. Such organisational changes can also facilitate accurate, consistent interpretation of data by defining terms and sometimes creating a lingua franca in the form of data dictionaries used throughout an organisation.

A few ways organisations can improve the quality of their data include the following:

Keep data up-to-date.
Many companies use automated, real-time quality checks to make sure employees are fully and accurately recording data on a continuous basis. Continuous updates are critical to strategic initiatives such as Honeywell Aerospace’s predictive maintenance service offering, says Dr Abhi Seth, senior director of data science
and advanced analytics at US aircraft parts manufacturer, Honeywell Aerospace. The Data Science team developing predictive maintenance models depends on colleagues in the company’s repair and overhaul operations and their customers to supply a steady stream of data to update existing predictive models and build new analytics capabilities.
Gather data from all sources.
Richer datasets have provided American Express with significant cost savings. The financial services company has started gathering more data on borrowers, enabling it to quantify loan default risks for individual customers. That level of precision reduces the amount of capital American Express has to reserve for potential losses under banking rules.
Verify accuracy continuously. A data governance initiative at Northern Trust, a large asset manager and custodial bank, provides associates with data quality dashboards that monitor the accuracy of data from various sources. Results can be revealing. For example, the dashboards showed that external sources often supplied incorrect security identifiers, critical data in the money management business. Internal transaction data, by contrast, turned out to be 99% accurate. As a result, Northern Trust officials worked with information vendors to improve accuracy.

The time is now

Business’s task of optimising data for use has become even more challenging, though increasingly vital, amid growing mistrust of corporate data-gathering tactics. A variety of companies, from internet giants Facebook and Google to large banks and healthcare providers, are under fire for data practices that often seem deceptive, intrusive and exploitative to customers. Public anger over these practices could translate into legislation limiting the ability of companies to capitalise on data.
Potentially more damaging than new legal restrictions is declining trust, from both internal and external stakeholders, that is critical to business decision-making and strategy.
Mr Arnold notes that “while data accuracy might be abstract to executives, they understand what data accuracy enables or disables—most notably trust, or trust in the ability to use your data to make reliable decisions.”
Even the highest-quality data lose value without trust. And no data-driven strategy can succeed if wary consumers decline to give companies access to meaningful information.
To rebuild trust, some companies are starting to disclose more information about how they gather, verify, safeguard and disseminate information. Greater data transparency may be the best approach for organisations hoping to get the most value from their data. 

Enjoy in-depth insights and expert analysis - subscribe to our Perspectives newsletter, delivered every week