Building Customer Success Systems That Cut Churn by 44%
A food subscription business was losing nearly a third of subscribers every month with no system for spotting at-risk customers. We built a customer success function from scratch: playbooks, tooling and measurement. Monthly churn dropped by 44%.
By the numbers
What we walked into
At 32% monthly churn, the business was essentially on a treadmill, acquiring customers only to lose them before they became profitable. The support function was reactive, response times were slow, and there was no NPS measurement or early-warning system for customers about to cancel.
How we did it
Understand who churns and why
We started with exit interviews, 40 calls with customers who had cancelled in the previous 60 days. The patterns were clear: poor onboarding experience, slow response to complaints, and no proactive outreach in the first two weeks. We also found that customers who received a personal check-in message by day 7 had 3× the D45 retention of those who didn't.
Create the infrastructure
We built a customer health scoring model in Google Sheets, integrated with Intercom to flag at-risk customers. We wrote onboarding sequences, complaint resolution playbooks, and escalation protocols. Every interaction type had a documented response framework, so quality didn't depend on who picked up the ticket.
Track what matters
We set up NPS measurement at D14 and D30, response time tracking in Intercom, and churn cohort analysis in Google Sheets. The weekly CS review became a data session: what are the numbers saying, and what do we do about it? Within 12 weeks, churn had dropped from 32% to 18%, and NPS had jumped from 22 to 61.
“The thing I appreciate most is that she speaks in data. Every recommendation had a measurement framework attached. That's rare in lifecycle marketing.”
Want results like these?
Let's talk about what's broken in your lifecycle and what it would take to fix it.