Attribution infrastructure for a national care provider

A national care provider running always-on recruitment campaigns across multiple UK regions, with no reliable way to tell which spend produced which hire. Applicant tracking systems record the application. They do not record what caused it. So the platforms optimised towards whatever they could see, which was rarely the thing that mattered.

Role types compounded the problem. A registered nurse and a care assistant are entirely different economics, entirely different candidate pools, and entirely indistinguishable to an ad platform reading a generic form submission.

The measurement layer was rebuilt before a single campaign was touched. Server-side tagging via Stape, Meta CAPI, GA4 cross-domain, and integrations into the ATS platforms in use, Jobtrain, Workday and CareTech, so an application could be traced back to the campaign, the region and the role type that produced it.

Halo attribution was then built on top: a model quantifying candidates who see a paid ad and later apply through an organic route, which conventional last-click reporting discards entirely. Campaigns were weighted by live vacancy data so budget followed genuine hiring need rather than last month's plan.

Reporting was consolidated into blended Looker Studio dashboards combining ATS and ad platform data, structured for client-facing use.

Cost per lead fell 41%. Over 50,000 applications delivered across care and nursing campaigns.

More importantly, the attribution work showed the campaigns were producing 3.7× more hires than the ATS alone had been crediting, and on one client the multiplier between directly attributed and total influenced applications reached 13×. Halo-adjusted cost per hire landed at £99 against a comparator platform charging £173.

That gap was not a reporting curiosity. It was the difference between a channel that looked marginal and one that was demonstrably the cheapest source of hires the business had.