Why the seminar did not work
The room liked it. Attendance was strong. Six months later the numbers had not moved. This is the most common outcome in workplace wellbeing, and it is not a failure of the session.
An Oxford study of more than 46,000 employees across 233 organisations found that individual-level workplace wellbeing interventions produced no measurable improvement in psychological wellbeing (Fleming, 2024, Industrial Relations Journal).
Mindfulness. Resilience training. Stress management apps. Coaching. Wellbeing seminars.
That finding gets used as an attack, and I do not use it that way, because it is not one. It is a statement about instruments, not about the people who chose them or the people who delivered them. A well-designed session, well run, in front of an engaged room, can still produce no measurable change if what is producing the problem sits at a level the session cannot reach.
Structural drivers require structural responses. That is the whole of it.
What “level” means here
An individual-level intervention changes a person’s capacity to cope with a condition. A structural intervention changes the condition.
Both are legitimate. They are simply answers to different questions, and the trouble starts when one is bought as though it were the other.
If someone is under sustained financial pressure - a real, ongoing, external condition of their life - then a session that improves their coping capacity is aimed at the wrong variable. It might genuinely help them feel better for a fortnight. It has not touched the pressure, so the pressure is still there in month three, doing what it was doing in month one.
This is why the evaluation looks so strange in hindsight. The session evaluates beautifully on satisfaction and terribly on outcome, and both results are accurate.
The regulator has taken the same position
You can read the same logic in Australian regulation, which is a useful thing to know if you are the one who has to defend the spend.
Since August 2025, the NSW WHS Regulation has required psychosocial risk to be managed through the hierarchy of control measures. Training, employee assistance and complaint procedures are not excluded, but they sit near the bottom of that hierarchy. Above them are the things that change the work itself: workload, resourcing, role clarity, job design, consultation.
SafeWork NSW’s Regulatory Statement 2026-27 directs its psychosocial focus toward high-risk workplaces and occupations and the use of higher-order controls.
Victoria arrived at the same place from a different direction. Its Occupational Health and Safety (Psychological Health) Regulations 2025, in force since December 2025, permit information, instruction or training to be the predominant control only where altering the work itself is not reasonably practicable.
The expectation is not that organisations should stop running sessions. It is that a session should not be counted as the control.
What actually went wrong, in order
The problem was named at the wrong level. It arrived as “our people are stressed”, which is a description of a symptom that has at least six plausible drivers, most of which are structural and none of which a seminar changes.
So the remedy was chosen at the level of the description. Stress, therefore stress management. It is a perfectly rational inference from a badly specified problem.
And then it was evaluated at the level of the remedy. Attendance, satisfaction, self-reported confidence. Every one of those measures is real, and none of them is the outcome the money was spent to buy.
Meanwhile nobody established what the driver was costing. Which is why, when it came time to renew, there was no number to argue with - only a feeling that it had been nice and had not changed much.
The version that works
Establish the driver before you choose the level. If a material share of the distress in your workforce is being generated by something structural, then the response has to reach the structure, and the individual-level work becomes a supplement rather than the answer.
Then price it. A driver with a dollar figure against it gets a structural response. A driver without one gets a session, because a session is what fits in the budget line that was already approved.
That is not cynicism about executives. It is how funding decisions get made everywhere, and it is entirely reasonable when the alternative arrives with no number attached.
Put a figure against the layer beneath the program
The Missing Line Item is a ten-minute worksheet that produces a defensible dollar figure for what financial pressure is costing your organisation. Your headcount, your own figures, published Australian research for the parts you cannot see. Every source dated. No email required.
