ResultSubscription churn fell from over 14% to under 8%
Subscription churn rate
Before
14%+
One quarter later
<8%
−43% reduction in subscription churn
The challenge
Fast growth, leaky funnel
The brand was growing fast, at more than 85% a year, but beneath that growth it became clear they were really operating what I call a “leaky funnel”. They were investing so much into customer acquisition, but only achieving sub-optimal lifetime value (LTV) because they were losing customers within the year — despite their product category being a sticky one by design, a product that meets a daily need.
Subscription churn had climbed above 14%, well above their vertical’s benchmark churn rate. At that growth rate, every point of churn meant real lost lifetime value and wasted acquisition spend. No one at the company had systematically and objectively identified why customers were churning and when. Cue my entry…
What I did
From guesswork to a shared plan
1
I found out why customers were actually leaving
I ran discovery across the organization: cancellation reasons, support tickets, subscription data and conversations with frontline CX agents — bringing in internal stakeholders (my peers) but also objective multi-source data into one view. Then I mapped the subscriber journey from first order to cancellation. The one insight that surprised people most: churn wasn’t going to be fixed by one person, it was going to require cross-functional alignment and effort.
2
I changed the way the founding team and company approached churn
I stood up a cross-functional task force across CX, marketing, product, operations and packaging. I had the data to show that no single team could fix it alone, and so I secured leadership’s buy-in and confidence on how we would reach record retention rates TOGETHER.
3
Turned the findings into testable ideas
Instead of one big overhaul, I ranked the friction points by impact and effort and ran a structured set of experiments, each with a clear success measure.
4
Fixed the three biggest friction points
Each one turned a reason to cancel into a reason to stay.
The fixes
Three friction points, three fixes
The subscription portal
Changing, pausing or skipping an order was hard enough that cancelling looked like the easy way out to a frustrated customer. Making it simple to adjust a subscription turned many would-be cancellations into pauses and swaps.
A new subscription portal design that prioritized clear actions and aligned to how customers thought (e.g. calendar dates, not weeks or months) was a high-impact change. And now, with AI, it’s an even lower-lift change.
Product information on the website
Customers were cancelling over questions the website should have answered, such as feeding amounts, switching between formulas and ingredients. Clearer product information prevented mismatched expectations before they turned into cancellations.
Packaging
Receiving leaky products was a real reason for customers churning. No one was comfortable using a product that had been compromised during shipment, and the last thing they wanted to do was now run to the store last-minute for their pet because the product was no good. Not to mention: Brand trust is compromised and the company’s CX costs are quickly driven up when 30% of call volume is attributed to damaged shipments.
I learned more about Shipping, Ops, and Warehouse practices than I ever thought I would. And it turned out, the easiest fix was one I observed when I received my son’s shampoo. Preventing a product leak during shipment was probably the most impactful thing that the company could have done, and we shipped (pun intended) this change within 4 weeks!
30% of call volume came from damaged shipments4 weeks from insight to shipped fix
The results
What changed in one quarter
43%Reduction in subscription churn
14%+ → <8%Subscription churn, within one quarter
93% · 75+CSAT held at 93% and NPS above 75 while the business kept scaling
4 → 10CX team scaled (including offshore) to keep up with 85%+ growth
How I’d run this today with AI
Same fundamentals. Much faster.
The core of this work hasn’t changed: listen to customers, find the friction, test your fixes. What has changed is speed. Today I’d use AI for retention to:
Sort cancellation reasons and support tickets automatically, so patterns show up in days, not weeks
Flag subscribers at risk of churning before they cancel, using their behaviour, and send them a personalized save or pause offer
Use agentic workflows to build segments, draft and check save campaigns, and report weekly, with a human approving what goes out