Expertise · B2B SaaS
A customer rarely leaves on a given day. It drifts away over months, and often nobody sees it. I build the function that spots these accounts early enough to act: the CSM role, churn prevention.
What the function has to protect in retention, and the metrics that will show whether it does.
Spotting accounts at risk early enough to act, with a response that no longer depends on one CSM's instinct.
The first months decide what follows. Gaps between the customer's process and the product are written down at the kick-off, and someone decides who adapts.
A score built after a year of renewals has been observed, not before. In the meantime, a list of at-risk accounts kept by hand and discussed every week.
Who covers which account, with what workload, and who settles it when Sales, Product and Customer Success disagree.
Customer lifecycle, account reviews, escalation path. These fixed meetings are what let a good result repeat beyond the account where it first happened.
Hiring, growing and leading a team across several markets.
Beyond about thirty recurring accounts, when repeated patterns of churn appear.
A function that exists with no definition of a renewal and no written boundary with sales.
When renewal is endured rather than steered.
When some rules standardise and others stay local.
The method has four stages, and the last one organises my departure. See it under Working together →
A pricing SaaS vendor serving Tier-1 retailers, with no established Customer Success or Account Management function.
The function was built from scratch, from portfolio segmentation to follow-up processes and expansion strategy, then run internationally with a team of five CSMs and AMs on individual quotas.
The health score, churn risk and team sizing tools are free to use, with no sign-up. See the tools →
The other side
Preventing the departure does not grow ARR on its own. Renewal, upsell and cross-sell are the Account Manager's job, covered on the Account Management page.
Explore the Account Management expertise →Thirty minutes is enough to tell whether your topic calls for a diagnostic, a build-out or something else. No strings attached.