Why doesn't the data we collect to run the organization answer our questions?
The data you collect to run the organization doesn't answer your workforce questions because it was designed for a different job. A payroll system records who was paid, how much, and for how many hours, and a scheduling system records who worked which shift, but neither was set up to tell you whether employees on one shift feel supported by their supervisor, or why a new hire left after six weeks.
Definitions are a common second problem. One system may count every departure as turnover, while another counts only resignations. The federal government's own survey of job openings and labor turnover, for example, separates quits (employees who left voluntarily) from layoffs and discharges and from other separations such as retirements and transfers. Your organization doesn't have to use those categories, but if your records lump them together, you can't tell whether people are choosing to leave.
A third problem is that the records that would answer a question together often live apart. For example, an exit date may sit in the HR system, the employee's shift in the scheduling system, and the supervisor's name in neither, which makes even a simple question about turnover by supervisor hard to answer.
What can our HR records tell us on their own?
On their own, your HR records can tell you a great deal about where and when something's happening, and much less about why. Used carefully, they're the right place to start, because they show you where to look before you ask anyone anything.
| HR records can usually tell you | HR records usually can't tell you |
|---|
| How many employees left, and when | Why they left |
| Which departments, shifts, facilities, or roles they left from | What working there was like for them |
| How long they had been with the organization | Whether a supervisor, a schedule, or a policy played a part |
| Patterns in overtime, absences, and open positions | How the employees who stayed are experiencing the extra work |
The left-hand column becomes much more useful once you break the totals apart. An organization can know that 20 percent of employees left last year without knowing whether they left because of pay, scheduling, supervision, workload, advancement opportunities, workplace culture, communication, job expectations, transportation, childcare, or something else entirely. Looking by group shows where to start asking. Turnover is an outcome, and the records can tell you where it's occurring before anyone can say what's producing it.
Why do we still need to talk to employees if we have the numbers?
You still need to talk to employees when you have the numbers because the numbers can show you a pattern and rarely explain it. Employees understand the day-to-day realities of their jobs, and supervisors see challenges that executives may not. People who stay can tell you what's working, and people who leave can identify the friction points that pushed them out.
Surveys, interviews, focus groups, HR records, and organizational measures become much more useful when they're used together. Your records might show that new employees on one shift leave within their first few months, and conversations with those employees and their supervisors may explain why. The goal is to identify patterns that leaders can act on, which is more than collecting employee opinions. How do I find out what my employees actually need? covers how to ask in a way that lets people answer honestly.
Where do we start with the HR data we already have?
Start with the HR data you already have by choosing one question your organization needs answered, ideally one connected to a change you could realistically make. Then gather the records that bear on that question, check how each one is counted, and break the totals apart by department, shift, role, and how long employees had been with the organization. Write down what the records show, what they don't, and what you'd need to learn from employees to fill the gap.
That last list tells you whether new information is needed, and what kind. Often the first useful step involves no new data at all. If you're weighing a larger decision on the strength of these records, We're about to make a big workforce decision. What should we know first? covers the questions to answer first.
If you're running a program with years of survey results from past grant cycles, the starting point is different. We have years of data and no idea what it tells us. Where do we start? covers that case.
The Grantibly perspective
Lauren's perspective
Administrative data are important, but numbers only provide part of the picture. Turnover numbers tell an organization what happened. They rarely explain why it happened, and an organization-wide total can hide important differences between departments, shifts, and groups of employees.
We don't believe organizations need more data simply for the sake of having more. Data collection should have a purpose, and that means starting with the question or problem before choosing a system or a survey. Data don't make decisions. People do, and our job is to help them turn information into understanding and action.
When we work with an organization's existing workforce data, we look for patterns across groups, compare what the records show with what employees and supervisors tell us, and look for strengths as well as problems. Understanding what makes employees stay can be just as important as understanding why they leave.
