Turning an account at risk into an experimentation engine
+61%ROI
4.5xmonthly spend
~3structured tests a week
An account close to offboarding
ClientMax Fashion (Landmark Group), a national fashion retailer across Tier 1–3 cities in India
My rolePrimary account manager
Timeframe2018–19
The account was close to offboarding. Performance swung week to week, every attempt to scale broke results, and decisions were reactive.
There was no system for learning. Changes were made several at a time, so nobody could say what had worked.
One variable at a time, about three tests a week
I set up an experimentation engine running about three tests a week, each isolating one variable: city tier, targeting logic, placement, creative format or optimization objective. Every test was documented for transferable lessons, not just that week’s result.
What the tests found
- Tier 1 cities justified 30–40% higher CPMs through stronger downstream conversion.
- Android consistently beat iOS on efficiency.
- Some placements failed regardless of budget or creative, so they were excluded by default.
- Creative structure moved CTR and conversion more than targeting did.
From near-offboarding to one of the agency’s biggest accounts
- ROI up 61% over my time on the account.
- Monthly spend grew 4.5x, making it one of the agency’s highest-spending accounts.
- The method was adopted as an agency-wide playbook.
- The account was whitelisted by Meta for Dynamic Product Ads and Broad Audience Ads.
A good learning system is worth more than any single insight it produces.
Hiring for a role like this?astitvasawhney@gmail.comMore ways to reach me
Next caseMore work