Creative-led Meta Ads
that convert.

How should a brand structure creative tests on Meta?

Test different concepts first, then test variations of the winner. A concept is a different reason to buy: a problem, a comparison, a testimonial, a founder story. A variation is a new hook, cut or caption on the same idea. Mixing the two in one test tells you very little.

I keep new concepts in a separate testing campaign with a fixed budget, so the main scaling campaign isn't constantly reset by fresh ads. Each concept gets enough spend to be judged on cost per purchase or qualified lead, not on CTR alone. Winners move into the scaling campaign. Losers get a short note on why they probably failed, which shapes the next brief.

Fatigue is the other thing to watch. When frequency climbs and cost per result rises on a winner, the answer is usually a fresh variation of the same concept, not a new audience. That's why the testing pipeline has to keep running even when the account is performing well.

What breaks Meta Ads tracking for Indian ecommerce brands?

Cash on delivery is the usual culprit. The purchase event fires when the order is placed, but a share of COD orders are refused or returned, so Meta optimises toward people who order and don't pay. Browser tracking also misses events, which makes reported results look worse or better than they are.

• Conversions API alongside the pixel, with deduplication
• A separate event for confirmed or delivered orders where possible
• Prepaid vs COD tracked as different values or events
• UTMs on every ad so Shopify and GA4 can cross-check Meta
• Weekly reconciliation of Meta purchases against actual orders

Here's what I set up so the numbers you scale on are believable. Without it, you can scale a campaign that looks profitable in Ads Manager while the courier returns quietly wipe out the margin.