About

I've seen workplace wellbeing sold on how it looks, with very little said about whether it works. This is where I write about that.

The day job is Alltoogether, an employee benefits broker with its workplace-health platform included. Before that I worked at a large corporate benefits consultancy, where I kept being told to talk to clients about wellbeing. Broking benefits policies I understood. Wellbeing I didn't, and I was being paid to sound as if I did. So I did an MSc in Workplace Health and Wellbeing at Nottingham, part time, while working. My thesis looked at how subjective wellbeing relates to the psychosocial work environment, and it left me obsessed with measuring properly: single-item measures, quantitative wellbeing, and the difference between a number and a finding. I also hold a law degree (LLB Hons).

Over fifteen years in this field, I've kept running into records that don't fit together. I wrote the Open Workplace Health Standard to give them a shared structure. It is an open draft, free for anyone to use.

Here you'll find longer notes on work, health, working with AI and where work is heading. Where I make a claim, the evidence is laid out with its limits. Some notes are rough, and corrections are dated.

The writing is mine. The checking is done with AI agents, working from the sources, and every claim they check shows the passage it rests on so you can check it too. Where they could not reach a source, the entry says so.

Nothing here speaks for Alltoogether.

How this site works

Notes are essays in progress. They are dated, they can be rough, and each one stands on its own.

Evidence reviews sit behind some notes, and a few stand on their own. They set out what the research actually supports, claim by claim, with the source passage, its limits, how much of the source was read and the date it was checked.

Corrections that change what a piece says are dated on the piece and listed on the corrections page.

What the verdicts mean

Supported

The source says what the claim says.

Overstated

The source supports a weaker version of the claim.

Misattributed

The finding is real but comes from a different source.

Unsupported

The source does not support the claim.

Unverifiable

The source could not be accessed or checked. The attempted source, the reason and the date are recorded.

In review

Not yet checked against the source.