How an OTC cough-and-cold brand ran Growth-tier UCPMP-clean ChatGPT Ads across pre-monsoon and monsoon peaks
The situation
Picture a mid-sized Indian pharmaceutical company with an established over-the-counter cough-and-cold portfolio — a syrup, a lozenge line and a fast-dissolve tablet, sold through the standard retail-plus-e-pharmacy mix that most OTC brands in this category rely on. The brand had run television and print for years, alongside a modest always-on Google Ads presence built around branded search and a handful of symptom-adjacent generic terms the compliance team had cleared long ago. Meta Ads had been tried twice and pulled back both times — the brand's marketing head was cautious about running symptom-targeted advertising on a platform where ad review didn't always seem to catch the subtler UCPMP issues before a competitor or a regulator did.
The business problem was seasonal and specific. Roughly 55-60% of the category's annual volume in India lands in a six-to-eight-week window spanning the pre-monsoon transition and the first six weeks of monsoon proper, when viral upper-respiratory complaints and allergic symptoms spike together. Outside that window, demand is steady but unremarkable. The brand's existing paid channels weren't built to flex with that curve — Google Ads budgets were raised manually each June with a rough multiplier, and by the time the account had relearned the season's keyword patterns, three to four weeks of the peak window had already passed at a higher-than-necessary cost per outcome.
The brand's regulatory affairs lead, brought into the very first scoping call, set the real constraint before any media planning began: this was an OTC product, not a prescription pharmaceutical, but that didn't loosen the UCPMP overlay the way the marketing team initially assumed. The 2024 code applies to marketing practices for pharmaceutical products broadly, and the brand's own legal counsel had advised treating OTC advertising with the same discipline as any prescription-adjacent messaging — no disease-cure claims, no implied superiority over named or unnamed rival formulations, and no promotional patterns that could read as incentivised placement. That constraint shaped everything ICG built into the Growth-tier account from day one.
The ChatGPT Ads campaign structure ICG designed
ICG structured the account around a single Growth-tier conversation bucket for the cough-and-cold portfolio, but built it with two distinct pacing modes baked in from launch — a baseline off-peak mode and a pre-monsoon-to-monsoon surge mode — rather than treating seasonality as something to react to once volume moved. This is the core design choice that separates a seasonal Growth-tier build from a flat always-on one: the bid ladder, the daily budget cap and the creative rotation were all pre-configured to shift on a calendar trigger roughly three weeks ahead of the historical demand curve, rather than waiting for live volume data to confirm the spike had arrived.
Conversation-stage bidding in the off-peak baseline stayed almost entirely in the early-exploration band — users asking general questions about cough types, home-remedy comparisons, or when a cough warrants seeing a doctor. These conversations bid low and the brand's OTC portfolio appeared as a cited, factual mention within the assistant's response rather than as an aggressive sponsored push, consistent with how ICG approaches this platform for any product that doesn't want to look like it's chasing symptomatic users mid-illness. As the pre-monsoon trigger date approached, the bid ladder shifted deliberately toward mid-intent conversations — users who had moved from "what is this symptom" to "what OTC options exist for this" — without following users all the way to late-intent, purchase-adjacent conversations, because that territory is where UCPMP's comparative-claims restrictions bite hardest and where ICG's compliance review is most conservative about approving copy.
Landing infrastructure was built as three short, machine-scannable pages — one per product line (syrup, lozenge, fast-dissolve tablet) — each stating factual, on-label information (active ingredient category, indicated use, standard dosing guidance framed generically rather than personalised) with no comparative language and no unverifiable claims about speed or efficacy relative to alternatives. Each page routed to a stockist-locator tool and, where available, a click-through to the brand's listed e-pharmacy partners, rather than to any kind of gated lead form — an OTC brand's ChatGPT Ads funnel doesn't collect personal health information the way a clinic's does, and ICG deliberately kept the funnel that light.
Creative rotation ran on a two-week cycle through the surge window specifically so that no single ad variant sat in front of the same regional audience long enough to start reading as repetitive symptomatic targeting — a pattern ICG has found regulators and platform reviewers alike are more sensitive to during high-volume illness seasons than during the rest of the year. Reporting ran monthly during the baseline period and moved to a weekly cadence for the eight-week surge window, reconciled against the brand's own stockist-locator click data and, where the e-pharmacy partners shared it, downstream listing-page traffic.
The compliance discipline
DCGI/UCPMP 2024 was the governing framework for every ad variant in this account, and the review discipline ICG applied treated the OTC status of the products as no reason to relax it. Every piece of copy went through a two-stage check before launch: first against the code's core restrictions — no disease-cure or symptom-elimination guarantees, no claims of superiority over named or clearly implied rival formulations, no urgency or fear-based framing designed to push a purchase decision during a period of genuine physical discomfort — and second against the brand's own regulatory affairs sign-off, which in this engagement was stricter on several points than the baseline code, particularly around any language that could be read as implying the product prevented illness rather than managed symptoms already present.
A specific discipline point worth naming: ICG's copy review explicitly excluded any claim pattern that implied the product worked faster or more completely than competing formulations, even in a generic, unnamed sense ("faster relief than other syrups"), because UCPMP's comparative-claims restriction doesn't require a competitor to be named for a claim to fall foul of it — implied superiority against the category as a whole is treated the same way. This meant several early creative drafts from the brand's own agency-of-record, carried over from television scripts, were rejected outright before they reached the ChatGPT Ads platform, because language that reads as acceptable in a 30-second broadcast spot doesn't automatically clear a stricter conversational-ad review standard.
ASCI Chapter III applied as a secondary layer alongside UCPMP, reinforcing the same no-guarantee, no-fear-framing principles from the advertising-standards side rather than the pharma-marketing side, and DPDP 2023 governed the stockist-locator tool's handling of any location data a user shared to find nearby retailers — explicit, minimal-scope consent, no retention beyond the immediate lookup, and no data-sharing with third parties. Because the funnel never collected personal health information, DPDP's more demanding obligations around sensitive health data didn't apply in the way they would for a clinic's enquiry form, but ICG still ran the locator tool's consent language through the same review standard it applies to any healthcare-adjacent data collection, since regulators and platforms alike tend not to distinguish precisely between "pharma" and "healthcare" data sensitivity in practice.
The 90-day outcome pattern
In the pattern ICG has observed across comparable seasonal OTC Growth-tier accounts, the baseline period (roughly the first four weeks, before the pre-monsoon trigger) ran at modest, steady volume — illustrative figures put completed conversations in the range of 15-20 a week, mostly early-exploration queries that mentioned the brand's products as one factual option among several within the assistant's response. This period existed less to drive volume than to season the bid ladder and let the compliance-cleared creative set accumulate enough conversation data that the surge-mode transition wouldn't be starting cold.
The pre-monsoon trigger, timed roughly three weeks ahead of the historical demand curve based on prior-season retail sell-through data the brand shared, produced a visible ramp within the first ten days — completed conversations rising from the baseline 15-20 a week into the 60-80 a week range as the bid ladder shifted toward mid-intent queries and daily budget caps lifted on schedule. By the peak monsoon weeks (illustrative weeks 7-10 of the 90-day window), completed conversations ran 140-160 a week, with a materially higher share landing in the mid-intent band the account had been deliberately built to capture — users actively comparing OTC options for symptoms they were currently experiencing, without the account chasing them into the late-intent, purchase-decision territory that UCPMP's comparative-claims restrictions made too risky to bid aggressively for.
Cost per completed conversation during the surge window ran higher than the baseline period by a factor of roughly 3-4x, consistent with the conversation-stage bid ladder mechanics — more of the spend sat in mid-intent bands that simply cost more per event than early-exploration ones. But cost per stockist-locator click-through (the tracked downstream outcome) actually improved slightly through the surge window relative to baseline, because a larger proportion of surge-window conversations converted to an actual locator lookup — users arriving mid-symptom with an immediate, practical need rather than general curiosity. In GA4, this traffic aggregated under the AI Assistant channel; illustrative figures put its click-through-to-locator rate at roughly double the brand's equivalent Google Ads generic-term traffic over the same eight-week window, though direct e-pharmacy conversion data remained outside ICG's visibility since the brand's e-pharmacy partners didn't share purchase-level reporting back.
The brand's regulatory affairs lead's post-campaign review, conducted at the 90-day mark, found zero ad variants across the full seasonal run required a post-launch takedown or revision — a result the brand's marketing head specifically flagged as a contrast with a prior season's television campaign, which had required one script revision mid-flight after an initial regulatory query.
What we'd do differently next time
The pre-monsoon trigger, set three weeks ahead of the historical demand curve, still left the account seasoning its mid-intent bid ladder during the first ten days of actual ramp rather than arriving at the peak with a fully learned ladder already in place. In retrospect, starting the surge-mode transition four to five weeks ahead rather than three would have let the platform's bid-learning process complete before the highest-cost, highest-value weeks arrived, likely improving cost-per-outcome during exactly the period where volume — and spend — was highest.
The compliance review process, while it produced a clean zero-takedown result, ran slower than ideal during the first two weeks of the engagement because the brand's regulatory affairs lead and ICG's compliance reviewer hadn't yet built a shared shorthand for which claim patterns would clear quickly versus which needed escalation. A pre-launch calibration session — walking through five or six borderline example claims together before any real creative was drafted — would have compressed that early friction considerably.
Finally, the decision to exclude late-intent, purchase-adjacent conversation bidding entirely, made for sound compliance reasons, meant the account likely left some genuinely high-intent volume uncaptured during the peak weeks — users who had moved past comparison and were actively asking where to buy. A narrower, more carefully compliance-reviewed carve-out for that late-intent band, rather than excluding it wholesale, is worth testing on a repeat engagement, provided the ad copy for that narrow band goes through an even stricter review pass than the rest of the account.
How this maps to your own pharma brand
If your OTC or pharma-adjacent brand carries a similar seasonal demand curve — allergy products through spring, digestive-health products around festival-season eating patterns, or any category with a predictable annual spike — the pre-built dual-pacing structure described here is the template worth adapting: a steady baseline mode and a pre-triggered surge mode, rather than a flat always-on campaign that reacts to volume after the spike has already started. If your product carries no strong seasonality, a simpler single-pacing Growth-tier build, closer to this account's baseline period alone, is likely the better starting point.
Whatever the shape, the UCPMP-clean discipline doesn't flex by product category or OTC status — DCGI/UCPMP 2024 and ASCI Chapter III apply to the full range of pharmaceutical and pharma-adjacent advertising, and treating an OTC product's compliance bar as lower than a prescription product's is the single most common mistake ICG sees brands make when they first bring a ChatGPT Ads account in-house or hand it to a generalist agency.
