Renewal Risk Scorer

---
name: renewal-risk-scorer
description: Use when scoring a customer's renewal risk and recommending specific interventions. Triggers on "renewal risk", "score this renewal", "is X going to renew", "renewal triage".
---

# Renewal Risk Scorer

A renewal risk score is only useful if it drives a specific action. "Medium risk, schedule an EBR" is not an action — it's a calendar invite. This skill scores AND prescribes.

## Required inputs

1. **Customer + ARR + renewal date**
2. **Usage trend** — last 90 days, by core feature (not aggregate logins)
3. **NPS / champion sentiment** — most recent score or pulse
4. **Support volume** — ticket count, P1 count, sentiment of resolutions
5. **Exec sponsor stability** — has the original buyer / champion changed roles, left, or gone quiet?
6. **Billing signals** — late payments, procurement push for discount, contract questions
7. **Competitive signal** — have they been talking to your competitor or just got acquired by one?

If 3+ of these are missing, stop and tell the user: "Risk score will be weak with this data. Pull X, Y, Z from Gainsight/Salesforce/Zendesk first."

## Risk dimensions and what they mean

Walk through each signal and weight it:

### Usage decay
- Core feature usage down >25% over 90 days = high signal
- Login count down but core feature usage flat = lower signal (different humans using it)
- New feature never adopted after 6 months in contract = expansion-blocker, not necessarily churn risk

### Champion / exec change
- Buyer left in last 90 days and successor hasn't been engaged = HIGHEST risk signal
- Champion changed roles internally = re-sell needed, treat as new buyer
- Champion still in seat and engaged = strong protection

### Ticket pattern
- Volume spike + low CSAT = active dissatisfaction
- Volume spike + high CSAT = engaged, not at risk
- Volume drop to zero = silent disengagement, often worse than complaints

### Billing / procurement
- Late payment, asks about month-to-month, procurement-led ask for steep discount = budget pressure
- "Send me a one-pager our finance team can review" = champion under pressure to defend

### Competitive
- Acquired by company using competitor = high risk, often non-controllable
- Mention of competitor in customer conversation = test for seriousness, often a leverage play

## Output

### Risk grade
**A / B / C / D** with one-line rationale.
- **A**: low risk, expansion-mode customer
- **B**: stable, monitor
- **C**: at risk, intervene this quarter
- **D**: actively churning, all-hands save mode

### Top 3 signals driving the score
Cite specifics: "Core workflow usage down 40% MoM since March; exec sponsor (VP Ops) departed in February; champion NPS was 7 in last pulse."

### Specific intervention plan (3-5 steps, owned)
NOT generic. Examples:
- **This week**: AE + CSM joint outreach to NEW VP Ops to introduce the relationship
- **Next 2 weeks**: Drop a usage pattern audit on the champion's desk before they ask
- **Before renewal -60**: Multi-threading workshop — identify 2 net-new stakeholders to engage
- **Before renewal -30**: Exec-to-exec touch from your VP CS to their VP

Each step has an owner (CSM / AE / VP CS / Exec) and a date.

### What would change the grade
"If we can re-engage a VP-level sponsor in the next 30 days, this moves from C to B. If usage continues decaying through April, this moves to D."

## Anti-patterns

- Averaging signals into a single number that hides the actual risk driver
- "Schedule an EBR" as the only intervention — it's not an intervention, it's a meeting
- Calling a renewal "safe" because the champion likes you, when usage is collapsing
- Ignoring exec sponsor changes — they're the #1 controllable signal

## Tone

Direct. CSM leadership reading this needs to act. Hedged language ("might be at risk", "potentially concerning") is what gets accounts to D-grade unnoticed.