Rehab and recovery centre Facebook ads: measure admissions without a programme name leaving your site

Rehab and recovery centres have one of the hardest measurement problems in advertising, and one of the most sensitive. Getting it wrong is costly in two directions: you can burn budget on enquiries that never convert, or you can leak information that should never leave.

The thing you want to measure is an admission. That's the moment that matters. And it almost never happens on the website. Someone enquires, speaks to an admissions team, maybe visits, maybe arranges funding, and then, days or weeks later, they're admitted. The ad platform sees the enquiry and nothing else.

Meanwhile, everything about the way people find you reveals something highly personal. A page about a specific programme, a condition or a substance says why they're there. That's not data you want travelling with an ad event.

So the setup has two rules, and they're worth stating plainly.

Rule one: keep programme and condition names out of the data. Cut page links to your domain, use neutral event names, and send only a hashed identity and the ad click. The form and the call notes stay in your own system.

Rule two: count the admission, or the stage close to it. When your team marks someone as admitted, send that stage back, with the click attached and the identity hashed. That's what tells the ad platform what a real admission looks like.

There's a useful way to think about the funnel. Enquiry, then a first call, then an assessment, then admission. Each step filters out more people. The ad platform, seeing only enquiries, optimises for the widest, cheapest group. If it sees "assessment completed" and "admitted", it learns to find people who go all the way, which is a much smaller and much more valuable group.

Something to be careful about, honestly: the timing. Meta rejects events sent more than 7 days after they happened, so the stage needs to be sent promptly. If your admissions process is slow to update the CRM, your ads will always be learning from old data. A simple rule helps: update the status the same day it changes. It costs nothing, and it makes every number downstream better.

A made-up example. 200 enquiries. 90 complete a first call. 40 complete an assessment. 25 are admitted. If the ads see enquiries, they chase volume. If they see admissions, they chase the 25.

The handoff from an admissions call to the CRM is where a lot of these setups quietly fail, so let me spell it out.

Define the statuses first, in writing: enquiry received, first call completed, assessment completed, admitted. Each should mean exactly one thing.

Decide who updates each status, and when. Same day, by the person who actually did the step.

Keep clinical notes out of any field that syncs to other tools. It's easy for a free-text field to end up in an automation, so keep the notes in a separate, non-syncing place, and let only the status and the identifiers travel.

Then test end to end with a dummy record: move it through each status and check that the right event reaches the ad platform, with a hashed identity and the click, and with nothing in it that describes the person's situation.

It's a small amount of process, and it's what keeps the whole setup safe when the team changes.

How do you measure admissions from ads today?