How a multispecialty hospital in tier-2 North India migrated 40% of its paid budget from Google Ads to ChatGPT Ads in Q3-2026
The situation
Picture a 200-bed multispecialty hospital operating in a tier-2 city in North India — the kind of regional anchor institution that draws patients not just from its own city but from a cluster of smaller towns within a two-to-three-hour radius, competing against two or three comparable hospitals in the same catchment for the same specialty consultations. By early 2026, this hospital's marketing team had run Google Ads Search campaigns across its five highest-volume specialties for over three years, and cost per click in that market had climbed steadily as competing hospitals bid more aggressively for the same keyword inventory, a familiar pattern in any tier-2 healthcare market once two or more well-funded institutions start competing on the same channel.
The marketing head's problem wasn't lead volume in the abstract — Google Ads was still producing enquiries — it was that cost per qualified consultation booking had risen roughly 35-45% over eighteen months while consultation-to-admission conversion held flat, meaning the hospital was paying meaningfully more for the same downstream outcome. Board-level budget conversations in early 2026 were pushing for either a flat marketing spend with declining lead volume, or finding a channel that could restore the cost-per-lead economics the hospital had enjoyed two years earlier.
ChatGPT Ads India's mid-2026 launch gave the hospital's marketing team an opening to test a channel with materially less competitive density in a tier-2 specialty market than Google Ads had by then accumulated. The hospital's own patient base skewed toward research-heavy decision-making — families comparing hospitals for cardiology or orthopedic procedures typically consulted several sources, including, increasingly, conversational AI assistants, before calling any hospital directly. That behavioural shift, combined with rising Google Ads costs, was the direct trigger for a structured 60-day parallel test that preceded the Q3-2026 budget migration this case describes.
The ChatGPT Ads campaign structure ICG designed
Growth-tier fit this hospital's needs during the parallel-test phase and carried through into the Q3 migration: five specialty conversation buckets — cardiology, orthopedics, gastroenterology, oncology and general surgery — run under one shared hospital-brand compliance track, with specialty-specific intent ladders rather than a single blended bucket. Each specialty's conversations behave differently; a cardiology conversation about chest pain symptoms carries a different urgency signal than an orthopedic conversation about a scheduled knee-replacement second opinion, and the bid ladder needed to reflect that rather than treating all five specialties as one undifferentiated demand pool.
Within each specialty bucket, the intent ladder separated symptom-research conversations from consultation-ready conversations. A user asking an assistant to explain a cardiac symptom was treated as early-stage and bid low; a user asking the assistant to help them find a cardiologist in their specific city or asking comparative questions about a hospital's cath-lab capability was treated as consultation-ready and bid 5-8x higher. This mirrors the logic Google Ads Search had struggled to replicate cost-effectively in this market, where broad symptom keywords had become expensive without reliably surfacing consultation-ready searchers.
Landing infrastructure routed each specialty's conversations to a dedicated, machine-scannable specialty page naming the relevant department head or senior consultant, the hospital's specific equipment or accreditation relevant to that specialty, and a direct consultation-booking form — deliberately avoiding a single generic "book a hospital appointment" page, since a user arriving from a cardiology-specific conversation expects confirmation that the hospital has a real cardiology programme, not a generic front door. Budget allocation across the five specialties wasn't equal; cardiology and orthopedics, the hospital's two highest-revenue specialties and the ones carrying the steepest Google Ads cost increases, received roughly 55% of the Growth-tier budget between them, with gastroenterology, oncology and general surgery splitting the remainder.
The migration mechanism itself ran as three tranches over six weeks once the parallel test confirmed the channel's economics: Google Ads Search budget for cardiology, orthopedics and gastroenterology was reduced by roughly 13-14 percentage points of the total specialty ad spend in each of three two-week windows, with each reduction timed against a week where ChatGPT Ads conversation-completion volume for that specialty had already demonstrated it could absorb the shifted demand without a corresponding dip in total consultation bookings. Oncology and general surgery stayed on their existing Google Ads allocation through the full quarter, both because their conversation volume on the parallel test had been thinner and because oncology copy required a longer compliance review cycle the marketing team wasn't ready to compress under a migration deadline.
The compliance discipline
A five-specialty hospital account carries five separate compliance surfaces rather than one, and the discipline that mattered most in this scenario was refusing to let a single shared review checklist stand in for specialty-specific scrutiny. NMC Section 6 and ASCI Chapter III applied uniformly as the baseline across all five specialties — no outcome guarantees, no comparative superiority claims against unnamed competing hospitals in the same city, no testimonial-as-advertisement framing — but the specific risk profile within that baseline differed meaningfully by specialty.
Cardiology copy was checked for language that could be read as implying guaranteed procedural success or minimised risk around cardiac intervention, a common drafting instinct when trying to reassure an anxious patient family that ASCI Chapter III explicitly prohibits. Orthopedic copy was checked for implied recovery-timeline promises, since "back on your feet in weeks" framing, while intuitively appealing marketing language, edges into an outcome guarantee the regulatory framework doesn't permit. Oncology copy, the specialty that stayed on Google Ads longest in this scenario, required the most extended review precisely because oncology advertising carries the highest sensitivity to any language that could be read as offering false hope or minimising the severity of a cancer diagnosis — the marketing team's decision to hold oncology back from the Q3 migration was as much about giving compliance review adequate time as it was about thinner parallel-test data.
DPDP 2023 governed every specialty's enquiry form consistently, with particular attention to the cardiology and oncology forms, where the symptom-description fields patients filled in constituted sensitive health data requiring explicit consent language distinct from a general contact-request form. Across the full Q3 migration, all cardiology, orthopedic and gastroenterology ad variants cleared review before the relevant tranche's launch date, and the hospital's own legal team, which had previously flagged concerns about how aggressively some Google Ads copy framed outcome language, signed off on the ChatGPT Ads compliance process as measurably more conservative than what the hospital's prior agency had delivered.
The 90-day outcome pattern
Across the 90 days following the start of the first migration tranche, the pattern ICG has observed on comparable tier-2 multispecialty hospital accounts held here as well: the three migrated specialties saw ChatGPT Ads conversation-completion volume climb from an illustrative combined 40-50 a week in the first two weeks post-tranche-one to roughly 110-130 a week by day 90, as both the intent ladder matured and the hospital's overall specialty-page presence in conversational search accumulated more citation history for the assistant to draw on.
The core economic result the migration was built to test held up: blended cost per qualified consultation booking across cardiology, orthopedics and gastroenterology ran illustratively 25-35% below what the same three specialties had been paying on Google Ads Search in the quarter immediately prior, restoring cost-per-lead economics closer to what the hospital had seen two years earlier before competitive bidding pressure escalated. Cardiology showed the largest gap, illustratively 32-38% lower cost per qualified booking, consistent with cardiology also having carried the steepest Google Ads cost increase pre-migration — the two specialties with the worst Google Ads economics turned out to have the most headroom on the new channel.
Total consultation booking volume across the three migrated specialties held roughly flat through the transition rather than dipping, which the marketing head considered the more important result than the cost improvement alone — a budget migration that improved cost per lead but cost the hospital net booking volume during the transition quarter would have been a harder result to defend to the board regardless of the unit economics. In GA4, the ChatGPT Ads traffic attributed to the AI Assistant channel showed a key-event rate consistent with ICG's broader healthcare book, meaningfully above the hospital's organic and direct traffic, which the marketing team used as supporting evidence when presenting the Q4 case for extending the migration to oncology once that specialty's compliance review concluded. Consultation-to-admission conversion for the ChatGPT Ads-sourced leads tracked in line with the hospital's existing Google Ads-sourced leads rather than showing a meaningfully different rate, suggesting the channel shift changed cost efficiency and channel mix without changing the underlying quality of patients who ultimately booked.
What we'd do differently next time
Running the initial parallel test at roughly 15% of specialty ad budget produced a usable but noisier signal than a larger sample would have; month one of the full Q3 migration still carried some residual calibration uncertainty in the bid ladder that a 20-25% initial test allocation, held for the same 60-day window, would likely have already resolved before the migration decision was made.
Holding oncology and general surgery entirely off the migration schedule for the full quarter, rather than running a smaller-scale parallel test on those two specialties in parallel with the main migration, meant the hospital entered Q4 planning without any data on how those specialties would perform on the channel — a missed opportunity to have oncology's compliance review running concurrently with, rather than sequentially after, the main three-specialty migration.
The three-tranche, two-week-interval migration schedule was conservative by design, and in hindsight the first tranche's two-week hold before moving to the second could likely have compressed to ten days once the cardiology data alone showed the channel could absorb shifted demand — a faster cadence would have reached full migration for the three specialties roughly two weeks earlier without materially higher risk.
How this maps to your own vertical
If your hospital or multispecialty group is watching Google Ads cost per click climb in a competitive local market, the structure here — a bounded parallel test at a meaningful budget share, specialty-specific intent ladders rather than one blended hospital-wide bucket, and a tranche-based migration timed against confirmed conversation-completion volume rather than a single budget cutover — is a defensible way to shift spend without risking a volume dip your board will notice before your cost savings show up in the numbers.
The specialty-by-specialty compliance sequencing matters regardless of city tier or hospital size: a migration schedule should follow compliance readiness, not the other way around, and holding a higher-sensitivity specialty like oncology back from an aggressive timeline is the right call even when the budget case for moving it looks strong on paper.
