Board Requisite for Data Quality and Oversight
Board Requisite for Data Quality and Oversight
Data quality is a critical factor for organizations of all sizes. Poor data quality can lead to inaccurate business decisions, missed opportunities, and even financial losses. As confidence in the quality of financial results are a concern for a board of directors, underlying data quality should be as well. While there are challenges to ensuring quality data, there are also game-changing steps that can be undertaken that can support stronger business decisions.
Data Quality Affects All Organizations
In 2018, an Experian survey produced some interesting yet conflicting results. in surveying employees with visibility of data management issues within the business:
|69% of respondents believe that inaccurate data undermines their ability to deliver an excellent customer experience
|Data management projects to attain a single view of a customer and improve holistic data quality were organizations’ two lowest priorities
|81% of respondents trust their data
|The average organization’s data is 30% inaccurate
So, one in three records are wrong, people have faith in that data anyway, and organizations do not want to do anything about it. More recently in 2022, a survey of 300 data professionals by Monte Carlo and Wakefield Research found that three out of four staff members needed over 4 hours to find a data quality problem, nine hours to resolve it, and that more than half of the respondents said that data quality has become worse over the past year. So not only is data quality a big problem and getting worse, but the costs of fixing data also add up to huge inefficiencies.
Why is Data Quality a Challenge?
If data quality is a pervasive issue with real consequences to business, why is it not more prominent in the headlines, such as when a data breach occurs? This is because:
- Data is an intangible asset and, unlike other assets, does not give signals that can be detected, such as changing color, giving off smoke, or changing smell. In fact, a single erroneous record normally can only be detected when someone who knows what the correct value should be notices the value is not correct in the data.
- It is not cost-effective to confirm the accuracy of every data record. It requires real-world observations or corroboration by another source. These are expensive endeavors.
- So much data is collected and processed so quickly now that bad records are not perceived as worth the effort to correct, as they will just be replaced soon.
- Unless the root cause of data quality issues is discovered and resolved, businesses will continue to admit poor data quality into their systems.
Another challenge to data quality is defining what it means for data to be fit for purpose. That definition can change not only across companies, but within business units of a single company as well. In general, in order for data to be of high quality, it needs to meet the following requirements – shown below by way of using zip codes, as an example:
|A data record presents what is found in reality without distortion.
|The ZIP code is the correct ZIP code for a customer.
|The data values present follow the correct format.
|A US-based ZIP code has five digits with no number or special characters.
|A data record has no missing values where values are mandatory.
|A ZIP code is present for all customer address records.
|There are no duplicate data records for the same entity or event.
|The list of valid ZIP codes in a system presents each ZIP code only once.
|There are no contradictions within a data record or across data records.
|A ZIP code is the correct ZIP code for the city and state in a customer record.
|The data record represents the most current known information.
|The ZIP code in a customer record exists for the customer’s current address, not the previous one.
|The parentage of the record can be traced so that the user knows from where the value was derived.
|The customer’s ZIP code was pulled from the sales database after a purchase was made last week.
Best Defense Against Poor Data Quality
Companies can take one of three stances regarding data quality:
- Do nothing. Consider poor data quality a cost of doing business and accept the inherent risk.
- Reactive remediation. When data quality problems are discovered—often too late to prevent a damaging business outcome—fix the data records and the root cause of the problem. Over time, data quality will improve.
- Proactive remediation. Pick the data records that are most critical to the business, usually meaning they are used by more than one business unit more than once. Define the data quality rules for those data records. Codify those rules into a dashboard that searches for violating data records and aggregates them into a scorecard. When a rule breaks an established threshold value, act quickly to fix the data records and the root cause of the problem.
Proactive remediation requires resources, but it can potentially find problems in the data and fix them before the data is used to inform a decision. Proactive remediation does not need to be resource-intensive, as the program can be focused on key data elements—the program does not need to extend to cover every single data element.
What Questions Should a Board Ask the Management Team?
Recognizing that no organization has perfect data, boards have a role to play in promoting data quality. As a board member, you can proactively probe what data policies and procedures your management team have established, whether they are being adhered to, and if they are effective at improving data quality:
- How is the trustworthiness of the company’s data determined? Unofficially, how do employees feel in general about the quality of their data?
- What data quality policies and protocols exist and how these are communicated, tested, and enforced from a compliance standpoint?
- How much time do staff spend cleaning data in preparation for a business analysis exercise?
- Does management know where the data quality problems lie? Do they know why the problems occur?
- What steps has management taken to detect poor data quality? When found, does management fix the data record, fix the problem at the source, or both?
Data Quality is at the Heart of Strong Governance
Data quality is a key component for good governance and effective oversight. Directors need accurate and timely information to make informed decisions about the company and its strategy. Poor data quality can distort decision-making, lead to missed opportunities and misguided investments, and cause compliance issues. Data quality is also critical for risk management, as poor quality can increase the risk of fraud, cyberattacks, and create other business disruptions.
Directors should support data quality programs that identify the company’s most critical data records, monitor those records for problems, and address problems at the source when they occur.
Next Steps for All
We invite you to explore additional resources of interest and educational programming via the BDO Center for Corporate Governance.
- What is Data Governance and Why is it Critical for your Organization
- The Board’s Role in Data Protection
- Data Governance in the Digital Transformation Age