When implementing a new Human Resource (HR) system, upgrading your organization’s existing system or simply reviewing your strategy for your current system, what does your organization focus on? Does data governance even make your top five-priority list? Often, data governance is an afterthought rather than being included in the initial process design.
HR data should be an enabler to an organization. Yet, often times, it becomes so messy that it becomes an inhibitor. Do you cringe when someone requests reports from your department because you know you will be solving a gigantic puzzle with missing pieces? When data is passed to downstream systems, do you spend time correcting issues due to bad or missing data? Or, are you confident that all data is defined appropriately, is a completed work of art and passes flawlessly to any system that requires it? If the former is true, then I suggest updating your existing HR data governance practices or implementing policies if they do not exist.
Data governance should boil down to three basic concepts:
- How your organization makes decisions about data
- Data definitions that span across the entire organization
- A go forward methodology on how to execute processes and reports that involve Human Resources data
Making Decisions About Data
Let’s examine how an organization might make decisions about data, and not just for protection and compliance. A good starting point is to outline which systems contain data, determine what the systems of record are and review how data flows throughout the organization.
Data quality issues often are discovered when reports are produced or downstream processes and systems are found to have errors. It is important to define the root cause of these issues and develop standards to avoid or mitigate them as soon as possible.
It is also important to review access rights for your organization’s data. Clear rules should be created for employee access to data. It is the duty of the organization to protect employee data, as it should be accessible only on a need-to-know basis. To do this, designate a specific role or team that is tasked with controlling and monitoring HR data. This designation should also include guarding against breaches of security.
Spanning Data Definitions
If you execute a Google search for “definition of data dictionary” it will say it is “a set of information describing the contents, format, and structure of a database and the relationship between its elements, used to control access to and manipulation of the database.” Today, organizations face multiple, contrasting, transactional and operational systems within the HR ecosystem. It is critical to define what each individual data element means and provide a “rulebook” that provides a common language across all systems.
These definitions will need to be communicated to the user population, which will eliminate confusion behind naming conventions and field labels for data elements that each individual system utilizes. This step will create a common organizational language, regardless of the system being used or analyzed.
Data Governance Going Forward
Additionally, it is also important for your organization to implement a go forward methodology on how to execute processes and reports that involve Human Resources data. Best practices suggest pushing quality audits to the end user at the beginning of the process. The sooner this can be accomplished, the more the initial steps can be controlled. This shift will prevent lengthy audits at the end of the process and spare cleanup tasks. The governance structure will also ensure that better decisions are being made from more reliable sets of data.
Data governance isn’t an easy task for organizations to feel confident enough that they have a good understanding of the complex data systems. However, by making these strides, it will be easier to get a clear picture of the current state and to make meaningful projections about the future. Don’t let your organization get stuck in the “now” with the aftermath of incomplete, disconnected, inaccurate and difficult to aggregate data.