All work
Core PM Consumer · iOS · Android · Wearables May 2025–present

Turning daily habits into durable retention

At Hevy, a fitness tracking app with 14M+ users, I own the Core product area end to end. Three initiatives shipped: Contact Sync, Streaks redesign, and Year in Review.

My role Senior PM, Core
Platform iOS, Android, Wearables
Tools Amplitude, Snowflake, Firebase
Timeline May 2025–present
TL;DR
Weekly sessions
+26% after Contact Sync rollout
90-day retention
+19% improvement across social cohorts
30-day retention
+9% from Streaks redesign, A/B tested
December acquisition
+21% from Year in Review, no paid spend

01 — Context

A product at scale with a retention problem

Hevy had strong acquisition and a passionate community of lifters, but engagement data told a more nuanced story. A significant portion of users activated, logged a few workouts, and drifted. The Core area I owned included the social layer, streaks mechanics, and seasonal touchpoints.

The question was not how to get more users, but how to make the product feel indispensable to the ones already there.

02 — Problem

Three distinct failure modes, three initiatives

Through Amplitude funnel analysis and retention cohort work, I identified three separable problems: users were not connecting with others on the platform (social graph was thin), streak mechanics had low awareness and weak recovery paths, and there was no end-of-year hook to drive both acquisition and reactivation.

Each problem mapped to a separate initiative with its own hypothesis and success metric.

◎
Key insight
Users with at least one connection on Hevy churned at significantly lower rates than isolated users. Social graph density was the single strongest leading indicator of 90-day retention in the dataset.

03 — Key decisions

What I chose to build, and why

01
Contact Sync prioritized first Contact Sync was prioritized first because social graph density is a leading indicator of long-term retention. Users with connections churned at significantly lower rates than isolated users.
02
Streaks redesign focused on recovery Streaks redesign focused on recovery mechanics, not just streak display. Most churn happened the day after a missed streak. Adding a grace period and a visible recovery path changed the loss framing entirely.
03
Year in Review as a distribution channel Year in Review was scoped as a zero-incremental-spend acquisition driver. Shareable personal stats turn the existing user base into a distribution channel.

04 — Results

Measured, not estimated

+26%
Weekly sessions after Contact Sync rollout, measured over 60-day post-launch window in Amplitude
+19%
90-day retention improvement across cohorts exposed to the social graph feature
+9%
30-day retention lift from Streaks redesign, A/B tested against control group
+21%
December acquisition spike driven by Year in Review social shares, no paid spend

05 — Reflections

What I would do differently

01
Contact Sync needed a stronger onboarding nudge Discovery was too passive and delayed time-to-value for users who would have benefited most. A proactive prompt during first-run would have accelerated social graph density.
02
Year in Review should be planned in Q2, not Q3 The engineering crunch in November compressed testing time and limited the share mechanics we could ship. Earlier planning would have unlocked a richer, more testable feature.
Next case study
Yendoplan — from 1.5% to 18% retention