The chain had grown to 48 locations on the strength of its clinics, but its social advertising had not grown with it. Every market, from the newest opening to the busiest flagship, was served by a single national Meta account with one shared budget and identical creative. Media naturally pooled wherever the algorithm found cheap clicks, not where clinics had appointment capacity to fill, and local teams had no lever to pull when their diaries ran empty.

Expansion made the cracks visible. Each new metro entered the same undifferentiated campaigns, competing with established markets for budget while showing adverts written for nobody in particular. Over eighteen months the cost per lead doubled, and every launch became a slow, expensive grind as the account tried to learn a new city with no dedicated structure behind it.

Measurement compounded the problem. Browser-side tracking losses meant a large share of conversions never reached the ad platform, and the signal that did arrive stopped at the lead. Whether that lead actually attended a consultation lived only in the CRM, so the platforms optimised toward form fills, including the ones who never showed up.

We rebuilt the programme as a metro-level engine rather than patching the national account. Each market received its own campaign structure with budgets set by local clinic capacity, so spend followed appointment availability instead of the algorithm's path of least resistance. Lookalike and retargeting audiences were rebuilt per metro, seeded from local data rather than a national blend.

The measurement fix came next. Server-side Conversions API tracking replaced fragile browser-side signals, and CRM consultation outcomes flowed back to the ad platforms, so optimisation targeted people who attend consultations, not just people who fill in forms. That single change redefined what the campaigns were learning from: every attended consultation became a training signal, and wasted spend on audiences prone to no-shows began to fall away.

Creative moved to a monthly testing cadence. A central library supplied concepts, each metro localised them, and winners rolled out across markets while fatigued variants retired. A launch playbook then standardised how the engine enters a new city: structure, budgets, audiences and creative are set up the same way every time, so new metros start from a proven template rather than from scratch.