← All work
ZeptoHead of Ads & Monetization2024 — 2025

Zepto Ads

₹240cr → ₹1,200cr ARR in 10 months
AI/MLMarketplaceAdTech1 → 10
Context

Zepto is one of India’s fastest-scaling quick-commerce platforms. The ads business existed but was early: revenue was real, the surface area was growing, yet monetization lagged the demand on both sides of the marketplace. I joined to own ads and monetization end to end — product, pricing, and the systems that decide what gets shown, to whom, at what price.

The Problem

A quick-commerce ad business growing fast but monetizing below its ceiling, with advertiser RoAS and platform take both under target.

Approach

The category treated ad-to-GMV ratio as a fixed constraint. I treated it as a design problem.

Reframed monetization as a marketplace-balance problem, not a sales problem — every incremental ad slot had to clear against organic relevance so the customer experience never paid for revenue.

Rebuilt the auction and placement logic so advertiser RoAS and platform take moved together rather than against each other. When advertisers win, they spend more; the system was tuned to make winning the default.

Instrumented the full DIFM loop (Demand → Inventory → Fill → Measurement) so every lever was observable and the team could reason about second-order effects before shipping.

Sequenced GTM so supply (advertiser onboarding, self-serve) and demand (high-intent surfaces) scaled in lockstep — avoiding the classic ad-platform failure of selling inventory that doesn’t convert.

Outcomes
5x
Ad ARR growth — ₹240cr to ₹1,200cr in 10 months
Built the monetization engine and GTM motion that compounded revenue without degrading the customer experience.
6.5%
Ad-to-GMV ratio, up from 3.5% (target was 5%+)
Beat the structural ceiling the category assumed was fixed.
4x
RoAS lift for participating advertisers
Better targeting and placement made spend efficient — so advertisers re-upped instead of churning.
What I learned

The hardest part wasn’t the auction — it was holding the line on customer experience while revenue was compounding 5x. Every ₹ of ad revenue is a small tax on relevance, and the temptation to over-monetize is strongest exactly when the numbers are going up. The discipline that worked: make the ad-to-GMV ceiling an explicit, defended number, not an emergent one. If I did it again I’d instrument customer-side guardrails even earlier — before the revenue curve made them inconvenient to add.

Related writing