Building the business case when the evidence is against you
Two things are true at once. Benefit changes get rejected for lack of data. And the best available studies say wellbeing programmes do not reliably produce the financial return vendors promise. A business case has to live with both.
- A third of proposed benefit changes are rejected for lack of data.
- Large studies found no significant effect on spend, absence or job performance.
- So do not build the case on health outcomes or ROI. It will not survive a CFO who has read them.
- Build it on cost per activated employee, coverage and take-up — all countable, none of it personal.
If your paper promises health outcomes or a financial return, someone in the room may have read these. Writing as if they do not exist is the fastest way to lose the argument.
| Claim | Evidence you can actually produce | Why it holds |
|---|---|---|
| People can use it | Coverage: how many have access | Countable, no personal data |
| People do use it | Activation: how many started | Countable, and the number that moves |
| It was worth starting | Return within a month | The only honest quality signal |
| It survived the launch | Spread across the year | Separates a benefit from a campaign |
| It costs this much per person who used it | Cost ÷ activated | The number finance actually wants |
Cost per activated employee is a better number than cost per head, and it is harder to game. A scheme nobody opens has an infinite cost per activation, which is exactly the right answer.
Engagement metrics measure whether someone opened a thing. An app with a daily streak will show rising engagement whether or not anyone is better off. None of it is an outcome, and a CFO who has been sold it once will say so.
- Contract on codes issued and codes activated
- Report counts, never individuals
- State what the scheme cannot prove
- Benchmark against comparable employers
- Promise reduced absence
- Quote a vendor ROI multiple
- Use satisfaction scores as evidence of effect
- Call engagement an outcome
Benchmarking is reportedly far more persuasive to approvers than ROI modelling — which is worth knowing before you spend a fortnight building a model.
A wellbeing benefit on devices people already own can be argued for in one paragraph: it costs this, this many people can use it, this many started, this many came back, and the organisation learns nothing about any of them. No health claim is made, so none has to be defended.
That is a smaller claim than most vendors make. It is also one you can evidence at renewal.
Is there any defensible outcome claim?
Take-up and repeat use are real and yours to measure. Health outcomes from a voluntary programme are hard to attribute, which is why the studies above found what they found.
What does renewal evidence look like without individual data?
Activation, repeat use and spread across the year, all aggregate. If a supplier can only prove value by naming people, that is a signal about the supplier.
Our board wants a cost-benefit number.
Give them cost per activated employee and coverage, and say plainly that a financial return on wellbeing spend is not well evidenced. Finance people generally respect that more than a multiple.