How an IVF chain across 5 metro cities used Growth-tier ChatGPT Ads to lift second-opinion conversions
The situation
Picture an IVF chain operating fertility centres in five metro cities across India — the kind of mid-sized network that isn't a single-clinic operation but isn't a national hospital brand either, competing for patients who are actively comparing 2-3 centres before committing to a treatment cycle. The chain's core acquisition challenge wasn't top-of-funnel awareness; IVF as a category is well understood by the audience searching for it. The challenge was capturing patients at the specific, high-value moment when they'd already had one IVF consultation elsewhere, weren't fully convinced by what they'd been told, and were actively looking for a second opinion.
This second-opinion segment mattered disproportionately to the business. Illustrative internal data the chain shared during scoping suggested second-opinion patients who converted to a consultation had a meaningfully higher rate of proceeding to a full treatment cycle than first-time enquirers, because they arrived pre-qualified — already diagnosed, already emotionally committed to pursuing IVF, just uncertain about where. The chain's existing Google Ads and Meta Ads spend wasn't distinguishing between these two audiences well; both channels were optimising for volume of enquiry, which meant budget was being spent chasing early-stage researchers at the same rate as ready-to-move second-opinion seekers.
The five-city footprint added its own complexity. Delhi NCR and Mumbai carried heavy existing competition from other fertility networks bidding aggressively on the same audiences across every paid channel. Bengaluru and Hyderabad had moderate competition. Pune was comparatively under-served. A single blended national campaign risked either overspending in the competitive metros or underspending in Pune relative to its actual opportunity — the same structural tension that shows up whenever one budget has to serve genuinely different local markets.
The ChatGPT Ads campaign structure ICG designed
Growth-tier gave this account a single conversation bucket built specifically around second-opinion intent signals, run across all five cities with city-level bid adjustments rather than five fully independent campaigns. This is the structural difference from the Scale-tier multispecialty build: one coherent service line, one compliance track, one attribution model — but geographic nuance layered on top rather than five parallel machines.
The conversation-stage bid ladder was tuned deliberately toward the second-opinion signal rather than general IVF exploration. Conversations that showed markers of prior consultation — a user referencing an existing diagnosis, asking comparative questions about protocols or success context, or explicitly asking for a second opinion — were treated as high-intent and bid accordingly, in the range 6-9x above baseline exploration-stage bids. General "what is IVF" or "how does IVF work" conversations, which the chain's organic content already served reasonably well, were bid low or, in some cases, deliberately excluded from the paid budget entirely so spend concentrated on the segment the business case was built around.
City-level bid adjustments ran on top of the intent ladder. Delhi NCR and Mumbai bids were set higher to compete for visibility in more contested conversation auctions, Pune's bids ran lower given the lighter competition, and Bengaluru and Hyderabad sat in between. This is a lighter-weight version of the geographic tuning that would, in a larger account, justify separate city buckets — at Growth-tier budget levels, a single bucket with city-weighted bidding captured most of the benefit without the reporting overhead of five independent campaigns.
Landing infrastructure routed every conversation to one of five city-specific, machine-scannable pages — short, factual, cited, each naming the specific centre location and its second-opinion consultation process, because a user arriving from a conversational assistant wants confirmation that a real, nearby option exists, not a national brand page. Each page's enquiry form asked directly whether the patient had a prior IVF consultation elsewhere, which let the chain's front-desk team route second-opinion leads to a specific consultation format distinct from first-time enquiries. Monthly reporting ran on conversation-completion attribution reconciled against the chain's CRM city by city, so the chain could see not just aggregate performance but which of the five cities was actually returning the second-opinion volume the campaign was built to capture.
The compliance discipline
IVF advertising in India sits under one of the strictest compliance overlays in the healthcare category, and a second-opinion-focused campaign carries a particular risk that a general awareness campaign doesn't: the temptation to imply, even subtly, that the chain's protocols would succeed where a prior consultation elsewhere had failed or fallen short. The ART (Regulation) Act 2021 prohibits exactly this kind of comparative outcome implication, and every ad variant in this account was screened specifically for language that could be read as "we'll get you the result the other clinic couldn't."
The Act's success-rate restriction applied directly here too — no bare success-rate figure could appear in ad copy without the mandated context (cycle count, age cohort, or whatever qualifier the figure depends on), and ICG's practice on this account was simply to avoid citing success-rate numbers in the ad copy itself, reserving that context-dependent information for the landing page where it could carry its full qualifying detail. The Act's sex-selection-advertising prohibition, while less directly relevant to a second-opinion angle than to general IVF promotion, was still checked on every variant as a standing rule rather than a case-by-case judgment call — nothing in the campaign referenced gender outcomes in any form.
NMC Section 6 and ASCI Chapter III applied as the baseline across all five cities' copy — no outcome guarantees, no testimonial-as-advertising framing, no urgency language pressuring a patient toward a decision. DPDP 2023 governed the enquiry forms, particularly given the sensitivity of the "prior consultation elsewhere" question each form asked — that field carried explicit consent language and a stated data-handling policy, since it's effectively health data about a relationship with another provider, not just contact information. Across the full engagement, every ad variant across five cities cleared compliance review before launch with zero post-launch takedowns, which the chain's own legal counsel flagged as notably cleaner than what they'd experienced with some of their earlier Meta Ads creative.
The 90-day outcome pattern
In the pattern ICG has observed on comparable Growth-tier IVF accounts, month one ran as a calibration period — the second-opinion intent signals needed real conversation volume before the bid ladder could reliably distinguish a genuine second-opinion conversation from a general exploration one that happened to use similar language. Illustrative volume for this hypothetical: combined conversation-completion events across all five cities started around 25-30 a week in month one, climbing to roughly 70-85 a week by month three as the ladder's intent classification sharpened.
The second-opinion targeting showed its value most clearly in conversion-to-consultation rate rather than raw volume. Illustrative figures: roughly 45-50% of conversations classified as second-opinion intent converted to a booked consultation within 10 days, compared with an illustrative 12-15% for the general-exploration conversations the chain had been running on Google Ads previously. This gap is the core thesis the campaign was built to test, and it held up across all five cities, though the absolute volume varied — Delhi NCR and Mumbai produced the highest raw numbers of second-opinion conversations given their larger addressable population, while Pune's lower competition meant a smaller absolute volume but a notably efficient cost per qualified lead, illustratively 30-40% below the Delhi NCR figure for the same intent tier.
Blended across the five cities, cost per qualified second-opinion lead from ChatGPT Ads ran illustratively 20-30% below the chain's blended Google Ads cost per lead over the same quarter, with the gap most pronounced in Delhi NCR and Mumbai where Google Ads competition among fertility centres was steepest. In GA4, this traffic attributed to the AI Assistant channel, showing a key-event rate consistent with ICG's broader healthcare book — meaningfully above organic and direct — reinforcing that conversation-completion attribution was capturing a genuinely higher-intent audience rather than just a cheaper one. Downstream, the chain's own consultation-to-cycle-start tracking (kept internally, reconciled with ICG only at an aggregate level) reportedly showed the second-opinion segment converting to a paid treatment cycle at a distinctly higher rate than the chain's first-time-enquirer segment, consistent with the original business case for focusing the campaign here rather than on broad IVF awareness.
What we'd do differently next time
Launching all five cities on day one, with a single shared bid ladder differentiated only by a city-weight multiplier, meant the more competitive metros absorbed disproportionate learning-phase budget before the system had enough data to weight them correctly. Launching Delhi NCR and Mumbai first, letting the intent-classification model mature on the highest-volume markets, and bringing Bengaluru, Hyderabad and Pune in during weeks three and four would likely have reached steady-state performance faster across the full footprint.
The "prior consultation elsewhere" form field, while operationally useful for routing leads, generated more front-desk hesitation among patients than expected — several patients left it blank or gave vague answers, suggesting the framing needed softening or needed to be optional rather than a required field. A lighter-touch version of that question, asked conversationally by staff during the first call rather than required on the form itself, would likely have preserved the routing benefit without the drop-off risk.
Finally, excluding general "what is IVF" conversations from paid budget entirely was the right compliance-conscious and cost-conscious call, but it meant the campaign built no top-of-funnel brand presence in the ChatGPT Ads channel at all — a decision worth revisiting once the second-opinion segment matures, since a modest allocation toward early-exploration conversations could compound into future second-opinion volume as those users progress through their own journey over subsequent months.
How this maps to your own vertical
If your fertility centre or IVF chain runs a single core service line across one or several locations, this Growth-tier structure — one intent-focused conversation bucket, city-weighted bidding rather than fully separate campaigns, and a landing page per location — is likely the right starting shape, whether you're operating in one city or five. The specific lesson worth taking regardless of city count is the value of targeting a defined intent segment like second-opinion seekers rather than optimising broadly for IVF enquiry volume; a narrower, higher-intent audience converts at rates that make a smaller budget go further than chasing scale on unqualified early-stage traffic.
The ART Act compliance discipline described above isn't optional or scenario-specific — it applies to any IVF or fertility advertising account regardless of tier or city count, and it's worth building that review process before launch rather than after a first compliance flag forces a mid-campaign pause.
