Reporting and Analytics¶
This document defines key KPIs and analytics for the Personal Agents Platform domain. It is written for product managers, business analysts, and engineers building reporting and analytics capabilities.
Reporting and analytics provide visibility into platform performance, user engagement, agent effectiveness, and business metrics. This document defines KPIs, data sources, and dashboard requirements.
Note
Analytics are built on events from domain services and Audit Platform. Real-time dashboards provide immediate visibility, while historical reports enable trend analysis and cost optimization.
Key KPIs¶
| KPI | Description | Calculation | Target |
|---|---|---|---|
| Active Users | Monthly active users | Count of unique users with at least one agent run in period | Track growth |
| Agent Runs per User | Average agent runs per user per month | Total runs / Active users | 10-20 runs/month |
| Suggestion Approval Rate | Percentage of suggestions approved | (Approved / Total) × 100 | > 60% |
| Pack Adoption Rate | Percentage of users with at least one pack | (Users with packs / Total users) × 100 | > 70% |
| Agent Run Success Rate | Percentage of successful agent runs | (Successful / Total) × 100 | > 95% |
| Average Run Duration | Average time for agent execution | Average(CompletedAt - StartedAt) | < 30 seconds |
| Token Efficiency | Average tokens per run | Average(TokensUsed per run) | Optimize based on cost |
| Cost per User | Average AI costs per user per month | Sum(Cost per run) / Active users | Track and optimize |
| Free-to-Pro Conversion | Percentage of free users upgrading to Pro | (Upgrades / Free users) × 100 | > 5% |
| User Retention (30-day) | Percentage of users active after 30 days | (Active after 30 days / New users) × 100 | > 40% |
Data Sources and Events¶
Domain Events¶
Events from Domain Services:
| Event | Source | Data Captured |
|---|---|---|
UserCreated |
Personal Identity & Profile Service | User ID, email, subscription tier, created date |
AgentRunCreated |
Agent Orchestration Service | Run ID, user ID, agent ID, trigger type |
AgentRunStarted |
Agent Orchestration Service | Run ID, started timestamp |
AgentRunCompleted |
Agent Orchestration Service | Run ID, completed timestamp, tokens used, cost, status |
AgentRunFailed |
Agent Orchestration Service | Run ID, failed timestamp, error details |
SuggestionCreated |
Suggestion & Approval Service | Suggestion ID, user ID, run ID, type, priority |
SuggestionApproved |
Suggestion & Approval Service | Suggestion ID, approval timestamp, decision |
SuggestionRejected |
Suggestion & Approval Service | Suggestion ID, rejection timestamp |
PackInstalled |
Pack & Workflow Service | Pack ID, user ID, installation timestamp |
ConnectorAuthorized |
Connector Service | Connector ID, user ID, connector type, authorization timestamp |
SubscriptionUpgraded |
Personal Identity & Profile Service | User ID, from tier, to tier, upgrade timestamp |
Audit Platform Events¶
Events from Audit Platform:
| Event Type | Description | Use Case |
|---|---|---|
| User Actions | User login, pack installation, suggestion approval | User behavior analysis |
| Agent Run Events | Agent execution, token usage, cost tracking | Agent performance and cost analysis |
| System Events | Pack installation, connector authorization | System activity tracking |
| Privacy Events | Data access, data export, account deletion | Compliance and privacy reporting |
See: Audit Platform for audit platform details.
Data Pipeline¶
ETL Flow:
- Domain services emit events to event bus (Azure Service Bus)
- Events consumed by Dashboard & Analytics Service
- Events transformed and stored in data warehouse (Azure SQL Database)
- Dashboards query data warehouse for metrics
- Real-time metrics calculated from event stream
See: Event-Driven Mindset for event patterns.
Dashboards and Views¶
User Dashboard¶
Purpose: User-facing dashboard showing their agents, packs, runs, and suggestions
Metrics:
- Active packs count
- Active agents count
- Recent agent runs (last 7 days)
- Pending suggestions count
- Usage this month (runs, tokens, cost)
- Usage vs. subscription limits
Widgets:
- Packs Overview - List of installed packs with status and last run
- Agents Overview - List of agent instances with status and last run
- Recent Runs - Timeline of recent agent runs with status and results
- Suggestions Inbox - Pending suggestions requiring approval
- Usage Summary - Runs/month, tokens used, cost, vs. limits
- Run History - Detailed run history with filters
Views:
- Dashboard home (overview)
- Packs detail view
- Agents detail view
- Runs detail view
- Suggestions inbox
- Usage and billing
Operations Dashboard¶
Purpose: Internal operations dashboard for platform management
Metrics:
- Total active users
- Total agent runs (today, week, month)
- Total tokens used and costs
- Suggestion approval rates
- Agent run success rates
- Subscription tier distribution
- User retention metrics
Views:
- Platform health overview
- User engagement metrics
- Agent performance metrics
- Cost and usage trends
- Subscription and revenue metrics
- Error and failure analysis
Agent Performance Dashboard¶
Purpose: Dashboard for analyzing agent performance and effectiveness
Metrics:
- Agent run success rate by agent type
- Average tokens per run by agent type
- Average cost per run by agent type
- Suggestion generation rate by agent type
- Suggestion approval rate by agent type
- Most used agents and packs
Views:
- Agent performance comparison
- Cost analysis by agent type
- Suggestion effectiveness
- Popular agents and packs
- Performance trends over time
Cost Analytics Dashboard¶
Purpose: Dashboard for cost tracking and optimization
Metrics:
- Total AI costs (daily, weekly, monthly)
- Cost per user by subscription tier
- Cost per agent run
- Token usage trends
- Cost by agent type
- Cost optimization opportunities
Views:
- Cost overview and trends
- Cost breakdown by user tier
- Cost breakdown by agent type
- Token usage analysis
- Cost optimization recommendations
Analytics Queries¶
User Engagement Analytics¶
Query Examples:
- Monthly active users (MAU)
- Daily active users (DAU)
- User retention cohorts
- Agent runs per user distribution
- Pack adoption by user segment
- Suggestion interaction patterns
Agent Performance Analytics¶
Query Examples:
- Agent run success rate by agent type
- Average run duration by agent type
- Token usage distribution
- Cost per run by agent type
- Suggestion generation and approval rates
- Most effective agents (by approval rate)
Business Analytics¶
Query Examples:
- Free-to-Pro conversion funnel
- Subscription tier distribution
- Revenue per user (RPU)
- Customer lifetime value (LTV)
- Churn analysis
- User acquisition sources
Cost Analytics¶
Query Examples:
- Total AI costs over time
- Cost per user by tier
- Cost per run trends
- Token usage efficiency
- Cost optimization opportunities
- ROI by subscription tier
Reporting Requirements¶
User Reports¶
Available Reports:
- Usage Report - Monthly usage summary (runs, tokens, cost)
- Run History Report - Detailed run history with filters
- Suggestion Report - Suggestion activity and approval rates
- Cost Report - Cost breakdown and trends
Operations Reports¶
Available Reports:
- Platform Health Report - System performance and reliability
- User Engagement Report - User activity and retention
- Agent Performance Report - Agent effectiveness and costs
- Business Metrics Report - Subscription, revenue, conversion metrics
Compliance Reports¶
Available Reports:
- Audit Trail Report - Complete audit log for compliance
- Data Access Report - User data access logs
- Privacy Report - Data retention and deletion logs
Related Documents¶
- Domain Overview - Domain overview
- Domain Vision and Scope - Vision and scope
- Processes and Flows - Business processes
- API and Integration View - API design
- High-Level Design - System architecture
- Dashboard & Analytics Service - Service details
- Audit Platform - Audit platform
- Event-Driven Mindset - Event patterns
- connectsoft.me Business Model - Business model