Measuring psychosocial risk (without another annual survey)

Key takeaways

• An annual survey gives you one photograph a year of something that moves every week. By the time you act on it, the picture is old.
• Good measurement isn't one org-wide number. It's broken down by team and role, because that's where hazards actually live.
• The free government People at Work assessment is a solid, neutral baseline. Continuous measurement is what you build on top of it, not instead of it.
• Continuous data does double duty. It tells you where to act, and it becomes the record that shows you reviewed your controls, which the Code expects.
• If your only measurement tool is a survey that lands in everyone's inbox once a year, you're not managing psychosocial risk. You're taking its temperature occasionally.

A leader said something to us recently that stuck. "We ran the engagement survey in March. It's August now, and I genuinely don't know if anything we did in response actually worked."

That's not a failure of effort. It's a failure of the tool. An annual survey was never built to answer that question, and measurement is only one part of managing psychosocial safety well.

Why the once-a-year snapshot doesn't work

Psychosocial risk doesn't sit still. A team can go from fine to overloaded in a quarter, because a project blew out, someone left and wasn't replaced, or a manager changed and took the culture with them. If your only read happens once a year, you find out about the overload months after it started and months after it's done real damage.

There's a second problem, which is response blindness. Say your March survey flags high role overload in customer support. You put in a control, maybe you hire, maybe you redistribute the workload. Good. But your next data point is March next year. You have no idea, for twelve months, whether that control actually worked or whether it quietly made things worse.

The Code of Practice expects you to review your controls, not just install them and hope. Reviewing on a twelve-month lag isn't really reviewing. It's guessing with extra steps. See what the Code of Practice expects, in plain English.

What good measurement actually looks like

Good psychosocial measurement has three things an annual engagement survey usually doesn't.

It's broken down by team and role, not one org-wide number. A single company-wide wellbeing score is close to useless for action. If your average masks a burnt-out operations team sitting next to a comfortable marketing team, the average tells you nothing you can act on. You need the data at the level where the hazard actually lives, which is the team, sometimes the individual manager.

It combines the hazards, not just one of them. Role overload, low job control, poor support, exposure to remote or isolated work, poorly managed change. These interact. A team with high workload and strong support can hold up fine. The same workload with a manager nobody trusts is a different risk entirely. Measuring engagement alone, or burnout alone, misses that interaction. For the full list of what you're actually looking for, see the full list of psychosocial hazards.

It tracks your response, not just the baseline. The point of measuring isn't the score. It's what changes after you act on it. Good measurement closes the loop: here's the hazard, here's what we did, here's whether it moved.

None of that requires more surveys. It requires shorter, more frequent, better-targeted ones, and a system that actually connects the dots between them.

The free baseline, and what it doesn't cover

If you haven't started measuring at all, don't overthink the first step. The People at Work assessment, backed by Comcare and Safe Work Australia, is free, validated, and a genuinely good place to start. It's built on solid psychometric ground, alongside other validated instruments like APHIRM and COPSOQ III that Australian and international regulators reference.

Here's the honest limitation. People at Work, like most validated survey instruments, is designed to be run periodically, not continuously. It's a proper baseline, not an ongoing operating system. It'll tell you where you stand today. It won't tell you whether the change you made last month is working, or flag a team that's sliding before it shows up in the next full run.

That's the gap continuous software fills. Not by replacing a validated instrument, but by running lighter, more frequent reads between the formal assessments, and connecting them so you can actually see a trend instead of a single point. It's the difference between an annual physical and knowing your resting heart rate every day. Both matter. They're not the same thing. We've laid out that contrast in more detail in annual survey versus continuous listening.

Continuous data does double duty

Here's the part that's easy to miss if you only think about measurement as a wellbeing exercise. It's also your evidence.

If a regulator or a court ever asks whether you took reasonable steps, "we ran a survey once, eighteen months ago" is a weak answer. "Here's our data by team over the last two years, here's where we identified rising risk, here's the control we put in, here's the follow-up data showing it worked" is a genuinely strong one. Duty of care isn't just about doing the right thing. It's about being able to show you did it, which we've written about in duty of care and what leaders now own.

Continuous measurement, done properly, builds that record as a byproduct of doing the work well. You're not creating a compliance file separately from managing risk. The record is the same thing as managing risk.

What this looks like in practice

In practice, most teams that get this right run something like:

• A short, frequent pulse at the team level, not a company-wide once-a-year event.
• A periodic deeper read, using a validated instrument like People at Work, to check the pulse data against a proper baseline.
• Trend tracking that flags when a team's risk is moving in the wrong direction, before it shows up as resignations or a formal complaint.
• A simple loop back to whoever owns that team, so the person closest to the hazard is the one acting on it.

That's the continuous listening approach we've built into the Culture module at Whyser Work. It's not a bigger survey. It's a different shape of measurement, built for how risk actually moves.

Frequently asked questions

Should I stop running an annual survey altogether?

Not necessarily. A periodic, validated instrument like People at Work is still worth running as a proper baseline. The problem isn't the annual survey existing, it's relying on it as your only source of truth for something that changes far more often than once a year.

How often should continuous measurement actually happen?

There's no single right cadence, and it depends on team size and risk level. What matters more than the exact frequency is that it's regular enough to catch a shift before it becomes a resignation, and specific enough at the team level to point you at where to act.

Is continuous measurement more expensive than an annual survey?

It's a different kind of investment, not necessarily a bigger one. A well-run continuous system replaces a lot of the reactive cost that shows up later, exit interviews, replacement hiring, and the time spent firefighting a problem that was visible months earlier if anyone had been looking.

Does continuous data actually count as evidence?

Yes, provided it's genuine, ongoing, and connected to what you did in response. A record that shows you identified a hazard, acted, and reviewed the outcome is exactly the kind of evidence the Code of Practice points to. This is general guidance to help you understand your obligations. It supports your duty of care, it doesn't replace your own legal and WHS advice.

So here's the question worth sitting with. If a regulator asked you today what you know about how your teams are actually going, right now, not eight months ago, what would you show them?