Seven Steps to Effective Data Governance in Financial Services
By Jon M. Deutsch, Vice President Financial Services, Information Builders
Data governance can help financial services organizations meet core requirements for financial reporting, regulatory compliance, and privacy policies. By defining and implementing a set of rules for the rights, accountabilities, and processes for taking action with information, companies can ensure the confidentiality, quality, availability and integrity of the information. This is particularly important because data quality problems absolutely plague banks and other financial services firms; for example, studies show that up to 25 percent of data in an average bank’s Customer Information File/Customer Information System is erroneous because of keying errors, information duplication, or incorrect transformations between systems.
Implementing governance rules can be complex, but companies can avoid bureaucratic delays – and the paralysis that sometimes attends overwhelming tasks – by taking an incremental approach. The following seven steps provide a clear mechanism for making steady progress without having to boil the ocean.
1. Prioritize areas for business improvement.
An incremental approach helps to avoid one of the biggest historical problems with data governance: lack of follow-through. Target a single area, such as marketing, and work with the existing organizational structure to take action and ensure accountability. Align information objectives clearly with business strategy, get stakeholders to achieve consensus about requirements, and identify key data entities so that policies about them can be outlined. For instance, marketers may want to use cleaner demographic and geographic information to increase their effectiveness in targeting potential clients with specific risk profiles. Projects like this help keep attention focused, which would be much more difficult with a large-scope project.
2. Maximize availability of information assets.
This may seem contradictory: How do you govern information better by making more of it more widely available? The availability of different forms of information – EDI transactions, data warehouses, CRM and ERP applications, legacy file structures, etc. – helps get stakeholders to understand the relationships of various pieces of information and how it all should be managed. For instance, imagine an employee pulling out an iPad to call up the record of a recent sale to find out why it’s not showing up on a report – and then think about how that could affect stakeholder attitudes toward real-time data cleansing.
3. Create roles, responsibilities, and rules.
Data governance is a partnership between business people, who know the data, and IT professionals, who understand how it can be manipulated. Profile the data to identify incorrect or inconsistent data elements. Analyze the impact of bad data on your organization and provide suggestions, or cleansing rules, as to what the data should look like. Pass the cleansing rules to IT professionals so that they can apply technology to cleanse the data based on the business professionals’ suggestions.
4. Improve and ensure information asset integrity.
Don’t make the mistake of considering data governance to be a one-time, set-it-and-forget-it system. Continuously improve and ensure the integrity of information assets in a four-step process: data profiling, parsing and standardization, data enrichment, and data monitoring. For example, profiling shows you the variation in your data; parsing and standardization makes sure the variation occurs in easily predictable ways; enrichment fleshes out information to make sure everyone who uses it has the proper context and detail; and monitoring ensures that bad data doesn’t have a chance to be spread from system to system. Together, these steps minimize the amount of bad data in the enterprise and mitigate the risk and possible damage done by any that happens to creep in.
5. Establish an accountability infrastructure.
Even with all of the processes in place to ensure information integrity, some questions will linger: What happens if the information is still inaccurate? What happens to those data elements that fall through the cracks of the automated processes? What if I want to make sure the changes are right before they are applied? Processes alone do not ensure the integrity of information. People do. Establish an accountability infrastructure that holds people accountable for information assets, and provide them with the technology and metrics they need to ensure the integrity of the assets remains high.
6. Convert to a master data-based culture.
Most organizations today focus on transactional data. In reality, though, individual transactions matter less than overall relationships; master data focuses holistically on the essential facts that define a business, which helps raise relationships to the position they deserve. For instance, in a master data-based culture, a discussion about a particular invoice elevates to a discussion about the customer. Master data exists everywhere in an organization – in different applications, systems, transactions, data warehouses, and messages. Master data management (MDM) decouples master data from those individual source applications and ensures consistent master data across transactional and analytical systems. Everyone sees the same information, providing one acceptable version of reality.
7. Develop a feedback mechanism for process improvement.
Everything discussed so far is part of a cycle, and there is always room for improvement. Build a feedback mechanism into the process that allows for continual process improvement. Monitoring information assets over time gives a clear picture of how initiatives are performing and provides a way to graphically depict both successes and failures in the process.
From a process improvement perspective, compliance standards indicate that companies will need to audit the processes that capture, cleanse, and manage information. While building their compliance processes, they should be careful to look beyond basic compliance issues and build in a feedback mechanism that helps them continually improve their processes. For instance, if a bank designs a given data governance initiative to improve services to minority-owned businesses, and the company has licensed information from a data provider to augment marketing records with demographic information, there should be a feedback mechanism that shows what percentage of records are being properly augmented with minority-owned status; if the percentage is too low, the company may need to research additional third-party data providers. Feedback mechanisms of this sort are essential to prove business value, improve processes, and keep efforts focused on areas that need improvement.
By following these steps on an iterative basis, your organization can take control of its information and begin to approach data governance in a way that increases oversight, reduces business and compliance risk, and improves business processes – all at the same time.