Introduction to Workforce Metrics

It is important for Human Resources (HR) and child welfare leaders to start with a question-based mindset when analyzing workforce data (i.e., being thoughtful about what you and/or your stakeholders want to know), but it is also important to leave room to explore the data as well. This can only be accomplished if you know what is possible. This blog post describes some of what is possible to explore within the realm of workforce metrics as they relate to employee well-being, performance, and retention.

In many ways, the concept of workforce metrics is still in its infancy and many things are not yet standardized. Many inconsistencies can be found across the numerous metrics that currently exist. There are often different labels for the same formula and different formulas that go by the same label. Additionally, there are no standard benchmarks for many metrics, which can make it hard to know if your numbers are good or bad. Given this, it is important to embrace learning as you go and resist the urge to automatically compare your agency with others.

Key Terms

To lay the foundation for discussing workforce metrics, it’s important that we provide definitions for some key terms and challenges that are associated with them.

PERCENT: a portion of a whole, expressed as amount per hundred (e.g., a tip of 20%, a price reduction of 40%, or a sales commission of 18%). The challenge is that percents are sometimes incorrectly labeled as rates.

RATE: the ratio of two numbers, which can be in different units (e.g., speed–miles/hour, heart rate–beats/minute, gas mileage–miles/gallon) or in different categories of the same unit (e.g., women/men, stayers/leavers). With workforce analytics, people are generally conceptualized as the “unit” and what we are analyzing is usually the categories that people fall into. The challenge is that rates are sometimes incorrectly labeled with the percent sign. Turnover rate is a common example; although we attach a percent sign after it, it is actually a rate since it is not a portion of a whole.

PERCENTAGE POINTS: the numerical difference between two percentages. An example of this would be a turnover rate that drops from 32% to 28%, which would be a reduction of 4 percentage points. This is often incorrectly described as a reduction of 4%; instead, a reduction of 4% would be a drop from 32% to approximately 31%. The actual percent reduction in this example of 4 percentage points would be a decrease of approximately 13%.

Workforce Outcomes: Performance

There are numerous ways to conceptualize individual-level performance, but two common ones are training and job performance. There are a variety of ways to measure training performance, including knowledge test scores, skills evaluation scores, results of field observations, and assignment completion (e.g., % of completed versus missing assignments or on-time versus late). Most child welfare agencies have a fairly significant initial training, and with good evaluation data, agencies can use training performance as an early indicator or precursor of job performance.

Another dimension to consider is individual versus aggregate (e.g., team, unit, department, and agency) performance.  When considering which of these to examine, keep in mind that they each serve different purposes. For example, you may look at individual performance when assessing whether an employee is eligible for a promotion and aggregate performance when evaluating the extent to which a specific group collectively achieved some specified goal.

It may also be helpful to conceptualize performance in terms of how it is measured, be it objectively or subjectively. Subjective measures are typically supervisor judgments about things such as performance of job duties or demonstration of competencies. Objective measures could include data elements in your child welfare information system (e.g., timeliness, task completion, disciplinary actions). A final dimension to consider regarding performance is process versus outcome data. Outcome data is information that can speak to the achievement of certain specified goals, whereas process data speaks to the steps that are designed to lead (either positively or negatively) to a particular goal. Having outcome data is ideal, but if you are unable to get to that, process measures can serve as precursors.

Workforce Outcomes: Retention

Retention is a workforce outcome that gets a lot of interest and attention. Agencies often examine retention by looking at their aggregate turnover rate, which = (the number of employees who left/number of employees) x 100. Though it is described in terms of a percent (e.g., 25% turnover rate), turnover is not actually a percentage in the customary sense (i.e., portion of a whole). That is why it can exceed 100%.

Breaking that turnover metric down a bit, we must acknowledge that there are many types of movement that could be considered “leaving”—leaving the position, the job, the office, the region, the division, the agency, etc. One thing to consider when looking at turnover is whether it was internal (also known as “churn” or “churnover”) or external (i.e., leaving the agency altogether). Agencies should carefully consider the type of movement that matters most for their purpose. Even though internal turnover can be a good thing, it still has costs and consequences that should be considered and examined. You can also look at turnover in terms of whether it was voluntary (i.e., employee initiated) or involuntary (i.e., agency initiated), avoidable (i.e., employee departure could have been prevented by the agency) or unavoidable (i.e., employee departure could not have been prevented by the agency), and functional (i.e., a benefit to the agency, often because the person who left was a poor performer) or dysfunctional (i.e., a loss to the agency, often because the person who left was a good performer).

Additionally, there are a number of ways to define “number of employees” (i.e., headcount) in the turnover equation such as the number at the beginning of the relevant time period (e.g., month, year), the number at the end of a relevant time period, the average of two days in time period, and the average across all days in a time period. A turnover rate should not be interpreted without knowing the time span it covers. Unless the number of employees overall changes drastically, the turnover rate will naturally increase across time as more people leave. Therefore, turnover rates should not be compared unless they are for the same time frame. For example, assuming steady conditions, a turnover rate over a 6-month span is likely to be about half the turnover rate over a 12-month span. Additionally, turnover rates for time periods that are less than one year can be annualized, as a means of standardizing or forecasting.

Further details on the metrics discussed in this blog (and additional metrics) can be found at the QIC-WD’s Institute page.