A ₹1,710 test that exposed the real growth constraint.
What was happening.
TaxByKK launched a new kind of product: an MCP connector at ₹125/month that upgrades Claude into a GST research assistant for tax professionals. It's a category nobody was searching for yet.
What was really in the way.
A new category can't be sold on features. The audience first needs to understand what an "MCP connector" is and why it matters — and we didn't yet know whether broad reach or a smaller, premium audience would respond.
“A tighter Tier 1 & 2 city audience will engage far more deeply with an AI-for-GST message than a broad 20-city audience — even at a much higher CPM.”
What changed.
We split the launch into two campaigns that ran the same three creatives, so geography was the only variable. The message ladder moved from reassurance ("doesn't replace Claude — it upgrades it") to a clear price anchor (₹125/month) to curiosity ("why is everyone talking about it?").
What we shipped.
What the data says.
| Metric | 20 Cities (broad) | Tier 1 & 2 (premium) |
|---|---|---|
| Impressions | 37,832 | 7,153 |
| CTR | 1.17% | 2.85% |
| CPM | ~₹22 | ~₹120 |
| Est. spend | ~₹832 | ~₹858 |
What it meant.
The hypothesis held: the premium audience clicked at 2.4× the rate, while broad targeting delivered reach at a fifth of the CPM. But the most valuable finding wasn't in the ad account — the Meta Pixel wasn't installed, so no conversions could be measured. The next move isn't more spend. It's the Conversion layer: pixel, a dedicated landing page, then scaling the audience that buys, not the one that clicks.
Want this kind of thinking on your numbers?
Tell us where growth is stuck. We'll tell you honestly whether we can fix it — and what it would take.