Data is as ubiquitous as cash for many companies, but that doesn’t necessarily mean they all treat it with the level of care and attention it deserves. As a result, the quality and reliability of data can vary not only from one organization to the next, but also with across the enterprise.
For example, at many companies, the same business data elements (BDEs) — such as “customer name” and “customer address” — reside in multiple instances throughout the organization, with small inconsistencies in the names of fields and data formatting. Consider how one customer, ABC Technologies, Inc., might be presented in different databases and columns across multiple functional areas of a fictional organization:
In this example (which is actually quite typical), the data quality problems are clear and numerous. The same two data elements, customer name and address, appear in differently named fields in different repositories. In addition, there is variation in the way the data values appear in the fields. Because of these discrepancies, one can imagine the difficulty in trying to aggregate all the data about this customer — let alone generate any reliable reporting around it. And keep in mind, this is just one specific BDE, of which the typical company maintains hundreds or even thousands.
If these discrepancies seem trivial, consider that as a result, this organization may not be able to gain accurate insights into important company metrics. If asked to gather billing history about ABC Technologies, for example, a company analyst might find some of the records listed above, but certainly not all. The CFO might report inaccurate data to external stakeholders, over-count some key pieces of data in an investor report, or miss other data altogether.
How data quality management helps
Data quality management is the practice of understanding how data is created, modified and consumed across the organization, and improving its consistency and availability to those who need it. This is not just a nice-to-have improvement — it’s actually a way the organization can better position itself to serve its customers, outmaneuver the competition and increase profits.
Data quality management plays a role in a variety of financial analyses an organization must do:
Give data scientists and analysts access to better data sources — The ability to aggregate data on customers, revenue, profitability and other areas is essential to meaningful analysis. But when data quality is not well managed, the analysts may be drawing conclusions based on duplicated or incomplete information. The larger the organization and its data assets, the more likely there are data redundancies — or data that should be the same, but contains inconsistencies. Better data quality leads to better analyses, and better analyses lead to better decisions.
Reduce the time needed to implement data-intense initiatives — There are potentially many financially-focused activities that can benefit from quality data, ranging from implementation of CRMs, ERPs and other data repositories, to Hadoop/big data, data warehouses, BI solutions and more. It goes without saying that the better the quality of the data that feeds into such initiatives, the less time and effort will be needed comparing and cleaning up the data before implementation (and financial insights) can begin.
Find where data is lowering profits — On another level, data quality management can help the organization gain insights into its data as an asset class. For example, the process could help management better understand the economic contributions or costs of specific data assets to the firm’s P&L, balance sheet and cashflow. It’s not uncommon to find that certain data elements are costing a firm in the course of day-to-day operations, based on re-work and misuse caused by data quality problems.
Despite these and other compelling scenarios that illustrate the importance of data quality management, most organizations manage to get by despite their data quality. It’s the rare company that bothers to quantify the financial impact of its data quality problems, let alone build a business case for addressing them. Instead, they may spend their resources on technology problems that are perceived rather than real, without a clear sense of whether their efforts are making a difference.