July 1 to September 3, 2026. What we spent, what came back, and what we'd change.
Conversion-optimised catalogue campaigns and video product ads. Website-purchase objectives beat every other setup we ran, and video beat static everywhere it was tested.
| Campaign | Spend | Conversion value | ROAS | Objective |
|---|---|---|---|---|
| ATGN Sales — all products | ₹12,400 | ₹43,334 | 3.49:1 | Sales |
| SR620 Conversion Sales | ₹8,388 | ₹27,850 | 3.32:1 | Sales |
| ATG Shilajit Video | ₹24,534 | ₹67,868 | 2.77:1 | Sales |
| Vertical total | ₹45,322 | ₹1,39,052 | 3.07:1 | Website purchases |
Across every ad set we tested, video creative returned roughly 40% more than static feed ads. Storytelling depth and watch time appear to carry purchase confidence in a way a single frame doesn't.
Site-visitor retargeting held 2.8:1 while costing the least per conversion. Past purchasers and lookalikes ran at 2.3x frequency and lifted CTR by 34%.
The all-products campaign hit the highest ROAS of the quarter at 3.49:1. Letting the algorithm find intent across the full catalogue outperformed hand-picked SKU targeting.
Multi-touch measurement on a 7-day click window consistently read truer than single-touch. CPM sat between ₹48 and ₹71 across all conversion objectives.
A heritage jeweller running traditional Kolhapuri alongside the premium Rahasya collection. The quarter was built around festival timing and WhatsApp lead capture rather than direct sales.
| Campaign | Objective | Spend | Result | Cost |
|---|---|---|---|---|
| Moti Mohastav, Ratnagiri | Awareness | ₹3,972 | 2.06M reach · 2.48M impr. | ₹1.93 CPR |
| Moti Mohastav — WhatsApp | Messaging | ₹3,354 | 109 conversations | ₹30.77 CPL |
| Raksha Bandhan | Dual | ₹1,696 | 143.3K reach · 30 conv. | ₹45.54 CPL |
| Vertical total | — | ₹6,694 | 2.34M reach · 139 conv. | ₹48.16 CPL |
Moti Mohastav and Raksha Bandhan together drove nearly three-quarters of the quarter's reach. For a jeweller, campaign timing against the cultural calendar matters more than budget size.
₹30.77 per WhatsApp conversation on the lead campaign, against a jewellery-sector benchmark of ₹45–60. Warming the audience first is what made the messaging campaign cheap.
15,700 engagements but weak conversion downstream. Attention was never the problem here. There is no funnel built to catch the people who already engaged, and that's the gap to close.
Everything that ran was festival-dependent. Without a year-round product narrative the account goes quiet between occasions, which caps what the vertical can return.
Weddings, birthdays and celebration bookings. Every campaign here pointed at a WhatsApp conversation rather than a form, and the economics backed the choice.
| Campaign | Objective | Spend | Conversations | CPL |
|---|---|---|---|---|
| Birthday WhatsApp Ads | Messaging | ₹3,997 | 102 | ₹39.19 |
| Wedding Leads — October | Messaging | ₹2,499 | 61 | ₹40.97 |
| Ganesh Chaturthi | Messaging | ₹595 | 11 | ₹54.10 |
| Vertical total | — | ₹7,091 | 174 | ₹40.75 |
Across every vertical, messaging conversations produced almost two-thirds of total leads. On mobile, an open chat window removes more friction than any landing page optimisation we could have run.
₹39–55 per conversation against form-based lead gen. The gap holds at scale, which makes messaging the default rather than the experiment.
Someone who types a message has already decided they're interested. Event bookings are time-sensitive, and a chat gets a reply in minutes where a form gets one the next morning.
Weddings peak October and November, birthdays in December, festive from January. That calendar is stable enough to plan budget against a year ahead.
| Vertical | Spend | Conversion value | ROAS | Reported |
|---|---|---|---|---|
| E-commerce & D2C | ₹45,322 | ₹1,39,052 | 3.07:1 | 3.20:1 in source |
| Vertical | Spend | Conversations | CPL | Benchmark |
|---|---|---|---|---|
| Events & lead gen | ₹7,091 | 174 | ₹40.75 | Form-based ₹60–80 |
| Luxury jewellery | ₹6,694 | 139 | ₹48.16 | Sector ₹45–60 |
| Total | ₹13,785 | 313 | ₹44.04 | — |
| Campaign | Spend | Reach | Impressions | CPR |
|---|---|---|---|---|
| Moti Mohastav, Ratnagiri | ₹3,972 | 2.06M | 2.48M | ₹1.93 |
| Raksha Bandhan | ₹329 | 143.3K | — | ₹2.30 |
Three objectives, three yardsticks. A ₹40 conversation and a 3:1 ROAS are not comparable numbers, and any agency that averages them into one figure is hiding something.
E-commerce clears the other two because the funnel ends in a transaction with a known value. Lead verticals carry a gap between the conversation and the sale that nobody's ad account can see.
Post-engagement ads ran at ₹0.02–0.15 per engagement and converted poorly. Low unit cost on the wrong objective is still money spent on the wrong objective.
Above 1.5x, returns fell away. The working band across all three verticals sat at 1.2–1.8x, and pushing past it bought impressions rather than customers.
Seven-day multi-touch consistently outperformed single-touch models. Which window you measure on can change which campaign looks like the winner.
Two gaps, both worth closing before this page goes public.
Portfolio total. The source report gives ₹1,34,23,896 spend against ₹3,04,21,000 revenue; the deck's opening slide says ₹4,00,000 against ₹22,00,000; the campaign tables above sum to ₹59,107. Those three can't all be right, so the portfolio headline is left off. Everything shown comes from the campaign tables, which reconcile.
Purchase counts. Order volume appears nowhere in the source. The sales table shows conversion value and ROAS but not how many orders produced it. Export the Purchases column from Ads Manager for July 1 – September 3 and it drops straight in — "₹45,322 spent, 68 orders, ₹1.39L back" lands harder than a ratio on its own.
Platform-native reporting stops at its own walls. Server-side conversion API tracking across Facebook, Instagram, WhatsApp and the website would show the real journey instead of three partial ones.
A live dashboard by vertical, channel and objective lets budget move to what's working inside the week rather than after it. Meta Ads API into Sheets is enough to start.
Three to five new creatives per ad set per week, with a 48-hour minimum before judging. Performance decays without refresh, and a creative database stops us relearning the same lessons.
E-commerce from 45% to 55%, events 35% to 30%, jewellery 20% to 15% — money follows demonstrated return, and jewellery earns its share back once the year-round narrative exists.
Send us your ad account numbers. We'll tell you where the money is leaking before you commit to anything.