Engagement Tracker

Tracks touchpoint frequency and recency across all communication channels for every stakeholder. Computes engagement health per contact and per account. Part of the Relationship Architecture pillar.

This is a measurement and alerting skill. It tracks frequency -- how often and how recently have you engaged. It does not assess quality -- a weekly call where nothing meaningful is discussed scores the same as a weekly call that transforms the relationship. Frequency is necessary for healthy relationships; it is not sufficient. The CSM provides the quality assessment.

When to Run

• Continuous: Logs touchpoints as they occur from connected channels
• Scheduled: Weekly engagement health computation and gap detection
• On-demand: When a CSM requests engagement data for a specific account or contact

Core Execution Logic

Step 1: Log Touchpoints from All Channels

Every customer interaction is a touchpoint. Capture from all available sources:

For each touchpoint, log:

• Timestamp
• Channel
• Direction: inbound (customer-initiated) or outbound (CSM-initiated)
• Contact(s) involved
• Type: scheduled (planned touchpoint) or unscheduled (ad hoc)

Step 2: Compute Per-Contact Engagement Metrics

For each contact in ra-stakeholder-mapper:

Step 3: Classify Engagement Health

For each contact, compare engagement metrics against the expected cadence for their stakeholder type (from ra-stakeholder-mapper):

Step 4: Compute Per-Account Engagement Health

Aggregate contact-level health into an account-level engagement score:

Weighted by stakeholder type:

• Champion engagement: 35% weight (the most important relationship)
• Economic Buyer engagement: 25% weight
• Executive Sponsor engagement: 15% weight
• Technical Lead engagement: 15% weight
• All other contacts: 10% weight

Account engagement score = weighted sum of contact engagement scores (each contact scored 0-100 based on health level: Strong=90, Adequate=70, Declining=45, Gap=20, Lost=0)

Account engagement classification:

Step 5: Detect Patterns and Generate Alerts

Step 6: Generate Engagement Report

Per-CSM weekly engagement summary:

{
  "csm": "Jane Doe",
  "report_date": "2026-03-10",
  "portfolio_engagement": {
    "total_accounts": 42,
    "engagement_distribution": {
      "healthy": 28,
      "adequate": 8,
      "declining": 4,
      "critical": 2
    },
    "avg_engagement_score": 72
  },
  "alerts_this_week": [
    {
      "type": "no_reply_streak",
      "account": "Gamma Corp",
      "contact": "Lisa Park (VP Eng)",
      "detail": "3 outbound emails in 14 days with no response",
      "recommended_action": "Try a different channel (call) or approach through another contact"
    },
    {
      "type": "champion_cooling",
      "account": "Beta Inc",
      "contact": "Tom Chen (Champion)",
      "detail": "Engagement moved from Strong to Declining. Touchpoint frequency down from every 10 days to every 28 days over last 60 days",
      "recommended_action": "Investigate. Schedule a direct conversation. Do not add more emails to the no-reply streak"
    }
  ],
  "engagement_movers": {
    "improved": [
      { "account": "Acme Corp", "score_change": "+12", "driver": "Executive sponsor engaged for first time in 90 days" }
    ],
    "declined": [
      { "account": "Beta Inc", "score_change": "-15", "driver": "Champion engagement cooling" }
    ]
  },
  "cadence_compliance": {
    "touchpoints_expected": 38,
    "touchpoints_completed": 34,
    "compliance_rate": 0.89,
    "overdue_touchpoints": [
      { "account": "Delta Ltd", "contact": "Sarah Kim", "expected_cadence": "monthly", "days_overdue": 12 }
    ]
  }
}

Handoff to Human

Confidence and Limitations

• High confidence for touchpoint logging and frequency computation -- structured event data from connected systems
• High confidence for engagement health classification -- deterministic comparison of metrics to defined thresholds
• Medium confidence for pattern detection (no-reply streaks, cancellation patterns) -- these are reliable signals but have multiple possible interpretations
• Medium confidence for inbound ratio as an engagement quality proxy -- a declining inbound ratio is a meaningful signal, but some customers are naturally passive communicators. Benchmark against the contact's own historical pattern, not a universal standard
• Low confidence for engagement quality assessment. This is the fundamental limitation: the skill tracks frequency, not depth. A 15-minute status call and a 2-hour strategic planning session count the same. The CSM provides the quality layer
• Cannot track engagement that happens outside connected systems (in-person meetings, conferences, informal conversations, personal phone calls)
• "No inbound engagement" is a strong disengagement signal in most cases, but some customers only engage reactively and have always done so. The baseline matters more than the absolute number

Dependencies

Required:

• CRM API (activity data, contact records)
• Calendar integration (meeting data)
• Email integration (email activity data)

Strongly recommended:

• ra-stakeholder-mapper (stakeholder types and expected cadence for health computation)
• Call platform integration (call data)
• Support platform (support interaction data)
• ca-community-monitor (community engagement channel)

Downstream consumers:

• bi-health-score (engagement data feeds the engagement component -- 25% default weight)
• ra-stakeholder-mapper (engagement status per contact)
• bi-risk-detector (declining engagement as a risk signal)
• lo-check-in-scheduler (cadence compliance data)
• cc-report-generator (engagement metrics for CSM and manager reports)

References

• references/engagement-methodology.md -- Channel definitions, health classification thresholds, and cadence baseline calibration