Growth roadmap: activation, evidence, and distribution

[cderinbogaz]

Objective

Turn Juror's strong initial developer interest into retained repositories completing successful multi-model reviews.

This epic tracks the product, research, trust, community, and distribution work identified in the August 2026 growth review.

Phase 1 — Activation and trust

• Add a one-command juror init onboarding flow #40
• Add a one-secret multi-model starter preset #41
• Add SECURITY.md and verifiable Action release provenance #47

Phase 2 — Representative proof

• Expand the adjudicated benchmark to 20–30 representative PRs #42
• Evaluate Juror in Martian's 50-PR review benchmark #43
• Publish a recurring cross-model Jury Report #46

Phase 3 — OSS adoption and PR-native distribution

• Run a 10-repository open-source lighthouse program #44
• Build an opt-in gallery of real Juror reviews #45

Phase 4 — Positioning and launch

• Reposition Juror around cross-vendor auditability #48
• Execute a post-proof launch and directory distribution checklist #50

Measurement and community infrastructure

• Define privacy-preserving activation, retention, and quality metrics #49
• Create contributor paths for jurors, harnesses, and benchmark cases #51

Intended outcomes

• A new repository can reach a successful multi-model review with minimal setup.
• Juror's quality claims are supported by a representative, reproducible benchmark.
• Public OSS usage creates an ethical, opt-in distribution loop.
• Security-sensitive adopters can audit the Action and its release chain.
• Growth is measured by retained reviewing repositories rather than stars.

North-star metric

Weekly active repositories completing successful reviews with at least two model families, measured with a privacy-preserving and preferably opt-in approach.

Recommended execution order

1. Ship onboarding and release-chain trust work.
2. Establish the measurement baseline.
3. Recruit lighthouse repositories while expanding the internal and external benchmarks.
4. Build the public gallery and first Jury Report from permissioned evidence.
5. Update positioning, then launch on evidence rather than the one-PR seed.

Non-goals

• Paid acquisition before activation and retention are understood.
• Inflated claims based on the current one-PR seed benchmark.
• Unsolicited bot installation or spammy promotion on third-party repositories.