"Aman, I've spoken to six agencies in the last year. Five of them said yes to everything I asked. That's why I don't trust them."
That was the founder of Derma Skin & Hair Solutions, sitting across from me in his Saket clinic in February 2024. He hired us three weeks later. After I told him on the second call that his Friday discount campaign idea would tank his average ticket size and I wouldn't run it.
By month 8 we'd hit 8.4x ROAS on a single campaign. By month 14 we'd helped him open his third location in Greater Kailash. This is how it actually went — including the month 4 disaster I still bring up when new clients ask why we charge what we charge.
What Derma Skin & Hair Looked Like When We Took Over
Two locations: Saket and Vasant Kunj. Both running for 6+ years. Strong walk-in traffic from the immediate catchment, weak everything else.
Their existing setup:
- Instagram: 8,200 followers, posting whenever someone in the team remembered
- Meta Ads: ₹25,000/month with a previous agency, ROAS hovering around 2.1x
- Google Business Profile: claimed but unmanaged — last post in 2022
- Website: built in 2019, mobile PageSpeed score of 41
- Conversion tracking: half-broken pixel, no Conversions API, no offline conversion uploads
Average treatment ticket size at the time: ₹6,800. Mostly HydraFacial, laser hair reduction, chemical peels.
The founder's brief was simple: "I want to open a third location in Delhi by year-end. I need the marketing to be predictable enough that I can plan around it."
That word — predictable — is the one I keep coming back to. Most clinic owners don't actually want 12x ROAS for one month. They want 5x ROAS every month for two years. Different goal. Different system.
The First Decision That Mattered
Before any creative, before any ad spend, we did three things in week 1:
Rebuilt conversion tracking from scratch. Old pixel was firing on every page view as a "Lead" event. Garbage data. We installed a fresh Meta Pixel + Conversions API setup, defined three real conversion events (form fill, WhatsApp click, call click), and started uploading offline conversions weekly from their CRM so Meta could see which leads actually became patients.
Audited the menu. This is the step nobody talks about. We pulled their last 12 months of revenue by treatment. HydraFacial was 31% of revenue. Laser hair reduction was 24%. PRP for hair was 18%. Everything else was noise. We decided to advertise only those three for the first 90 days. Three campaigns. Not seventeen.
Killed the discount addiction. They had been running 20% off promotions every month. We stopped all of it. The founder hated this for the first four weeks. By month 3 he stopped asking.
Total cost of week 1 work: zero rupees in ad spend. We hadn't launched a campaign yet.
Month 1 to Month 3 — The Climb
Launched mid-February with a ₹45,000/month budget split across the three campaigns.
Week 1: ROAS 1.4x. Founder texted me asking if this was normal.
Week 3: ROAS 3.2x. He stopped asking.
Week 6: ROAS 5.1x. He asked if we could scale the budget.
By end of month 3 we were at 6.8x sustained ROAS, ₹62,000/month spend, generating about 95 qualified consultation leads per month across both clinics. Cost per qualified lead had dropped from ₹820 (the previous agency's number) to ₹430.
What made the difference wasn't the ad spend. It was the creative volume. We produced 24 ad variants in the first 90 days using Midjourney + Higgsfield Cinema Studio + Kling AI for motion. That's roughly 8x what a traditional creative team would have produced at the same budget. About 4 of those 24 variants became the workhorses. The other 20 we killed within 14 days each.
This is the part most clinic owners don't understand about AI in marketing. AI doesn't make better creative. It makes more creative, faster, so you can find the few that work before you run out of money.
Month 4 — The One That Almost Killed The Account
May 2024. ROAS dropped from 6.8x to 2.4x in 19 days.
I'll admit it — I genuinely thought we'd done something wrong. I went through every campaign three times looking for what we'd broken.
Here's what had actually happened:
- Meta released their new Advantage+ Audience targeting and our existing ad sets started bleeding budget into low-intent audiences
- A competitor in Saket had launched a ₹999 HydraFacial offer (we found out from a patient who walked into Vasant Kunj asking why we were "more expensive")
- The clinic itself had a 2-week shortage of one of their imported HydraFacial serums and the front desk was rescheduling consultations
Three problems wearing the same costume.
We fixed them in this order:
The serum supply. Not our problem to fix, but we paused the HydraFacial campaign for 12 days while supply came back. This alone recovered 35% of the ROAS drop because we stopped paying for leads we couldn't convert. (Front desk wale bole: "sir, slots toh hain par serum nahi hai." That's the kind of problem no ad campaign solves.)
The competitor pressure. We did NOT match the ₹999 offer. Instead we launched a "Why we don't run ₹999 HydraFacial offers" Reel from the founder himself, talking about serum quality and post-care. It got 47K organic views in a week. Patient enquiries from Saket actually went UP that week, not down.
The Advantage+ leak. Migrated to Advantage+ Shopping Campaigns with explicit creative exclusions and tightened the geo radius from 8km to 5km around each clinic.
By end of June we were back to 5.9x. By July we hit 8.4x on a single campaign — the highest we've ever pulled for a derm clinic. Sustained 6.5x average for the rest of the year.
Opening Location 3
October 2024. Founder asked: "If I open a third location in Greater Kailash, can your system handle it?"
We took 4 weeks to plan it. Not because the marketing was complicated, but because we wanted to launch with full creative + tracking + GBP optimization on day 1, not retrofit it.
Launch plan:
- Cloned the campaign architecture from the existing two locations
- Built a 5km zip-code targeted audience around the new location
- Created location-specific creative (the GK clinic interior shot integrated into existing AI-generated sequences)
- Set up the GBP for the new location with 30 photos and 12 service descriptions before opening day
- Pre-loaded a "Now open in GK" creative across all three locations to push patients toward the closest branch
The GK location hit 4.2x ROAS in week 6. By week 12 it was at 5.8x. By March 2025 it was tracking the same numbers as the older locations.
The replication worked because we had 8 months of conversion data on the Meta pixel teaching the algorithm what a "good Derma Skin & Hair patient" looks like. New location, same patient profile, transferable signal.
The Year-End Numbers
Derma Skin & Hair Solutions — 2025 Performance (3 locations)
- Total Meta Ads spend ₹14.4L
- Attributed treatment revenue ₹89.28L
- Sustained ROAS 6.2x
- Peak single-month ROAS 8.4x
- Avg cost per qualified consult ₹520
- Consult-to-treatment conversion 41%
- New patients acquired 1,847
The 8.4x number is the one we put in the case study deck. The 6.2x sustained number is the one I quote on sales calls. They're both real, but only one of them is what a clinic owner should plan around.
Three Things I'd Do Differently If Starting Today
One. I'd start the offline conversion uploads from week 1, not week 4. We waited because we were "focused on getting ads live first." That cost us roughly 3 weeks of optimisation time because the algorithm was learning from form-fills instead of from actual paying patients.
Two. I'd build the WhatsApp auto-reply flow before launching the ads, not after. The first month we generated 95 leads but the front desk could only follow up with about 60 of them in time. The other 35 went cold. We built the auto-reply flow in month 5. Should have been month 0.
Three. I'd push back harder on the founder's request to advertise everything. The first version of his brief had us promoting 9 different treatments. We negotiated down to 3. Should have been 2 for the first 60 days, 3 from day 60, and never more than 5 at once. Spreading creative spend across too many campaigns is the most common reason clinics see flat ROAS in their first 90 days.
Should You Try This Approach For Your Clinic?
Honest answer: depends on three things.
Your average ticket size matters first. If your average treatment ticket is below ₹3,000, the math doesn't work at our agency rates. Meta ads cost roughly the same regardless of what you sell. You need a high enough ticket to make the cost-per-lead profitable. ₹4,000+ is where it starts working. ₹6,000+ is where it sings.
Your clinic's conversion capacity is the second filter. We can generate 100 qualified leads a month. If your front desk converts 12 of them because nobody's calling back inside 4 hours, you'll burn money and blame the ads. Fix the front desk first or hire us to fix it for you (we now do this for 4 of our 12 clients).
Your appetite for the first 60 days is the third. The Meta pixel needs about 50 conversion events to exit the learning phase. During this time, your ROAS will look bad. If you can't hold your nerve through that, paid ads aren't for you. Run organic content for 6 months, build a base, come back when you can commit.
Want to talk about your clinic?
Free 30-minute strategy call. No pitch — just an honest read on where you are and whether what we do fits.
Most agency content is written to make the agency look smart. I tried to write the version that would have actually helped me when I was running Derma Skin & Hair's first campaign and staring at a 1.4x ROAS in week 1 wondering if I'd over-promised.
— Aman