How an ophthalmology chain across 3 cities structured Growth-tier ChatGPT Ads across LASIK + cataract intent mix
The situation
Picture an ophthalmology chain running dedicated eye-care centres in three cities, with two procedures making up the large majority of its surgical revenue: LASIK refractive surgery and cataract surgery. On paper both are "eye surgery" and both had historically been marketed through one undifferentiated ophthalmology campaign across Google Ads and Meta Ads. In practice the chain's own consultation data showed these were two almost entirely separate businesses sharing an operating theatre and a brand name.
LASIK enquirers skewed younger, typically 22-40, self-paying, and actively comparing three or more providers on price, the specific laser technology used, and recovery time before committing — a purchase decision that often stretched over weeks or months as the enquirer weighed whether now was the right time financially and whether to wait for a newer technology generation. Cataract enquirers skewed considerably older, or were adult children enquiring on behalf of an ageing parent, and were typically working against a real functional deadline — vision had degraded to the point where driving, reading or daily tasks were becoming difficult — which meant a shorter, more urgency-driven decision window and far less price comparison shopping, with far more weight placed on surgeon experience, post-operative track record and how quickly a surgery date could be scheduled.
The chain's blended campaign was, by its own admission, better tuned to the LASIK buyer's price-comparison instincts than to the cataract buyer's urgency-and-trust instincts, which the chain suspected was leaving cataract volume — its higher-margin, more consistently scheduled procedure — under-served relative to its actual demand across all three cities.
The ChatGPT Ads campaign structure ICG designed
Growth-tier gave this account one ophthalmology service line running across all three cities, but split internally into two procedure-tagged conversation buckets — LASIK and cataract — each with its own bid ladder, its own landing page per city, and its own conversion-tracking tag, while sharing one compliance review process and one monthly reporting cadence. This is the internal-split pattern that distinguishes a well-run Growth-tier account from a naive single-bucket build: one service line, structurally, but two buyers underneath it that get treated separately where it matters.
The LASIK bucket's bid ladder weighted toward conversations showing later-stage comparison signals — a user naming a specific laser technology, asking about cost ranges, or asking how LASIK compares to newer refractive procedures — since these signals correlated with buyers closer to a decision than users asking only "what is LASIK." Bids on early-exploration LASIK conversations ran low, reflecting the long, comparison-heavy decision cycle this segment typically works through before converting, and reflecting that the chain's existing content already answered basic LASIK-education questions reasonably well organically.
The cataract bucket's bid ladder ran differently by design, weighted toward urgency and trust signals rather than price-comparison signals — conversations mentioning declining vision, difficulty with daily tasks, or asking specifically about surgeon experience or how soon surgery could be scheduled were bid at the account's highest tier, illustratively 5-7x baseline, since this segment's shorter decision window meant a well-matched conversation was disproportionately likely to convert quickly if the chain could be the first credible, reassuring option the enquirer found.
Landing infrastructure ran two page templates per city — six pages total across the three-city footprint — with LASIK pages leading on technology detail, transparent price ranges and recovery-time specifics the comparison-shopping buyer needed, and cataract pages leading on surgeon credentials, typical time-to-surgery-date and what the post-operative recovery period involves for an older patient, in each case written as short, factual, cited pages a conversational assistant could summarise accurately rather than long-form marketing copy. Conversion tracking tagged every enquiry by procedure at the point of form submission, which let city-by-city, procedure-by-procedure reporting run cleanly without the chain having to reconcile a blended lead list against its own patient records afterward.
The compliance discipline
NMC Section 6 governed both buckets as the baseline, but the two procedures carried different specific risks. LASIK copy across all three cities avoided any outcome guarantee — no "perfect vision guaranteed" framing, no implied promise that a specific laser technology produced uniformly superior results, since NMC guidance treats technology-superiority claims in advertising with the same scrutiny as outcome guarantees when the claim isn't substantiated by cited clinical evidence available to the reviewer.
Cataract copy carried a different risk given the segment's urgency-driven buyer: early creative drafts leaned toward language implying that delaying surgery risked permanent vision loss, which is medically inaccurate framing for the vast majority of cataract cases and reads as exactly the kind of fear-based urgency pressure ASCI Chapter III and NMC Section 6 both flag. Every variant carrying that framing was rewritten to state factual, non-alarmist information about when cataract surgery is typically recommended, without implying a false medical emergency to accelerate a booking decision.
Before-after visual claims, a common LASIK marketing convention on other channels, were excluded from this account's ChatGPT Ads creative entirely, since NMC Section 6 restricts this format closely enough that ICG's practice across ophthalmology accounts generally is to avoid it in paid advertising rather than build case-by-case justification for each instance. DPDP 2023 governed both buckets' enquiry forms, with the cataract form's age-related and vision-symptom fields treated as sensitive health data carrying explicit consent language, given the segment frequently involved a family member submitting the enquiry on a parent's behalf. Across the full six-page, two-bucket, three-city account, every ad variant cleared compliance review before launch with zero post-launch takedowns.
The 90-day outcome pattern
In the pattern ICG has observed on comparable Growth-tier ophthalmology accounts, the two buckets diverged sharply in both volume and conversion speed, exactly as the underlying buyer behaviour predicted. Illustrative figures: the LASIK bucket generated a higher raw volume of conversation-completions across the three cities, climbing from roughly 30-35 a week in month one to 60-70 a week by month three, but with a conversion-to-consultation rate that stayed comparatively modest throughout, illustratively 18-22% within a 14-day window, consistent with LASIK's longer, multi-provider comparison cycle.
The cataract bucket generated lower raw volume — illustratively 15-20 conversation-completions a week by month three — but converted to a booked consultation far faster and at a far higher rate, illustratively 50-55% within a 7-day window, and a large share of those consultations scheduled a surgery date within the same visit, consistent with the segment's urgency-driven decision pattern. This gap between LASIK's high-volume-low-conversion pattern and cataract's low-volume-high-conversion pattern is precisely why the two buckets needed separate bid ladders rather than one blended ladder optimising toward a single average.
Blended across both buckets and all three cities, cost per qualified lead from ChatGPT Ads ran illustratively 20-25% below the chain's prior blended Google Ads cost per lead over the same quarter, with the improvement concentrated almost entirely in the cataract bucket, where the previous undifferentiated campaign had been bidding cataract-intent conversations at the same low rate as early-exploration LASIK conversations and therefore losing auction visibility against more aggressive competitors. In GA4, both buckets attributed to the AI Assistant channel, with the cataract bucket's key-event rate running notably above the account's LASIK bucket and above ICG's broader healthcare-book average, consistent with a well-matched urgency segment converting cleanly through a conversational channel. Downstream, the chain's own scheduling data showed cataract surgery bookings sourced from this campaign scheduling, on average, faster than cataract bookings sourced from other channels, reinforcing that the urgency-weighted bid ladder was reaching genuinely ready-to-book patients.
What we'd do differently next time
Starting with a single combined ophthalmology landing page per city before splitting into procedure-specific pages in week three cost the account roughly two to three weeks of under-optimised conversion, since the combined page's compromise content served neither buyer as well as a dedicated page would. Building the split landing pages from day one, even at the cost of more upfront production work, would have avoided that early dip in both buckets.
The cataract bucket's rapid conversion-to-scheduling pattern meant the account could plausibly have absorbed a higher bid ceiling than initially set, and by the time month-two data made this clear, the bucket had already been running below its likely-optimal spend level for several weeks — a more aggressive initial bid ceiling on the cataract bucket, with room to pull back if volume outpaced capacity, would likely have captured more of the available demand earlier.
Finally, the account's monthly reporting initially blended both buckets into one topline number for the chain's leadership review, which obscured exactly the divergence described above until ICG restructured the report to split every metric by procedure starting in month two — a lesson worth applying from day one on any account carrying two structurally different buyer segments under one brand.
How this maps to your own vertical
If your ophthalmology practice, or any single-specialty practice, runs two or more procedures with meaningfully different buyer profiles — different age groups, different urgency levels, different price sensitivity — under one service line, the lesson here is to split your conversation buckets and bid ladders along that buyer-behaviour line rather than along a purely clinical or marketing-convenience line, even while keeping one shared compliance process and one shared reporting account at Growth-tier.
The NMC Section 6 discipline around outcome guarantees, technology-superiority claims and before-after visual content applies to any surgical-ophthalmology advertising account regardless of tier or city count — building that review process in before launch, procedure by procedure, avoids the mid-campaign creative rework a compliance flag forces.
