Search and Social Solve Different Clinic Demand Problems

Channel-native measures remain useful. Search terms, impression share and landing behaviour can diagnose response to query-led demand. Audience, format, video and placement reports can diagnose how visual discovery is delivered. Demand Gen, for example, allows network and ad-format segmentation. Its reporting guidance is a reason to inspect the distribution, not a reason to convert every view or click into the same business meaning.

The allocator needs a common later state. Google distinguishes an initial lead from advertiser-defined qualified and converted leads. That funnel taxonomy can support deeper comparison if the clinic's definitions are stable. It still does not independently verify an appointment, attendance or contribution. Those states belong to the clinic's operational system.

• Use platform metrics to diagnose delivery and response within a channel.
• Use a stable clinic outcome to compare mature cohorts across channels.
• Segment by service, branch, language and relevant catchment before reading an average.
• Keep unattributed and mixed-source outcomes visible rather than forcing certainty.
• Add capacity and patient-fit guardrails so volume cannot win by overwhelming operations.

1. Name the missing state: insufficient response to existing intent, insufficient informed discovery, weak contactability or poor downstream yield.
2. Choose the smallest channel change that can test that state without changing service, geography, offer and operations at the same time.
3. Set the primary mature outcome, leading diagnostic, safety or capacity guardrail and minimum useful window before launch.
4. Let the cohort mature. Recent reports can look inefficient while later conversions are still arriving.
5. Keep, reverse or refine the shift based on the stated hypothesis—not on whichever dashboard moved most dramatically.

Google warns that conversion delay can make recent periods show spend before later conversions arrive. The timing guidance is especially relevant when a clinic compares an immediate form event with a later qualification or booking state. A seven-day report and a 30-day operational outcome are not competing truths; they are different clocks.

Suppose a dermatology service has strong eligible search coverage and stable cost per qualified enquiry, but new-patient volume has stopped growing. Social video generates inexpensive landing visits, yet only a small share reach a qualified contact. The answer is not to crown search or cancel social. Search may be doing its response job near the current intent ceiling; social may be producing attention without enough informed progression.

This frame protects the budget from false symmetry. Search does not need to imitate social's cheapest attention metric, and social does not need to receive last-click credit for every later branded query. Each channel keeps the measures needed to understand its job, while the clinic uses one mature outcome and one capacity boundary to decide marginal investment.

• [ ] The demand state is named in plain language.
• [ ] The proposed channel can plausibly address that state.
• [ ] Eligibility and approval are already settled outside this decision.
• [ ] The clinic can observe a common mature outcome without unsafe data use.
• [ ] The review window covers the relevant conversion delay.
• [ ] Capacity, contact quality and patient-fit guardrails have owners.

The useful output is not a permanent channel split. It is a documented reason for the next marginal test. When the journey changes, the allocation can change. That makes the plan responsive without turning every weekly fluctuation into a reallocation ritual.