I spent a few days last month at the MIT CDOIQ Symposium a gathering of mainly chief data officers, data scientist types, and academicians. It was inspiring to be in a room of really smart people who are all trying to help organizations make better decisions with data. Much of the conversation centered around data quality and data management issues. Just about every organization involved in analytics faces these challenges – and it can seem daunting, especially in large organizations with many disparate systems and silo’ed analytics groups.
I have talked with many HR leaders who are struggling to get past the stage where their stakeholders are constantly challenging them on data, asking “Is this data right?” instead of the more important questions of “What does this data mean?” and “What should I do about it?” As a first step, HR analytics team can start by prioritizing their most important metrics based on what decisions their organizations are trying to make. HR teams may have hundreds of data fields in their systems, but not all will be critical. Early on it will be important to conduct an audience analysis: identifying the target stakeholders and the problems they are trying to solve –and what data is needed to solve them. The analytics team must understand the needs of each audience and select a set of the most important metrics. Of course, the required data elements will vary by audience and by the type of analytics project.
To understand the drivers of employee engagement, for example, you can take data from your engagement survey and employee data such as job role, tenure, location, demographics, and compensation. You can also pull data on each employee’s supervisor, their tenure, and characteristics. Coupled with the HR data will be operational and financial data from their business unit or location, such as revenue, profitability, productivity, safety incidents, worker compensation claims, theft or leakage, and customer satisfaction data. Many types of analysis will also require external data such as economic indices, labor statistics, market compensation, and industry benchmarks.
So from your mass of data, you will need to prioritize which elements will be most critical. Then the process begins of improving the quality of this data and setting up a longer-term data governance model. We have written a report on data quality and governance with more details.
Figure 1. (Click here to view our infographic.)