Implementation worksheet · 6 min read
A Subscription Renewal-Risk Lifecycle Playbook
Build the risk signal from usage change rather than usage level, and act on it with a ladder that escalates from automated to human rather than jumping to a discount. A low-usage account that has always been low-usage is a pricing fit question, not a churn signal; an account whose usage halved in three weeks is the one worth intervening on. The ladder: notice and understand, offer help, involve a human, and only then discuss commercial terms. Leading with a discount to a wavering customer teaches them that wavering produces discounts, and the ones who learn that lesson are the ones you most want to keep.
Renewal-risk programmes typically segment on low usage and send a sequence. That reliably reaches accounts that are small and happy, while missing the large account whose champion left and whose usage is declining from a high base. Level and trend are different signals, and only one of them predicts churn.
Put it into practice
1. Use change over time, not absolute level
Week-over-week or month-over-month decline in active users, core actions, or whichever metric represents real use. An account down 60% from its own baseline is at risk regardless of whether that baseline was large or small. Absolute thresholds select for small accounts.
2. Add the signals that precede usage decline
The champion's account going inactive while others continue. A support escalation that took a long time to resolve. A failed payment. A seat count reduced at a previous renewal. Several of these appear before usage drops, which is the point — by the time usage has halved, the decision may already be made.
3. Understand before you intervene
The first step is not a message to the customer. It is looking at what changed: a feature they relied on, a team reorganisation, an incident. Sending 'we noticed you have been quiet' to an account whose decline followed a week-long outage on your side is worse than sending nothing.
4. Escalate from automated to human, in that order
Automated help content for a specific observed problem, then an offer of a session with someone, then a human reaching out directly for accounts above a value threshold. Each step is more expensive and more effective; starting at the most expensive does not scale and starting at the cheapest for a large account is insulting.
5. Keep commercial terms off the ladder until the end
A discount is the last step, not the first. Offered early it converts a solvable product problem into a price negotiation, and it teaches the account that reducing usage before renewal is profitable. The accounts sophisticated enough to learn that are usually your largest.
6. Suppress the entire programme for accounts already in a conversation
If a human is actively working the account, automated risk messaging should stop. Receiving a generic retention email during a live renewal negotiation undermines the person having that conversation, and it happens constantly because the two systems do not talk.
7. Record what the intervention was and what happened
Which signal fired, what was tried, and whether they renewed. Without it the programme cannot improve, and you cannot distinguish an intervention that worked from an account that was never going to leave.
The intervention ladder
Copy this structure into your review document and record your observed result for each row.
| Signal | Step | Owner | Escalate if |
|---|---|---|---|
| Usage down >40% from own baseline | understand the cause internally | lifecycle owner | cause unclear |
| Cause identified as product friction | targeted help content | automated | no change in 2 weeks |
| No change after help | offer a session | automated | no response |
| Account above value threshold | human outreach | account owner | no response |
| Champion inactive, others active | human outreach | account owner | — |
| Failed payment near renewal | billing-led contact | billing | unresolved |
| Support escalation unresolved >X days | human outreach | support + account owner | — |
| All steps exhausted | commercial conversation | account owner | — |
| Human conversation in progress | SUPPRESS all automation | — | — |
A failure worth checking
The discount offered first. An account shows declining usage, the programme fires, and the message leads with a retention offer. Sometimes it works, and what it taught is that usage decline before renewal produces a better price. The accounts most capable of learning that are the large, sophisticated ones — so the programme trains exactly the customers you can least afford to train. The ladder exists so the commercial conversation is the last resort rather than the opening move.
Common questions
What if we cannot tell why usage dropped?
Then ask, plainly, without an offer attached. 'We noticed things have changed — is there something we should know?' is a better message than a discount and a much better one than a generic re-engagement sequence. The answer is often something you can fix.
Should low-usage accounts be contacted at all?
Consistently low usage is a fit or pricing question rather than a churn signal, and it deserves a different conversation — possibly a smaller plan. Putting them into a renewal-risk sequence generates volume and very little insight, because they are not at risk in the way the programme assumes.
Basis and scope
This is a proposed implementation method using illustrative examples, not a measured benchmark or a customer case study. Prepared with AI assistance. Validate product-specific behavior against current documentation and your own test environment.