# Cs Operations > Save churning customers through proactive health scoring and recovery plays. Use when red accounts pile up unpredictably, renewal conversations turn contentious, or expansion-ready accounts slip away. Builds a 5-dimension health score (product usage, relationship quality, support friction, commercial trends, outcomes) updated weekly, triages red accounts into rescue mode or renewal-risk mode, and activates save plays with executive sponsorship, backed by the full post-sale operating model (onboarding, QBRs, CS-Sales handoffs, renewal governance). Produces a weekly red-account triage list and recovery playbooks by churn cause. Rule: a customer does not turn red on one signal; red means 3 or more health dimensions breaking at once. Trigger phrases: we keep losing customers, red accounts, health scoring, save play, churn autopsy, renewal price increase resistance. - Skill: `swan-gtm/cs-operations` (Agent Skill, multi-file: 8 files) - Install (CLI): `npx skillmds@latest add swan-gtm/cs-operations` - Raw SKILL.md: https://api.skillmd.com/api/skills/swan-gtm/cs-operations/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: AI & ML - Author: swan-gtm (https://skillmd.com/u/swan-gtm) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/swan-gtm/cs-operations --- # Customer Success Operations You are a CS operations architect. CS Ops is the post-sale revenue engine. It sits in the ICP Value Loops layer of the revenue operating system, owning Adopt → Realize Value → Renew → Expand. Without it, you're relying on individual heroics instead of systems. ## The CS Ops Operating Model Four pillars. Neglect any one and the system breaks. ``` PROCESS DESIGN TECHNOLOGY & SYSTEMS ├─ Playbooks ├─ CS Platform: Gainsight (Horizon AI, 2024+), ChurnZero, ├─ Workflows │ Vitally; all support real-time ML scoring (2026) ├─ Escalation paths ├─ Product analytics (Pendo, Amplitude) ├─ Governance ├─ Support (Zendesk, Intercom) ├─ Swimlanes ├─ Survey tools (NPS, CSAT; note: Delighted shut 30 Jun 2026) └─ AI agents └─ Data warehouse integration; reverse ETL (Hightouch, Census) DATA & ANALYTICS ENABLEMENT ├─ Health scores ├─ CSM onboarding (2 weeks, not 2 months) │ (ML-powered, real-time) ├─ Playbook documentation ├─ Churn prediction ├─ Tools training │ (ML+NLP ensemble) ├─ Certification on key processes ├─ Expansion scoring ├─ Knowledge base ├─ Cohort analysis └─ AI-assisted customer signal synthesis └─ Team efficiency ``` ## CS Maturity Model ``` ┌────────────┬──────────────┬────────────────┬───────────────┬──────────────┐ │ Stage │ Health Score │ Renewals │ Expansion │ Governance │ ├────────────┼──────────────┼────────────────┼───────────────┼──────────────┤ │ REACTIVE │ None. React │ Ad-hoc, │ Accidental │ CS isolated, │ │ │ to churn. │ customer-led │ │ no loops │ ├────────────┼──────────────┼────────────────┼───────────────┼──────────────┤ │ DEFINED │ Manual, CSM │ Playbook, 60d │ Signals ID'd │ Handoff doc │ │ │ judgment │ before expiry │ inconsistently│ exists │ ├────────────┼──────────────┼────────────────┼───────────────┼──────────────┤ │ PROACTIVE │ Composite, │ Automated 90d │ Scoring model │ Quarterly │ │ │ weekly auto │ process + save │ + targeted │ CS-Product │ │ │ alerts │ playbooks │ outreach │ forums │ ├────────────┼──────────────┼────────────────┼───────────────┼──────────────┤ │ PREDICTIVE │ ML model, │ Preventive │ Expansion │ Real-time │ │ │ >80% churn │ interventions, │ scoring feeds │ feedback │ │ │ accuracy │ <5% churn │ Sales pipeline│ loops │ └────────────┴──────────────┴────────────────┴───────────────┴──────────────┘ ``` If you're Reactive: stop everything else and move to Defined first. ## Health Score Framework The single most important CS Ops tool. Composite scores with weighted dimensions: ``` HEALTH SCORE = Weighted average of 5 dimensions PRODUCT USAGE (35%) Login frequency, feature adoption, depth of use, usage trend Weight rationale: usage is the strongest churn predictor RELATIONSHIP (25%) Executive sponsor engaged, NPS response, meeting cadence, org adoption % Weight rationale: champions save accounts SUPPORT (18%) Ticket volume trend, severity, resolution time, CSAT Weight rationale: pain signals churn COMMERCIAL (17%) Contract value trend, payment history, expansion conversations Weight rationale: money doesn't lie OUTCOMES (5%) Achieving stated goals, ROI realization, milestone completion Weight rationale: data often incomplete early on ``` ### Traffic Light System ``` GREEN (80+): Expand. Growth partner posture. 2-5% churn risk (practice-based). YELLOW (40-79): Intervene. Increase touchpoints. 15-30% churn risk (practice-based). RED (<40): Save mode. Executive intervention. 60%+ churn risk (practice-based). ``` ### Health Score Anti-Patterns ``` Binary scores ("healthy" vs "at risk") → Use 0-100 scale with thresholds Usage-only scores → Composite with multiple signals Manual, updated quarterly → Automated, real-time or weekly minimum Score says green but they churn → Calibrate against actual outcomes Nobody acts on red accounts → CS Manager reviews reds weekly ``` **EU Data Governance Note (GDPR Article 22):** Health scores incorporate personal data (NPS responses, engagement metrics from personal profiles). If automated scoring drives material customer-facing decisions (e.g., renewal discounting, reduced support tier) without human review, you must establish documented lawful basis and provide individuals with the right to human review per Article 22. Recommended: reserve automated action for internal escalation only; require CSM/manager review before any customer-visible decision. ## Segmentation & Coverage Models ``` ┌──────────────┬─────────────────┬─────────────────┬────────────────────┐ │ Segment │ Criteria │ Coverage Model │ CSM Ratio │ ├──────────────┼─────────────────┼─────────────────┼────────────────────┤ │ HIGH-TOUCH │ >$50K ARR │ Named CSM │ 1:10-50 │ │ (Strategic) │ Complex product │ Quarterly EBRs │ (ChurnZero, 2026) │ │ │ High expansion │ Dedicated onboard│ AI-assisted: +25-50%│ ├──────────────┼─────────────────┼─────────────────┼────────────────────┤ │ MID-TOUCH │ $10-50K ARR │ Pooled CSM │ 1:100-200 │ │ (Scaled) │ Moderate complex│ Templated touches│ (ChurnZero, 2026) │ │ │ Growth potential│ Triggered playbks│ AI-assisted tools │ ├──────────────┼─────────────────┼─────────────────┼────────────────────┤ │ TECH-TOUCH │ <$10K ARR │ Automated │ 1:500-1000 │ │ (Digital-led)│ Self-serve │ In-app guidance │ (ChurnZero, 2026) │ │ │ Simple use case │ Community/webinar│ AI agents primary │ └──────────────┴─────────────────┴─────────────────┴────────────────────┘ ``` Reassess tiers quarterly based on usage growth, health, and expansion potential. ## Renewal Management Start 120 days before contract end. Not 30. Starting at T-30 is too late for anything except discounting. This is the single canonical renewal timeline. ``` T-120 days: PREP Renewal flag auto-created in CRM CSM reviews health, usage, and expansion signals; Finance pulls contract terms; Product flags blockers If health < 40 (Red): escalate immediately, don't wait for T-90 Output: Renewal strategy (standard / at-risk save play / expansion) T-90 days: HEALTH CHECK / KICKOFF QBR conducted, executive alignment, pulse on value delivery CSM prepares ROI / value-delivered summary Internal alignment: any pricing changes, product updates? If discount expiring: begin value-led conversation (see Renewal Discount Governance) Score: Likely / At-risk / Needs work T-60 days: PROPOSAL / CONVERSATION Standard: Renewal proposal sent. Present value delivered, roadmap alignment, terms At-risk: Executive sponsor conversation, barrier removal; activate save playbook Expanding: Sales/CSM joint upsell conversation, present expansion proposal alongside renewal T-30 days: NEGOTIATION Contract sent; handle objections, escalate to CS leader if not signed If discount negotiation: use graduated step-up (see Renewal Discount Governance) Daily tracking of renewal pipeline T-0: CLOSE Contract signed or churned. Revenue booked, post-renewal thank you, reset for next cycle If churned: initiate churn autopsy within 7 days Feed learnings to Product T+7: POST-RENEWAL CHECK-IN Confirm satisfaction with terms, document win/loss reason Set next QBR date, update health score ``` **Track:** On-time renewal rate, early renewal rate, save rate (red accounts renewed), renewal velocity. ## Expansion Playbooks Expansion is a process, not an event. Most expansion dies because nobody owns the handoff from CS to Sales or because the signal was visible but nobody acted on it. ### Expansion Signals ``` USAGE: Hitting seat limits, new feature adoption, DAU increasing, usage approaching plan limits (>80% utilization) RELATIONSHIP: Champion promoted, new stakeholder/exec sponsor engaged, positive NPS, new departments/teams using the product COMMERCIAL: Org headcount growth, new budget cycle, cross-functional interest, feature requests for premium-tier functionality OUTCOMES: Value achieved ahead of schedule, ROI exceeded, new use cases, positive health trend (score increasing 3+ months), referral/case study ``` ### Expansion Ownership Model ``` Small expansion (<20% ACV): CS owns end-to-end Medium expansion (20-50%): CS identifies, Sales closes Large expansion (>50% ACV): Sales owns, CS supports ``` CS-to-Sales handoff: CS prepares customer context, decision-maker, timeline, needs. Sales accepts or declines. If accepted, CS introduces Sales to champion and stays involved through close. Expansion motion: CS documents the signal in CRM → warm intro from CSM to AE/expansion AE → AE runs the commercial conversation while CSM provides context and stays on early calls → post-expansion, CS implements the expanded scope and tracks value on the new commitment. For the 0-100 expansion scoring model (usage/stakeholder/feature/health/headroom weights), score-based ownership thresholds, account readiness triggers, and whitespace analysis, see `references/expansion-scoring-model.md`. ## Onboarding Orchestration The handoff is where retention is won or lost. A weak handoff creates context loss, broken promises, and a customer who feels like they're starting over. For the step-by-step Sales → CS handoff process and checklist, see the summary under "Sales → CS Handoff" below and `references/sales-cs-handoff-template.md`. This is the single canonical onboarding model. ``` DAY 1-7: ACTIVATION / KICKOFF Welcome email (within 2 hours), license activation, kickoff scheduled Kickoff call with key stakeholders; assign customer-side project owner Success criteria and timeline confirmed: "What does success look like?" (Use SPICED Impact as baseline) DAY 8-30: IMPLEMENTATION / TECHNICAL SETUP Configuration / environment setup, data migration, integration setup Admin training, user training, weekly status calls, blocker resolution Document known issues or customization needs Milestone: Go-live readiness confirmed DAY 31-60: STABILIZATION / USER ENABLEMENT Go-live, first-week support surge (daily check-ins) End-user training sessions, internal documentation/guides Identify power users / internal champions, validate adoption metrics Usage monitoring, issue resolution, user adoption tracking First value delivered: "Here's the impact you're seeing" DAY 61-90: OPTIMIZATION / FIRST VALUE Confirm first value milestone achieved; capture success story (even informal) Performance vs. success criteria, advanced feature rollout Executive check-in, health score baseline set, first QBR/check-in scheduled Transition to business-as-usual CSM cadence ``` **Automation triggers:** No kickoff by day 3 → reminder. Not go-live by day 30 → alert manager. Usage <20% by day 60 → intervention. Health <50 by day 90 → yellow playbook. **Track:** Time to kickoff (<3d), time to first login (<7d), time to first value (TTFV) (<30d), go-live on schedule (90%+), onboarding completion rate (>90%), onboarding CSAT (90%+ top performers; recommended), feature adoption at T+30 (>3 core features). ### Sales → CS Handoff When a deal hits Closed Won, an automated notification routes the account to the assigned CSM with a CRM record containing the SPICED summary, stakeholders, contractual commitments, known risks, product/tier config, and any discount terms with expiry dates. The CSM reviews it, runs a short internal handoff call with the AE (what was promised, who's the champion/blocker, what success looks like, landmines), then sends the welcome email and schedules kickoff. A handoff quality check confirms the SPICED summary is complete, all stakeholders are in CRM, and discount terms are documented with expiry. For the full step-by-step timeline, Sales pre-handoff responsibilities, the three-way handoff meeting, and checklist, see `references/sales-cs-handoff-template.md`. ## Event-Based Playbooks ``` USAGE DROP (WAU drops >30% week-over-week) Action: CSM outreach within 24 hours Root cause: Change, vacations, budget freeze, support issue? NPS DETRACTOR (Score <6) Action: Call within 48 hours Document specific complaint, not generic dissatisfaction CHAMPION DEPARTURE Action: Identify new stakeholder within 1 week Risk mitigation: Involve next-level executive immediately SUPPORT ESCALATION (Sev-1 OR 3+ tickets in 7 days) Action: Triage within 2 hours, CS + Support Manager ``` ## Renewal Discount Governance The #1 cause of contentious renewals is a discount that was never documented to expire. At renewal, the customer anchors on what they paid, not what the product is worth. This is a structural governance failure: every discounted deal must carry a `Discount_Expiry_Date` in CRM, the T-120 workflow must flag expiring discounts automatically, and the CSM must prepare value justification before the conversation. The conversation framework leads with value (T-90), frames the step-up (T-60), and offers a graduated step-up only as a save play (T-30). GPO/consortium accounts get coordinated tier/volume validation from T-120 through T-30 with Deal Desk. For the full discount sunset problem, the pricing conversation framework, the save-plays-with-pricing table, and GPO renewal coordination, see `references/renewal-discount-governance.md`. For discount governance policy (approval matrices, ACV tiers), see pricing-monetization-ops skill. ## CS-Product Feedback Loop Install the structural mechanisms that turn customer signal into product action: a quarterly advisory board (3-5 customer execs + Product VP), churn autopsies on 100% of churned accounts (CS Manager + Product Manager), feature request tagging with ARR/use case/competitive/expansion/blocker fields, and a monthly adoption blocker log feeding a dedicated Product queue. For the full feedback loop architecture (owners, questions, outputs, and review cadence for each mechanism), see `references/cs-product-feedback-architecture.md`. ## Key Metrics ``` REVENUE TARGETS GRR 84% median (75th %ile 91%+; Optifai, 2026) NRR 106% median (top performers >130%; Optifai, 2026) Logo retention rate 80-90% (aspirational; industry median 70-80%) Expansion rate (% ARR growth) 40% of new ARR from expansion (ZoomInfo, 2026) OPERATIONAL Time to first value <30 days (recommended) Health score distribution >70% green (recommended) QBR completion (high-touch) 100% (recommended) Renewal on-time rate >90% (recommended) Save rate (red accounts) >40% (practice-based) EFFICIENCY Accounts per CSM (by tier) See coverage model (sourced, 2026) Admin time (% of total) <30% (recommended) CSM attrition rate <15%/year (recommended) LEADING INDICATORS Accounts with declining usage Track weekly (recommended) Accounts without exec sponsor <5% (recommended) Days to complete renewal <60 days (recommended) Core feature adoption rate >70% (recommended) ``` ## AI and Automation in CS Ops (2026) Real-time health scoring, autonomous churn prediction, and AI-assisted signal synthesis are now standard in modern CS platforms. This is not "future state"; it's operational today. ### Real-Time Predictive Scoring Modern platforms (Gainsight Horizon AI, ChurnZero, Vitally) update health scores in real time rather than batch weekly. Triggers are now event-driven: usage cliff detected → CSM alert; NPS detractor submitted → auto-flag; champion departure → relationship score drops immediately. Churn prediction with ensemble ML plus NLP on unstructured signals (call transcripts, support tickets, email sentiment) cuts churn 15-30% within 12 months (2026 vendor research). ### Autonomous Agents and Workflow Renewal and expansion processes no longer require manual CSM review at every step. Autonomous agents monitor thresholds (usage >80% of plan, score drop >15pts), initiate escalation workflows, and in some cases send templated outreach. The CSM remains accountable but the agent handles routine hygiene: flag at-risk accounts, surface expansion candidates, suggest playbook actions. This frees CSMs to focus on high-touch strategy and executive relationships. ### Dynamic Segmentation Static ARR tiers are outdated. Modern platforms segment dynamically by usage trajectory, health trends, and expansion readiness. A sub-$10K account with 200% month-over-month growth and executive sponsor engagement moves into a higher-touch coverage tier automatically. Weekly re-segmentation based on behavioral signals means coverage shifts to where the opportunity is. ### AI-Assisted Churn Autopsy and Insight Synthesis Gainsight Horizon AI (2024+) and peers now offer automated call transcription, sentiment analysis, and account summary synthesis. Rather than manual CSM + PM churn autopsies, the system ingests call recordings, extracts decision-driver signals, compares against similar accounts, and surfaces likely product gaps. QBR prep no longer requires manual data gathering; AI-generated benchmarking insights ("You are 40% below similar customers in feature X") are delivered automatically to both CSM and customer. **Implementation rule:** Deploy gradually. Start with real-time scoring; add autonomous escalation after 2 weeks; enable dynamic segmentation once you have clean usage data feeding the platform. Full AI adoption (transcription, synthesis, autonomous agents) requires data quality baseline of 80%+ field population. ## Diagnostic Assessment Score your CS Ops maturity. Count YES answers: ``` PROCESS: Documented playbooks? Defined renewal timeline? Health scoring framework? Escalation paths? QBRs for all high-touch? TECHNOLOGY: Dedicated CS platform? Integrates with product analytics? Automated health scoring? Usage data visible to CSMs? DATA: Can forecast churn 30+ days out? Health dashboards? Track expansion signals? Segment by tier? Measure team efficiency? ENABLEMENT: Structured CSM onboarding? Accessible playbooks? Process certification? Monthly training? GOVERNANCE: CS-Product feedback loop? CS-Sales handoff process? Exec reviews CS metrics monthly? Churn autopsies? Expansion ownership clear? Discount sunset tracking in place? 0-6 YES = Reactive. Fix: Define playbooks, build basic health scoring 7-13 YES = Defined. Fix: Integrate tech, automate playbooks 14-19 YES = Proactive. Fix: Build predictive models, optimize coverage 20-25 YES = Predictive. Fix: AI-assisted coaching, autonomous workflows ``` ## Process Operating Cadence Health scoring and renewal frameworks answer "who is at risk?" and "how do we retain them?" They depend on a structured review cadence and escalation discipline. A health score without a review cadence is a dashboard no one acts on. The cadence is the operating system; the score is the signal. **Manual review cadence (foundation for all tiers):** - **Weekly:** CSM reviews accounts with a score drop >10pts; flag any Green→Yellow or Yellow→Red (CSM + CS Lead). - **Monthly:** Full health score review of all accounts; any Red account reviewed with CS Lead + AE; scores updated on new data. - **Quarterly:** Portfolio health review, QBR prep for T1 accounts, Red account recovery planning. **Event-driven automation (2026+, layered on top):** Modern CS platforms trigger actions in real time on usage cliffs (WAU drop >30%), NPS detractors, champion departures, and score transitions. Autonomous agents notify CSMs of threshold breaches and recommend escalation. This does not replace the manual cadence; it augments it. The CSM still owns accountability and strategy, but routine monitoring is automated. **Red accounts: escalate within 24h.** When a customer turns Red, the CSM (or automated alert) notifies the CS Lead, who reviews the health breakdown and jointly decides with the AE between rescue mode (>90 days to renewal) and renewal risk mode (<90 days; activate the renewal process immediately, don't wait for T-90). For the full Red Account Playbook (rescue mode, renewal risk mode, commercial options) and the complete health monitoring cadence table, see `references/red-account-playbook.md`. QBRs are the highest-leverage CS touchpoint for T1 accounts and most fail by looking backwards rather than forwards. For the 45-60 minute agenda structure and cadence-by-tier guidance, see `references/qbr-process-template.md`. --- ## How to Use This Skill **"We keep losing customers but don't know why":** Start with health scoring. Build the 5-dimension composite score. You'll see patterns within 4 weeks. **"Our CS team is drowning":** Segmentation problem. Build the 3-tier coverage model. Most teams try to high-touch everyone. That doesn't scale past 50 accounts. **"Expansion is left on the table":** Build expansion signals and scoring. Define the CS-Sales handoff. Most expansion dies because nobody owns the handoff. **"Product doesn't listen to CS":** Install the feedback loop: churn autopsies, feature request tagging with ARR impact, quarterly advisory board. Product responds to revenue data. **"We need to build CS Ops from scratch":** Run the diagnostic. Score your maturity. Fix the weakest pillar first. Month 1: playbooks + basic health scoring. Month 2: renewal process. Month 3: automation. **"Discounts are expiring and customers are pushing back":** Install the renewal discount governance framework. Document every discount with an expiry date. Lead with value at T-90, frame the step-up at T-60, and have graduated step-up as a save play only at T-30. For discount governance policy (approval matrices, ACV tiers), see pricing-monetization-ops skill. **"GPO/consortium accounts coming up for renewal":** Check GPO master record. Validate actual vs. committed volume. Coordinate with Deal Desk. No ad-hoc terms outside the GPO framework. --- ## Reference Files | File | When to read | What's inside | |------|-------------|---------------| | `references/customer-lifecycle-plays.md` | Choosing the right CS action for an account's phase | Four-phase classification, health interpretation per phase, trigger plays, expansion timing rule | | `references/renewal-discount-governance.md` | Discount expiring / contentious renewal pricing | Discount sunset problem, pricing conversation framework, save-plays-with-pricing table, GPO renewal coordination | | `references/expansion-scoring-model.md` | Quantifying/prioritizing expansion | 0-100 scoring model, score-based ownership thresholds, readiness triggers, whitespace analysis | | `references/sales-cs-handoff-template.md` | Designing the Closed-Won → CS handoff | Step-by-step T+0 to T+5 timeline, Sales pre-handoff duties, handoff meeting, quality checklist | | `references/red-account-playbook.md` | A customer turns Red | Rescue mode, renewal risk mode, commercial options, full health monitoring cadence table | | `references/qbr-process-template.md` | Designing or running a QBR | 45-60 min agenda structure, cadence by tier | | `references/cs-product-feedback-architecture.md` | Building the CS-Product loop | Advisory board, churn autopsy, feature request tagging, adoption blocker log | --- ## Operator Templates: Forecasting Worksheet (Renewals Tab) Use a forecasting worksheet with dedicated renewal modeling tabs to model: renewal cohort sizing, churn impact on ARR, expansion uplift scenarios. ## What good looks like - Every account carries a health score across five dimensions, refreshed weekly. - Red accounts surface on a triage list with an owner and a named recovery play before the renewal window opens. - Renewal conversations start from documented value evidence, not a pricing surprise. - Churn postmortems feed a playbook per churn cause instead of anecdotes. > Built by [Neon Triforce](https://neontriforce.com)