Twelve campuses, bidding against each other.

When we started, the group’s twelve regional campuses each operated an independent paid search account, set up at different times by different local teams, all bidding on substantially the same set of course-related terms. Because catchment areas overlapped, several campuses were competing for the same auction on the same keyword, from the same parent organisation, which meant the group was inflating its own cost per click without any corresponding gain in total enrolments. No one at head office had visibility into this — each campus reported its own cost per enrolment, and none of the twelve numbers were built the same way.

We consolidated the twelve accounts into a single structure with shared negative keyword lists to stop internal competition, campus-level campaigns underneath for attribution, and a rebuilt enquiry form that captured qualification data — programme interest, funding eligibility, prior qualification level — at the point of submission. That qualification data was passed back into the bidding layer, letting the algorithm optimise toward enquiries that actually converted to enrolment rather than toward enquiry volume alone.

The 38% reduction in cost per enrolled student was measured against the group’s student information system, not against platform-reported cost per lead — the qualification data closing the loop between what an ad platform called a good result and what the registrar’s office called a good result.