Nine Clinics, One Campaign, One Misleading CPA
A dental group runs nine clinics across three metros on $30,000/month. Everything sits in one Search campaign targeting all nine clinic cities with the location setting "presence or interest", one shared budget, and location bid adjustments of +20% on two locations. All clinics share a single conversion action: appointment form submissions plus calls over 60 seconds. Blended CPA is $95 against a $70 target. Clinic-level reporting shows two suburban clinics at $180 CPA with 8% impression share lost to rank, while the downtown flagship is limited by budget and converting at $48. A tenth clinic opens next month in a city where the group has no history or reviews.
1. Diagnose why blended CPA is hiding the problem, and propose a structure and budget logic that would surface it.
Read the worked answer
A single campaign with a shared budget forces every clinic to compete for the same pool, and the bid strategy optimises to whichever clinic converts most cheaply. The flagship at $48 wins, the suburbs at $180 starve. Blended $95 looks acceptable while two of nine locations are uneconomic.
- Segment by clinic or by metro tier so budget and targets can differ. Each location has a different capacity, margin and competitive density.
- Give the underperformers their own budget and target, so their CPA is visible and managed instead of averaged away.
- Check capacity first: a clinic with two hygienists cannot absorb flagship-level lead volume. Bid strategy pushes spend where conversions are cheapest, not where they are most valuable.
- Report per-location CPA, appointment show rate and revenue per location, not one blended figure.
8% impression share lost to rank at $180 CPA also suggests the issue is landing page and relevance, not budget.
2. What location targeting setting should the tenth clinic use, and what else must be true before you fund it?
Read the worked answer
"Presence or interest" will serve ads to people researching the city from elsewhere and to people merely showing interest in the area - wasteful for a clinic that needs patients who can physically attend. For a physical clinic, presence is the safer default, with a tight radius around the address and possibly a negative list for the neighbouring metro.
Before funding, three things must be settled:
- Conversion and tracking: the new location needs its own conversion actions or at least location-tagged values, or it will be invisible inside the blend again.
- Realistic target: with no history, no reviews and no brand recognition, a $70 CPA copied from an established clinic is not a forecast. Start with a learning budget that reflects ramping demand and revisit after enough conversions to be meaningful.
- Capacity floor: understand how many new patients per week the clinic can actually see; below that, an aggressive CPA target buys appointments it cannot serve.
3. How would you decide whether the +20% location bid adjustments are helping or just raising costs?
Read the worked answer
A location bid adjustment boosts bids in a geography, but it does nothing about ad relevance, landing page or the local competitive set - so it is the wrong lever if the problem is quality or conversion rate.
Test it explicitly:
- Compare each adjusted location's CPA, conversion rate and impression share lost to rank against unadjusted locations with similar demographics. If the adjustment is not raising impression share in a competitive auction, it is only raising cost per click.
- Note that bid adjustments and Smart Bidding interact: with a portfolio or target-based strategy, manual adjustments can be overridden or absorbed. Confirm the strategy is honouring them before crediting them with anything.
- Run a clean before/after across a stable period, or split locations into adjusted and unadjusted groups. If the adjusted group shows no lift in served impressions or appointments, remove the adjustment.
The decision rule: keep an adjustment only if it buys incremental reach you can serve, not simply a higher average CPC.