All work

PLG · Subscription · Activation

Finding the aha moment that grew subscription ARR by $7M

Cohort analysis surfaced an unglamorous activation signal — first box in 6 days, 50%+ worn. Every initiative thereafter pointed at it.

Company
Gwynnie Bee
Role
Director of Product Management
Timeframe
2013 — 2020
Focus
PLG · Activation · Experimentation · ML personalization
60% → 50%
Free-trial churn (now profitable)
$7M
ARR lift
+10pp
Trial-to-paid lift
−7%
CAC reduction

Summary

Built the activation engine behind a fashion subscription platform — defining the trial aha-moment, then re-pointing the whole roadmap at reaching it faster.

Context

30-day free trials looked healthy at the top of funnel but churned hard. The team had a dozen 'most important' levers and no shared definition of what activation actually was.

The insight

Activation wasn't sign-up or first box. It was first box within six days plus 50% of garments actually worn. Everything before that was noise.

Approach

  1. 01

    Started with the user, not the funnel: interviewed 8 trial members — 5 who converted, 3 who churned — then validated quantitatively. Receiving the first box within 6 days and wearing 50%+ of the garments was what predicted conversion; that became the activation aha-moment.

  2. 02

    Turned the aha-moment into cross-functional alignment — marketing, business and product had each been pulling different levers, so re-pointed every initiative at one question: does this get a member to the aha-moment faster?

  3. 03

    Prioritised three levers with RICE: surfacing in-stock garments so boxes shipped sooner (the highest-reach lever, touching every trial member), a style quiz to capture preferences for personalised recommendations, and a Size Advisor for fit.

  4. 04

    Rebuilt the Size Advisor after a failed first cut — asking for body measurements was garbage-in, garbage-out, since members didn't know their hip/waist/bust. Pivoted to asking their sizes in brands they already wore and mapping those to new ones.

Outcome

The first round of aha-moment work cut free-trial churn by 500 bps, validating the thesis. Applying the same method to the downstream sub-aha moments took churn from 60% to 50% — where the trial finally turned profitable — adding $7M in incremental ARR. The higher trial-to-paid conversion also pulled CAC down by 7%.

Reflection

"The lesson: when teams are pulling in different directions, go to the user. Concrete behavioural data — not opinions — is what finally aligned marketing, business and product on a single activation target."

Next case

Closing the AI software development context gap and winning Salesforce and Airbnb