Retention Risk Command Center
A churn-risk operating console that scores accounts, identifies risk drivers, assigns playbooks, and models expected saved revenue and payback.
The problem
Retention work often starts after the account is already in trouble. Usage, support, payment, renewal, and relationship signals sit in different places, so intervention work becomes reactive and hard to measure.
The thesis
Retention improves when risk scoring, driver diagnosis, playbook selection, intervention ownership, and save-rate economics live in the same operating loop.
System flow
Step 1
Account risk inputs
Maintain editable MRR, usage, support, NPS, renewal, payment, sponsor, touch, and trend signals.
Step 2
Risk scoring layer
Classify account risk and identify the primary churn driver behind each recommendation.
Step 3
Playbook assignment
Map risk drivers to intervention playbooks with save-rate lift, cost, discount, and SLA assumptions.
Step 4
Save economics
Persist expected saved revenue, intervention cost, payback ratio, and portfolio recommendation.
Case-study artifacts
Risk scoring
Account signals combine into interpretable churn-risk scores.
Saved revenue
Runs estimate preventable churn and expected saved revenue.
Interventions
High-risk accounts can be routed to owner-assigned playbooks.
Decision frame 1
Input
Editable account health, MRR, renewal, payment, sponsor, support, NPS, playbook, and policy data.
Decision frame 2
Decision
Score churn risk, identify the primary driver, and select the strongest playbook.
Decision frame 3
Output
Risk run, account recommendations, expected saved revenue, payback, and intervention queue.
Evidence links
Shows the command-center framing and operating sequence.
Shows editable accounts, playbooks, and policy thresholds.
Shows model controls, risk/save visualization, and driver mix.
Shows owner-assigned playbooks and intervention workflow.
Shows expected saved revenue, payback, and account recommendations.
Guardrails
High-risk accounts require owner and response window assignment.
Discount/save offers are constrained by payback and margin rules.
Closed interventions remain monitored for repeated risk signals.
System architecture
1. Account risk inputs
Maintain editable MRR, usage, support, NPS, renewal, payment, sponsor, touch, and trend signals.
2. Risk scoring layer
Classify account risk and identify the primary churn driver behind each recommendation.
3. Playbook assignment
Map risk drivers to intervention playbooks with save-rate lift, cost, discount, and SLA assumptions.
4. Save economics
Persist expected saved revenue, intervention cost, payback ratio, and portfolio recommendation.
KPI callouts
Risk band
Every account receives an interpretable risk score and driver.
Expected saved
Portfolio runs model preventable churn and saved revenue.
Payback
Playbooks are evaluated against intervention cost and minimum payback policy.
Commercial framing
The system turns retention from reactive triage into economic prioritization: which account is at risk, why, what action should happen, and whether the save motion is worth the cost.
What I built
I built the retention schema, editable account/playbook/policy inputs, risk simulation endpoint, run visualization, intervention queue, output economics, audit trail, and docs.