Data Quality: Can I Trust My Client Data?

"Can I trust my client data?" 

This is a common question asked by the data professionals, risk officers, audit groups, sales teams, and more throughout our clients' organizations. Unfortunately, it's a question that is rarely asked until backs are against the wall. When you have a regulatory deadline to meet or a critical risk-based decision to make, you need to be confident in the quality of your client data.

To generate that confidence, data quality needs to be a constant, consistent focus of your data management strategy. Data quality needs to be evaluated, monitored, and acted upon every single day. Only then can you confidently answer that lingering question:

"Can I trust my client data?"

At Kingland, we have a 25 year track record of providing data management technology and solutions, and in that time, we've observed a multitude of data quality strategies. The most successful strategies always employ a comprehensive set of rules supported by real-time validation scans. Rules are commonly configured to improve data quality across the four primary data diagnostic areas: completeness, conformity, consistency, and duplication. Let me take a moment to describe each diagnostic area and how the Kingland Platform identifies, analyzes, and corrects data quality issues within your data set. 

• Completeness - Are there gaps in the data or missing values where values are expected? Does your client data have a first name but no last name or a city but no country? This could be caused by incomplete data integration - maybe an incorrect mapping during data import. It could also be caused by insufficient data validation checks upon data entry. And while it's not too difficult to determine where data is missing, correcting the incomplete data can be incredibly challenging and could require manual research, data extraction, and data entry.
• Conformity - Ensure data is aligned to predetermined standards. This could be rules such as states are always two-letter abbreviations or U.S. postal codes are always 5 or 9 digits. This is usually caused by free-form fields and insufficient data validation checks. Standards exist at a varying level of complexity and are completely dependent on the use case of your data set.
• Consistency - Check for inter-record inconsistencies within the data set. This is most common with hierarchy data. For example, a rudimentary data management system may allow a parent record to not be present as a base record. This tends to occur as data changes over time or additional data sources are integrated. As your data set grows, it can become incredibly challenging to maintain your data consistency. 
• Duplication - A duplicate occurs when two separate records represent the same 'thing' (legal entity, natural person, agreement, etc.). Occasionally, these are very easy to detect - especially when two records have the same name. For example, two separate records with the name Blackrock, Inc. are an obvious duplicate to identify. But duplication is much more complex than that. The name of one or both records could contain a typo ('Blckrokc, Inc.') or an elongated legal structure ('Blackrock, Incorporated') and still represent the same entity. Duplicates tend to be generated by manual data entry activities or after the integration of additional data sources and are a major cause of inaccurate analysis and unrecognized risk.

Using spreadsheets filled with formulas may work well enough with smaller data sets, but as data sets begin to grow, this quickly proves to be an unsustainable approach. As additional sources are integrated, processes evolve, data changes are implemented, and manually-entered inaccuracies occur, a manual approach to data quality can prove to be overwhelming, particularly with the nuances of client data. Data quality will deteriorate rapidly without a controlled, systemic rules approach for completeness, conformity, consistency, and duplication.  

So continue to ask yourself, "Can I trust my client data?"

And when your client data set has grown to the point that you can no longer confidently answer, consider the Kingland Platform's intuitive Rules, Workflows, Dashboards, and more features which support the data quality and data management of enterprise-sized data sets.