Executive Summary

The Executive Summary is a four-card KPI panel that displays the most critical factory health indicators at a glance. It appears at the top of the Factory dashboard and provides an instant snapshot of your SDLC pipeline performance.

The Four KPIs

Total Issues

The count of work items (e.g., Linear issues, GitHub issues) that were completed during the selected time window.

Why it matters: Throughput is the primary measure of factory productivity. A healthy factory sustains consistent throughput week-over-week.

Interpretation:

• Increasing → team is shipping more; can indicate improved automation or scaling
• Decreasing → potential bottleneck; check station metrics for blockages
• Zero → no work completed yet; dashboard may need longer time window or data hasn't ingested

Source: Counted from factory_events with type work-item.completed.

Total Cost

The sum of all costs (in USD) incurred to complete the work items in the time window. Cost includes:

• LLM provider costs (Claude, GPT, Gemini, etc.)
• Compute costs (sandbox execution, e2b, Vercel, etc.)
• Estimated human time (based on handoff events and human review durations)

Why it matters: Cost per issue is the "price of productivity." Optimizing the ratio of cost-to-throughput drives business efficiency.

Interpretation:

• $X.XX for N issues → cost per issue = $X.XX ÷ N
• Rising cost, flat throughput → agents are using more expensive models or loops; check provider routing
• Stable cost, rising throughput → automation improving; agents are getting better

Format: USD formatted to 2-3 decimal places (e.g., $1.23, $0.045).

Source: Sum of factory_metrics rows with type cost_per_issue.

Avg Cycle Time

The median time elapsed from when an issue entered the workflow (research station) to when it was deployed (deploy station completed).

Why it matters: Cycle time is a measure of agility. Shorter cycle times mean faster feedback loops and faster time-to-value.

Interpretation:

• Low (<24h) → healthy velocity; agents are moving work through quickly
• High (>3d) → potential bottleneck; check which stations are slow via the Station Pipeline panel
• Zero or "--" → no completed items with time data; data may be incomplete

Format: Human-readable duration (e.g., 42m, 3h, 2.5d).

Source: Calculated from work-item.started and work-item.completed events.

Avg Yield

The percentage of work items that passed first time without requiring rework (dev → QA → acceptance → deploy with no rework loop back to dev).

Why it matters: Yield is a quality metric. Higher yield means fewer defects, faster throughput, and lower cost per issue.

Interpretation:

• ≥90% → excellent quality; agents are producing production-ready code consistently
• 70-89% → acceptable; some rework happening but under control
• <70% → quality alert; investigate via rework panel; may indicate SDLC template issue or agent skill gap
• 0% or "--" → insufficient data or no rework tracking enabled

Format: Percentage (e.g., 87.3%).

Source: Calculated from ratio of direct completions vs. rework events.

Reading the Summary

The four KPI cards should be read left to right, top to bottom:

┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐
│ Total Issues    │  │ Total Cost      │  │ Avg Cycle Time  │  │ Avg Yield       │
│ 47              │  │ $256.82         │  │ 14h             │  │ 92.1%           │
└─────────────────┘  └─────────────────┘  └─────────────────┘  └─────────────────┘

A healthy dashboard snapshot looks like:

• Throughput: 5-10 issues/day (depending on issue size)
• Cost: $1-$5 per issue (varies by complexity and provider)
• Cycle time: <24h (for most SDLC templates)
• Yield: >85% (showing mostly first-pass completions)

Time Windows

The Executive Summary aggregates data over your selected time window (7d, 30d, or 90d). Click the time-range selector to update:

Data Freshness

Executive Summary metrics are updated by the factory aggregation job, which runs hourly. Raw events (work-item.completed, etc.) are ingested in real-time, but aggregated KPIs have a ~1-hour lag.

API Access

Fetch the Executive Summary metrics programmatically:

curl -H "Authorization: Bearer $RENSEI_API_KEY" \
  "https://app.rensei.ai/api/factory/metrics?metricType=aggregate&timeRange=30d"

Response:

{
  "metrics": {
    "totalIssues": 47,
    "totalCost": 256.82,
    "avgCycleTimeMs": 50400000,
    "avgYield": 92.1,
    "throughputPerDay": 1.57,
    "costTrend": [12.5, 13.2, 11.8, ...]
  }
}

See Metrics API for full details.

Next Steps

• For per-station breakdown, see Station Pipeline
• For cost attribution by stage, see Cost Breakdown
• To configure cost caps and provider routing, see Agent Cards