Stable Clicks, Falling Bookings at a Six-Location HVAC Company

The evidence points away from creative and toward audience saturation plus a qualification leak. Stable 1.4% CTR means the ads still attract clicks; what changed is what happens after the click and who is clicking.

Three signals converge:

• Frequency of 7.2 in 14 days on a local service area is high. A six-location footprint has a finite pool of homeowners with a broken furnace, so the same people see the same ads repeatedly and click out of curiosity rather than need.
• Two overlapping campaigns (broad and interest-stacked) almost certainly bid against each other for the same people, driving up cost and frequency. Interest targeting does not expand reach here; it narrows an already small geo pool.
• Out-of-area leads suggest a targeting or form problem, such as radius settings that cross market boundaries, or a form that does not verify ZIP code before submission.

Conversion rate falling while CTR holds also points at landing page or offer mismatch: the page may promise emergency service while the traffic is largely research-stage. Before rebuilding audiences, check page load speed, form length, and whether the page matches each campaign's promise.

Decline the 15-audience plan with reasoning, not just a no. Adding interest layers to a saturated local audience shrinks delivery pools and increases overlap, and it does not solve the qualification leak.

Recommended structure:

1. Consolidate to one broad prospecting campaign per service line, geo-restricted by service radius rather than by interest. Broad targeting plus creative variety outperforms stacked interests at this budget and geography.
2. Set frequency controls appropriate to a considered local purchase, and monitor 7-day frequency as a leading indicator.
3. Build proper exclusion layers: existing customers, open opportunities, and leads from the last 30 to 60 days.
4. Add a small retargeting campaign for site visitors and form abandoners, which is where the conversion drop may be recoverable.
5. Fix lead qualification at the source: ZIP validation in the form, service-area copy on the landing page, and a clear next-step promise. If Meta lead forms are used, add a qualifying question on service type.

Test one variable at a time. Changing audience structure and adding 15 interests simultaneously makes it impossible to know what worked, and the manager will be back in six weeks with the same symptom.

CTR is a diagnostic, not a decision metric for a business that sells appointments. Replace it with a ladder that maps to revenue:

• Cost per qualified lead: leads inside the service radius with a real service need, which requires passing ZIP and service type back from the CRM or call center.
• Cost per booked appointment and appointment-to-job conversion rate, since booked appointments that never convert are expensive noise.
• Lead quality by campaign and creative, so the manager can see whether the sales complaints concentrate in one audience or one offer.
• Frequency, reach, and audience overlap as guardrails that explain efficiency changes before they become emergencies.

Set a review cadence: weekly on spend, frequency, and lead volume; monthly on cost per booked job and revenue by location. Segment by location, because a six-market footprint can hide that two markets are subsidizing four.

Finally, use a geo holdout in one or two markets to test whether Meta spend is incremental to Local Services Ads, which often captures the same high-intent searches. That answers the budget question that vanity metrics cannot.