All work
Founder · Product Lead Marketplace · Social · B2C 2021–2023

From zero to 60K users, and a retention turnaround

Yendoplan was a social ticketing marketplace I co-founded and led product on. The hardest problem was not growth — it was keeping users after they bought their first ticket.

My role Co-founder, Product Lead
Type B2C Marketplace
Tools Amplitude, Firebase, SQL
Timeline 2021–2023
TL;DR
Total users
60K+ acquired across three markets
Monthly retention
1.5%→18% after social and post-event redesign
Scope
Full marketplace built from zero
Markets
3 countries at peak operation

01 — Context

A marketplace with strong acquisition and almost no retention

Yendoplan was built to make discovering and attending events a social experience. Users could see what events their friends were going to, buy tickets together, and share plans. Acquisition was not the problem — events created natural spikes of interest.

The problem was that after the event, users disappeared. Monthly retention sat at 1.5%, which meant we were essentially refilling a leaking bucket with every campaign.

02 — Problem

The product had no reason to exist between events

Through cohort analysis in Amplitude, I identified the core issue: the product lifecycle ended when the event ended. There was no persistent value layer. Users had no reason to open the app on a Tuesday with no upcoming plans.

The social graph was also weak — most users had attended one event alone or with one friend, so there was no network pulling them back. The retention problem was structural, not a marketing problem.

◎
Core diagnosis
The app was a ticketing tool, not a social product. Users had no reason to return between purchases, and purchases were infrequent by nature. Retention required rethinking what the product was for.
◈
What the data showed
Users with two or more connections on the platform retained at 4× the rate of isolated users. The social graph was the lever — it just wasn't being built during acquisition.

03 — Key decisions

What I chose to build, and why

01
Persistent social layer around plans, not tickets We built a persistent social layer around plans, not just tickets. Users could create plans for future events, invite friends, and track who was going. This gave the app a reason to exist before the event, not just during purchase.
02
Social feed showing what friends are planning We introduced a social feed showing upcoming events from friends. This created a passive engagement loop — users opened the app to see what their network was planning, not just to buy.
03
Post-event experience as a social artifact We redesigned the post-event experience to prompt reviews, photo sharing, and follow-up plans. The event became a social artifact, not a transaction endpoint.

04 — Results

Retention rebuilt from structural zero

60K+
Total users acquired across three markets
1.5%→18%
Monthly retention after redesigning the social and post-event layers
0→1
Full marketplace built from zero: discovery, ticketing, social graph, payments
3
Countries where the platform operated at peak

05 — Reflections

What I would do differently

01
Underinvested in social graph at launch Retention would have been stronger if we had required connecting with at least one friend during onboarding rather than making it optional. The data was clear in retrospect — connected users retained at 4× the rate.
02
Payments infrastructure delayed iteration The payments infrastructure took longer than expected and delayed our ability to test retention mechanics. In hindsight I would have used a third-party ticketing integration for the first six months and built payments later.
03
Three countries too early We operated in three countries too early. Focus on one market first would have given us cleaner data and faster iteration cycles. Multi-market noise made it harder to read which retention changes were actually working.
Next case study
Nova Hiring — LLM pipeline, 75% faster shortlisting