Factory Economics and Operating Metrics
Agent activity is easy to measure and easy to mistake for value. Tokens, sessions, generated lines, tool calls, and pull request volume can all rise while customer outcomes slow, defects increase, and engineers spend more time recovering or
A focused view of boundaries, contracts, state, authority, failure paths, and tradeoffs drawn from this chapter.
Reconstruct and defend this chapter’s architecture.
Reconstruct the architecture, name each boundary, and defend the tradeoffs.
Open the source exercise
Draw a metric tree from customer outcome down to delivery, factory, and runtime diagnostics. Add a team that doubles PR throughput while review time and change failure rise. Explain why the factory has not improved and which bottleneck to address.
4. Tradeoffs and alternatives
Comprehensive measurement can become surveillance. Measure workflows and systems, not individual developer productivity. Lines of code, commits, hours online, and prompt volume are especially harmful as performance targets.
Business outcomes can take weeks to observe, while teams need fast feedback. Use leading indicators, but label them as proxies. Do not quietly substitute “PR merged” for customer value.
Cost attribution will be imperfect. A transparent range is better than a precise but incomplete number. The measurement system itself has operational cost and should remain proportional to the decisions it supports.
5. Current Mission Control Implementation
At commit
af414acfaa7ea793cb43de8ab2617f343d922f23,
Mission Control implements the metric-lineage foundation and operator reporting
surfaces. This is current product architecture, not a hypothetical future
capability.
Convex retains canonical Task events plus Mission, Plan, WorkOrder, Attempt, candidate, evidence, quality-gate, pull-request, audit, run, workflow, and cost records. Its effectiveness projection reports verified completion, autonomous completion, and cost per verified outcome while labeling provenance as observed, projected, or insufficient. The operator product exposes portfolio, Factory Health, analytics, trace inspection, WorkOrder, and evidence views over those records.
The current implementation already supports durable lineage, duration and cost observations, verification state, human interventions, policy denials, retries, blocked work, and pending decisions. Factory Configuration versions bound maximum cost, runtime, and attempts. These capabilities make flow and operating diagnostics inspectable today.
Curriculum maturity is not implementation proof.
This chapter defines architecture or practice. It does not by itself prove a corresponding production implementation.
Review this chapter.
Challenge a claim, boundary, missing failure mode, unclear term, or unsupported evidence statement.
- Claim
- Boundary
- Failure
- Evidence