Mental health at work, beyond the wellness app

We bought a tool that was never designed to fix the problem.

A friend who runs his own firm mentioned recently that he was looking into wellness app subscriptions for his employees. Headspace, Calm, something in that space. He genuinely cares about his team, and the gesture was coming from a good place.

I didn’t say much in the moment. The conversation stayed with me, because I have watched the same impulse play out across organisations of every size for years, and the result is almost always the same.

Companies have spent enormous amounts on workplace mental health over the last decade. Meditation apps. Mental health days. Resilience training. Webinars on managing stress. Subscriptions to coaching platforms. The corporate wellness market is now worth around 55 billion dollars globally.

And yet, by every measure, things are getting worse. Gallup’s 2025 State of the Global Workplace report puts employee engagement at 20 per cent worldwide, the lowest since the pandemic, with disengagement costing the global economy 438 billion dollars a year. Workplace wellness was the only one of eleven wellness sectors to shrink in 2024, by 1.5 per cent. The product category designed to fix the problem is the only category losing money.

So either every company buying these programmes is uniquely bad at procurement, which seems unlikely, or the solution itself was never designed to do what we asked of it.

It is the second one. The friend buying subscriptions for his team is not wrong to care. He is just about to spend money on the wrong thing.

The strongest evidence we have

In 2024, an Oxford researcher named William Fleming published a study in the Industrial Relations Journal that should have changed the conversation. He looked at survey data from 46,336 workers across 233 organisations in the UK. He compared people who used their employer’s wellness programmes (mindfulness training, stress management classes, resilience apps, sleep programmes) with people at the same company who did not.

The result: no measurable benefit.

Stress, job satisfaction, sense of belonging, mental health. None of them improved for the people who used the programmes. The only intervention that showed a positive effect was charity and volunteering work, which is not really a wellness intervention at all.

Fleming’s conclusion was blunt. Organisations have to change the workplace, not just the worker.

That single sentence is the whole argument.

Why the wellness app was always going to lose

There is a model in occupational health called Job Demands-Resources. It has been validated across thousands of studies and the idea is simple. Every job has demands (workload, deadlines, emotional labour, conflict) and resources (autonomy, support, fair pay, decent management, time to recover). When demands consistently outweigh resources, people burn out. Not metaphorically. Measurably.

Now ask what a meditation app does to that equation.

It does not reduce your workload. It does not give you more autonomy. It does not make your manager listen to you. It does not fix the fact that three people left and were not replaced. It does not shorten the deadline. It does not change how performance is reviewed. It just teaches you to breathe more calmly while the same forces continue to grind.

It is like giving someone a better umbrella while ignoring the leaking roof. The umbrella is real. The umbrella works, in a narrow sense. But it was never going to dry the house.

That is not a knock on meditation. Mindfulness, used personally, has real evidence behind it. The problem is what happens when an organisation buys it as the response to an organisational issue. The category error is the problem, not the product.

Why companies bought the umbrella anyway

This is the part worth being honest about, because the wellness industry did not con anyone. Companies bought what they wanted to buy.

A wellness app is cheap, fast, and visible. You can announce it on a Monday. It generates a press release. It gets a tile on the careers page. It can be measured in download counts and login rates, which look like progress on a dashboard.

Fixing the actual environment is none of those things. It is slow. It is expensive. It usually requires telling someone senior that the way they run their team is part of the problem. It involves redesigning workloads, retraining managers, changing incentives, and sometimes hiring more people. There is no clean launch date and no logo to put on the slide.

Given the choice between a hard fix and a visible one, organisations under pressure will pick the visible one almost every time. That is not cynicism. It is how budgets and quarterly reviews work.

The wellness industry did not create that incentive. It just sold into it.

A way to tell the difference

If you want a quick test for whether your company is treating the symptom or the cause, look at where the mental health budget actually goes. If most of it is spent on things employees do alone, in their own time, on their own phone, you have a symptom-treatment programme. If a meaningful share of it is spent on changing how work is structured, you have something closer to a real intervention.

A more concrete checklist, drawn from what the research consistently identifies as the real drivers:

• Are workloads reviewed and adjusted, or are they assumed to be fine until someone breaks?
• Do people have genuine control over how and when they do their work, or only the appearance of it?
• Are managers trained to manage humans, or only to hit numbers?
• When someone raises a concern about pace or pressure, what actually happens?
• Are people who take their full leave treated the same as people who do not?
• Are layoffs, restructures, and “doing more with less” treated as wellbeing events, because that is what they are?

If most of your answers favour appearance over substance, no number of meditation app licences will move the needle. Fleming’s data already tells you that.

What actually moves the needle

The same body of research that shows wellness apps do not fix burnout also shows what does. None of it is mysterious. It is just harder.

Workload that matches the size of the team. This is the single most consistent finding. People can absorb intensity for a stretch. They cannot absorb permanent overload, and most burnout cultures are just permanent overload with better lighting.

Real autonomy, not posted autonomy. Letting people decide how to do their work, when to do it, and which problems to prioritise. The Job Demands-Resources research is clear that autonomy buffers people against high demands. A high-demand, high-autonomy job is tolerable. A high-demand, low-autonomy job destroys people.

Managers trained to be managers. Most managers were promoted because they were good individual contributors. Then they were handed a team and given a Slack account. Gallup’s 2025 data shows only 44 per cent of managers globally have ever received management training, and basic training alone cuts active disengagement in half. Manager quality is one of the strongest predictors of team mental health, and almost no organisation invests in it seriously.

Schedule and life flexibility. Not as a perk, but as a structural acknowledgement that humans have lives. Fleming specifically called this out as one of the most cost-effective changes available.

Job design. Many jobs are simply badly designed. Too many meetings. Unclear ownership. Overlapping demands. Constant context switching. You cannot meditate your way out of a calendar that has no time to think in it.

These changes are not soft. They are harder than buying an app. That is the point.

What you can actually do, depending on where you sit

If you are an employee who feels this in your bones, the most useful thing is to stop framing your exhaustion as a personal failure of resilience. It almost never is. When you hear “we are rolling out a new wellness platform” and you feel a small wave of irritation, that reaction is information. It is telling you the organisation is responding to symptoms it created. You do not have to fix the company to protect yourself. You can name what is wearing you down (workload, unclear priorities, a manager who does not shield the team, no recovery time) and decide whether to push for changes, lower your output to a sustainable level, or leave. Those are the real options. The fourth option, trying harder to be okay inside a system that is not, is the trap.

If you are a manager, you have more leverage than you think and less than you would like. You cannot redesign company-wide policy, but you can do three things that matter. Protect your team’s workload by saying no to work that does not fit. Give people genuine autonomy over how they work, not just where. Run one-to-ones that go beyond status updates, where people can actually say what is hard. Most managers skip this because it is uncomfortable. The ones who do not skip it have noticeably healthier teams.

If you are an HR or people leader, the hardest part of your job is probably explaining to executives that the dashboard-friendly intervention is not the effective one. Fleming’s study is your friend here. So is your own engagement data, if anyone has bothered to look at it honestly. Push for budget on manager training, workload audits, and job redesign before another app subscription. The first one is harder to sell. It is also the one that works.

If you are a senior leader, you set the conditions everything else lives inside. You decide whether the company runs lean to the point of breakage, whether managers are promoted on people skills or only on numbers, whether layoffs are treated as a financial event or also a mental health event. None of this is solved by procurement. The most useful question to ask in the next budget cycle is: of the money we are about to spend on employee wellbeing, how much is going to change the work, and how much is going to help people cope with work we have decided not to change?

That ratio is the real strategy.

The honest version of the conversation

I am not saying remove the meditation apps. If people use them and find them helpful, fine. The argument is not against individual tools. The argument is against pretending those tools are the answer to a structural problem.

The honest version of the conversation inside an organisation sounds something like this. We have a workforce that is anxious, exhausted, and quietly looking for the door. The reasons are not mysterious. The work is too much, the support is too thin, and the management is too inconsistent. We can keep buying things that signal we care, or we can change the things that wear people down. The first is cheaper this quarter. The second is cheaper across every quarter after that.

The wellness app was never the villain. It was just a convenient way of avoiding the harder work. Once you see that clearly, the next move is obvious. You do not need another subscription. You need to fix the environment that made the subscription feel necessary in the first place.

If this resonated, the most useful thing you can do is share it with the person in your organisation who actually has the authority to redesign work. They probably already know.