The consistency and accuracy of employment records is a critical first step in any workforce analytic analysis or review. With over 20 years of experience working with contractor data, I’ve seen a lot of data exports from HRIS and payroll systems over the years. Employer recordkeeping has improved dramatically in the last two decades, but it is still common to find inaccuracies in employment data.
Having a system that accurately captures point in time data is critical to completing accurate analyses. Let’s say we wanted to identify starting pay and current pay for certain employees, for the purposes of comparing salary growth - can your system do that? Accurate point in time data also comes into play when analyzing selection data. Sometimes I see the new hire job title changed to the current job title when a promotion or a job change happens. For example, Jane is hired as an Admin, and then promoted to an Analyst, her new hire record data will reflect her current job title of Analyst, not the Admin job title she was actually hired into. This could affect accuracy in any kind of analysis run on the hiring data. If the HRIS data that gets exported for workforce analytics does not capture historical movement correctly, any analyses run on the data risks being flawed.
Another common data disconnect that I see is when an applicant tracking system uses a different job title than the HRIS that houses the new hire data. For example, the applicant tracking system might capture the ‘posting title’ for the job, but this is different than the job title the employee is hired into. This can be a recordkeeping concern if there is not a way to link the applicant pools with the actual selection in the new hires. How would a company know if there is selection bias, if we can’t identify the candidates that applied to the hired job? Using requisitions in the applicant data and the new hires data allows that link to be formed, even if the job titles used in the systems are not the same. Any analysis done on applicant selection activity is dependent on accurate, consistent data, and processes.
Inconsistent use of termination reasons or having vague termination reasons is another data accuracy concern. Can the company easily identify voluntary or involuntary terminations? If workforce analytics are run on involuntary terminations, is that the accurate list? Companies should periodically review their termination reasons to ensure that they are specific and accurately classified as voluntary or involuntary.
Workforce analytics are only as accurate as the data used to create them. Ensuring your systems generate accurate and complete data should be the first step in creating useful workforce analytics.