Workforce Analytics
Workforce analytics is the practice of analyzing aggregated data about how work gets done, in order to make decisions about staffing, capacity, workload and process. Its subject is the work and the system around it, not the individual.
That last clause is the whole definition in practice, and it is also where the term is most often stretched. A good deal of what is sold as workforce analytics is individual activity monitoring with a broader label.
Workforce, people and HR analytics
These three terms are used inconsistently, including by the vendors selling them, so any confident-sounding taxonomy should be treated with some caution. The distinction that survives contact is one of subject matter.
| Term | Subject | Typical question |
|---|---|---|
| HR analytics | HR processes | How long does hiring take, and where do candidates drop out? |
| People analytics | People and their effect on outcomes | Which factors predict retention in this population? |
| Workforce analytics | Work, capacity and staffing | Where does this team's time actually go, and can it absorb this project? |
In practice the boundaries overlap and few organizations use the labels precisely. If you are evaluating a product, ignore which of the three words is on the box and ask what unit the data is about: a process, a population, or a named individual.
The questions it is good at
Workforce analytics earns its place on questions that are about the system rather than the person, and that a manager could not answer reliably from memory.
- Where does the team's time actually go, by category of work?
- How much of the week is consumed by meetings, and has that changed?
- What is our real capacity for the next period, after leave, support load and recurring commitments? Our page on capacity planning covers the arithmetic.
- Where does work wait, and for how long, between the steps of a process?
- Which teams carry a support burden that is not reflected in their commitments?
Notice that every one of these is answerable from aggregated data, and that acting on any of them changes a schedule, a process or a staffing level rather than an individual's standing. That is the test.
Where it stops being analytics
The same data can support either practice. What separates them is the unit of analysis and what the output is used for.
| Workforce analytics | Individual monitoring |
|---|---|
| Aggregated to a team or process | Attributed to a named person |
| Answers a question decided in advance | Reported routinely, then questions are found for it |
| Changes a schedule, process or headcount | Changes how someone is judged |
| Would survive being shown to the team | Usually is not shown to the team |
The last row is the most reliable single test. If the analysis could be presented to the people it describes without difficulty, it is analytics. If it could not, the label is doing work the practice does not support.
This matters legally as well as ethically. Once the data is attributed to individuals, the monitoring notice statutes engage: four states now require written notice before electronic monitoring begins, as our guide to monitoring laws by state sets out. Our page on workplace analytics without turning it into surveillance covers the design rules.
Why most implementations disappoint
Three failures account for most of it.
The data arrives with no question attached. A dashboard built because the tool could produce one gets used for whatever it happens to reveal, which is how analytics becomes surveillance without anyone deciding.
The measures do not constrain each other. A single headline number can be improved at the expense of something unmeasured. Several measures that pull against each other make that visible, which is the property our performance tracking guide covers.
Nobody can act on the output. If the finding is that a team is overloaded and no one is empowered to change scope or headcount, the analysis was an expensive way to document a known problem.
Key points
- Workforce analytics analyses aggregated data about how work gets done, to inform staffing, capacity and process decisions.
- Its subject is the work and the system, not the individual. That is the definition doing the work.
- The three terms are used inconsistently, including by vendors. Ask what unit the data is about rather than which word is on the box.
- It suits questions about where time goes, real capacity, meeting load and where work waits.
- The reliable test: if the analysis could be shown to the people it describes without difficulty, it is analytics.
- Once data is attributed to named individuals, monitoring notice statutes engage in four states.
- Most implementations fail because the data arrived with no question attached, or because nobody can act on the answer.
Frequently asked questions
What is workforce analytics?
The analysis of aggregated workforce data to understand how work gets done and to inform decisions about staffing, capacity, workload and process. It is concerned with the work and the system around it rather than with assessing individuals.
What is the difference between workforce analytics and people analytics?
Broadly, workforce analytics is about work, capacity and staffing, while people analytics is about people and their effect on business outcomes. In practice the terms are used inconsistently, including by vendors, so it is more useful to ask what unit the data is about: a process, a population, or a named individual.
Is workforce analytics the same as employee monitoring?
It does not have to be, and the distinction is the unit of analysis. Aggregated to a team or a process, answering a question decided in advance, and used to change a schedule or a staffing level, it is analytics. Attributed to named individuals and used to judge them, it is monitoring, whatever it is called.
What questions is workforce analytics good at answering?
Where a team's time actually goes by category, how much of the week meetings consume, what real capacity is after leave and support load, where work waits between process steps, and which teams carry an unacknowledged support burden. All are answerable from aggregated data.
Does workforce analytics require employee notice?
It depends on what is collected. Aggregated analysis of process data generally does not engage the monitoring notice statutes. Once activity is attributed to identifiable individuals it is electronic monitoring, and four states now require written notice before it begins, with Maine requiring notice to be repeated annually.
Why do workforce analytics projects fail?
Usually because the data arrived without a question attached, so the dashboard gets used for whatever it happens to reveal. The other two common causes are single headline measures that can be improved at the expense of something unmeasured, and findings nobody is empowered to act on.
What is the simplest test of whether it has gone too far?
Whether the analysis could be shown to the people it describes without difficulty. If it could, it is analytics. If it could not, the label is doing work the practice does not support.