← Case files
Case File 02Ad-Tech

Yektanet

Scaling product systems across ad-tech, automation, experimentation, and 0→1 growth

Associate PM → Product Manager · 2019–2022

Worked across CRM automation, advertiser UX, retargeting optimization, and a 0→1 e-commerce product at scale.

~30% growth in average daily ad revenue
25M+ unique users on the platform
5B monthly impressions
~1,000 e-commerce customers converted
Signal

Iran's largest ad network at 25M+ users had manual CRM, a leaky advertiser UX, rule-based retargeting, and an un-monetized creator tool.

System

There was no automated lifecycle, no structured way to convert support friction into product fixes, and no experimentation muscle around delivery.

Decision

Treat each phase as a distinct product surface — automation, UX, algorithm, then 0→1 — and prioritize what would actually move advertiser outcomes.

Build

Designed lifecycle automation for 20,000+ accounts, shipped UX changes informed by ticket-pattern analysis, ran dynamic segmentation experiments to move retargeting from rules to algorithm, and built a 0→1 e-commerce motion on top of a creator tool.

Outcome

~30% growth in average daily ad revenue, ~7% drop in key support ticket categories, ~10% lift in sign-up conversion, and ~1,000 customers converted before a strategic pivot.

Learning

Track downstream metrics — advertiser retention, renewal — from the start, not after the algorithm proves out short-term revenue.

Detailed file

Context

Yektanet is Iran's largest online advertising network — 25M+ unique users, 5 billion monthly impressions. Over nearly 3 years I moved across 4 distinct product areas.

Phase 1 — CRM & marketing automation

Problem

20,000+ advertiser accounts with no systematic way to onboard, nurture, or re-engage them. Communications were manual and inconsistent.

What I did

Built the company's marketing automation system from scratch. Defined automation rules across the full customer lifecycle — welcome sequences, activation nudges, re-engagement flows — covering 20,000+ business accounts via email and SMS.

Lesson: Automation at scale requires thinking in segments, not individual messages. The hardest part wasn't the tooling — it was deciding which user behaviour should trigger which response.

Phase 2 — Advertiser panel UX

Problem

High volume of repetitive support tickets signalling friction in the advertiser self-serve experience.

What I did

Combined Google Analytics funnel data with support ticket analysis to locate friction points. Shipped UX improvements including a structured FAQ before ticket submission — reducing ticket volume for key complaint categories by ~7%.

Lesson: Support tickets are underrated as a product signal. The key is categorising them at scale to find patterns, not treating each one individually.

Phase 3 — Retargeting & algorithmic optimization

Problem

The retargeting product was delivering ads using manual frequency and timing rules. Revenue was being left on the table.

What I did

As sole PM working with a data science team, designed experiments around ad frequency caps, timing windows, and impression pacing. Rather than fixed A/B splits, we segmented users dynamically across scenarios — shifting traffic allocation as performance signals emerged, similar to a multi-armed bandit approach. Transitioned delivery from manual rules to an algorithm, contributing to ~30% growth in average daily ad revenue.

Technical highlight: Dynamic user segmentation across experiment scenarios — traffic allocation shifted in real time as performance signals emerged.

Lesson: The shift from rule-based to algorithm-driven delivery is as much a stakeholder challenge as a technical one. Getting sales and ops to trust the model required showing them the experiment data, not just the outcome.

Phase 4 — 0→1 e-commerce platform

Problem

A link-in-bio tool for Instagram creators had thousands of users but no monetization path.

What I did

Built tracking dashboards to understand the conversion funnel. Improved sign-up conversion by ~10% through data-informed UX changes. Designed and executed the 0→1 e-commerce motion — converting ~1,000 users before a strategic pivot shut the project down.

Lesson: Project shutdowns aren't failures — they're strategy. We validated that conversion was possible. The business decision to stop was separate from the product outcome.
What I'd do differently: In the retargeting phase, I'd have pushed to track downstream metrics earlier — advertiser retention and renewal rates, not just CTR and short-term revenue. In the e-commerce phase, I'd instrument the funnel from day one rather than building dashboards after the fact.