Same packet → same rank. Reproducible scoring matrix.
AI advises; you record the authoritative call.
Immutable trail for board, legal, and diligence.
Onboarding
Your first governed decision (~15 min)
Follow these steps for a complete trade study with audit-ready output.
- 0Inspect a finished decision
See ranking, why the runner-up lost, and a hashed receipt — no sign-in.
Go → - 1Start this trade
Open the Harbor CRM packet you inspected. Community is manual-only; Founding Pilot adds AI assist.
Go → - 2Complete options & criteria
Add 2–7 options, weights that sum to 1.0, and scores 0–100.
Go → - 3Pass readiness gates
Resolve blocking items on the readiness panel before scoring.
Go → - 4Score, review, finalize
Deterministic ranking → critique → brief → human ACCEPT/OVERRIDE/REVISE.
Go → - 5Export audit trail
PDF on all tiers; Founding Pilot audit bundle with integrity digest (HMAC when configured).
Go →
Packet architecture
Seven steps · one scoreable matrix
Each step maps 1:1 to Genius navigation. Complete all seven before Run score.
- GoalProblem · Explore
- OptionsPrepare
- ConstraintsPrepare
- CriteriaEvaluate
- ScoresEvaluate
- AssumptionsPrepare · Evaluate
- ReviewApprove
After Genius
Score → brief → record
Packet completeness unlocks scoring. Every downstream step preserves provenance.
EngineRun scoreDeterministic rank EvaluateScorecardOption matrix ProCritiqueAI review (optional) ApproveBriefExecutive summary HITLAccept · Override · ReviseHuman gate CommitFinalizeLock record ReceiptAuditExport bundle
Catalog
Seven worked trade studies
Field-accurate copy for Genius. Select by complexity, margin, or governance path.
Ex 1
Clear winner
Analytics platform vendor
Best forFull trade study; strong Accept path.
Ex 2
Narrow margin
Cloud deployment model
Best forClose call; weak winner margin; possible **Override**.
Ex 3
Cost ceiling
Office lease
Best forEncoding must-haves in criteria scores when you skip constraints.
Ex 4
Risk in text
Acquire vs build vs partner
Best forDocumenting integration risk in **Top risk**; compare rankings after **Run score**.
Ex 5
Engineering trade study
API gateway
Best forIdeal 3–5 option shortlist; fill **Why? (optional)** on close scores.
Ex 6
Fast path
Hire FTE vs contractor
Best forFirst decision in under 15 minutes; **Standard** governance.
Ex 7
Portfolio pick at ceiling
Portfolio pick at ceiling
Best forMaximum shortlist; watch the **Shortlist is long** pill on the Options step.
Field sync
Genius — live UI mapping
Labels below are generated from production Genius. When the UI changes, this table updates automatically.
| Step | Nav | What you fill in |
|---|---|---|
| Draft | Start screen | Title, optional Template (optional) → Start |
| 1 | Goal | What are you deciding?, Background (optional), Decide by. Advanced: Decision type, Governance level |
| 2 | Options | Per option: Name, Trust in scores, Notes (optional). Add option / Remove (2–7; 3–5 ideal) |
| 3 | Constraints | Add constraint for must-haves, or Next to skip |
| 4 | Criteria | Per row: Name, Weight, Notes (optional). Weights sum to 1.00 |
| 5 | Scores | Score (0–100), optional Why? (optional) |
| 6 | Assumptions | Key assumption, What would prove you wrong?, Best evidence, Top risk, Mitigation (optional) |
| 7 | Review | Decision window, Outcome confidence, Check again in (days), What success looks like (optional) |
Footer: Back · Next · Save draft · Run score — Run score appears when the packet is 100% complete.
Trust in scores — Pick any quick level first (High — strong evidence, Medium — reasonable estimate, Low — mostly guesswork, or Not sure). The full preset list appears next — set the value from the guide table (e.g. High (0.8) or Custom… → 0.75).
Quick picker: High — strong evidence; Medium — reasonable estimate; Low — mostly guesswork; Not sure. Then presets: Very low (0.2), Low (0.4), Moderate (0.6), High (0.8), Very high (1), or Custom….
Decision window — choose 30 days, 60 days, 90 days, 6 months, 1 year, or Custom… and enter a value like 18_months.
Governance level — Standard, Consulting, Enterprise. Standard allows 2 criteria; Consulting and Enterprise expect fuller packets.
Seven worked examples for /decisions/new. Each lists exactly what to enter on each Genius step screen.
Leave Template (optional) on Blank trade study unless an example says otherwise.
Autosave: Progress saves on every step. Use Save draft or return via Home → Your drafts → Continue.
Between steps: Footer actions are Back, Next, and Save draft. Run score appears on Review when the packet is 100% complete.
Trust in scores (Options step): Pick any quick level first (High — strong evidence, Medium — reasonable estimate, Low — mostly guesswork, or Not sure). The full preset list appears next — set the value from the guide table (e.g. High (0.8) or Custom… → 0.75).
Analytics platform vendor
Draft
| Field | Value |
|---|---|
| Decision type | Vendor / procurement |
| Template (optional) | Blank trade study |
| Title | Q3 Analytics Platform Vendor Selection |
Click Start.
1
Goal
Open Advanced on this step.
| UI field | Value |
|---|---|
| What are you deciding? | Select the analytics platform we will standardize on for product, marketing, and finance reporting through FY2027. |
| Background (optional) | Current Mixpanel + spreadsheets fragment metrics. Need SSO, warehouse-native queries, and under $120k ARR Year 1. Evaluation committee: Product, Data, Finance. |
| Decide by | 2026-07-15 |
| Decision type | vendor |
| Governance level | Consulting |
Click Next.
2
Options
The Genius starts with two empty option rows. Click Add option twice (4 options total).
For each option, set Name, Trust in scores, and Notes (optional):
| # | Name | Trust in scores | Notes (optional) |
|---|---|---|---|
| 1 | Amplitude | High (0.8) | Strong product analytics; weaker finance cubes |
| 2 | Looker (Google) | High (0.8) | Warehouse-native; heavier admin burden |
| 3 | Metabase Cloud | Custom → 0.75 | Low cost; limited enterprise governance |
| 4 | Heap | Moderate (0.6) | Auto-capture; privacy review pending |
Click Next.
3
Constraints
No constraints required for this example. Click Next (or Add constraint if you have must-haves).
4
Criteria
The Genius starts with two empty criteria rows. Click Add criterion twice (4 total). Set weights to sum 1.00 (watch the Total pill).
| Name | Notes (optional) | Weight |
|---|---|---|
| Business impact | Revenue/retention insight quality | 0.35 |
| Implementation feasibility | Time-to-value ≤ 90 days | 0.25 |
| Total cost of ownership | 3-year TCO | 0.25 |
| Security & compliance | SOC2, data residency | 0.15 |
Click Next.
5
Scores
Fill each card — Score (0–100) and optional Why? (optional). Enter every score before Run score (unscored cards block ranking).
| Option | Business impact | Implementation feasibility | Total cost of ownership | Security & compliance |
|---|---|---|---|---|
| Amplitude | 88 | 82 | 70 | 85 |
| Looker (Google) | 85 | 68 | 72 | 90 |
| Metabase Cloud | 72 | 90 | 92 | 65 |
| Heap | 80 | 75 | 78 | 70 |
Example Why? (optional) (Amplitude / Business impact): Strong product analytics depth.
Click Next.
6
Assumptions
| UI field | Value |
|---|---|
| Key assumption | Warehouse (Snowflake) remains our system of record. |
| What would prove you wrong? | A warehouse migration RFP launches in Q4. |
| Best evidence | Vendor demo scorecards (June 2026) |
| Top risk | Heap privacy auto-capture — Severity 3 Likelihood 3 |
| Mitigation (optional) | Complete privacy review before shortlist cut |
Click Next.
7
Review
| UI field | Value |
|---|---|
| Decision window | 90 days |
| Outcome confidence | Custom → 0.78 |
| Check again in (days) | 90 |
| What success looks like (optional) | Platform live for 2 squads; dashboard adoption >60% of PMs. |
Click Run score when the packet shows 100% complete.
Cloud deployment model
Draft
| Field | Value |
|---|---|
| Decision type | Architecture / technical |
| Title | Production Deployment Strategy FY2027 |
Click Start.
1
Goal
| UI field | Value |
|---|---|
| What are you deciding? | Choose how we run customer-facing APIs in production for the next 18 months. |
| Background (optional) | Team of 8 engineers; 99.9% SLA; EU data residency for some tenants. Debated 3 weeks without convergence. |
| Decide by | 2026-08-01 |
| Decision type | architecture |
| Governance level | Consulting |
2
Options
Click Add option once (3 options total).
| # | Name | Trust in scores | Notes (optional) |
|---|---|---|---|
| 1 | Stay on AWS EKS | High (0.8) | Current state; known ops pain on upgrades |
| 2 | Fly.io + managed Postgres | Moderate (0.6) | Faster deploys; smaller enterprise footprint |
| 3 | Multi-cloud AWS + GCP DR | High (0.8) | Resilience; highest complexity |
3
Constraints
Click Next (or Add constraint if you have must-haves).
4
Criteria
| Name | Notes (optional) | Weight |
|---|---|---|
| Reliability & SLA | Uptime and failover posture | 0.40 |
| Engineering velocity | Deploy frequency and lead time | 0.30 |
| Operational cost | 18-month run-rate | 0.20 |
| EU residency | Data residency compliance | 0.10 |
5
Scores
| Option | Reliability & SLA | Engineering velocity | Operational cost | EU residency |
|---|---|---|---|---|
| Stay on AWS EKS | 85 | 70 | 65 | 88 |
| Fly.io + managed Postgres | 75 | 92 | 80 | 72 |
| Multi-cloud AWS + GCP DR | 90 | 55 | 50 | 90 |
6
Assumptions
| UI field | Value |
|---|---|
| Key assumption | Traffic stays under 50k RPS through 2027. |
| What would prove you wrong? | August load test exceeds 55k RPS sustained. |
| Best evidence | Q2 capacity planning spreadsheet |
| Top risk | Fly.io enterprise support gap — Severity 2 Likelihood 3 |
7
Review
| UI field | Value |
|---|---|
| Decision window | Custom… → 18_months |
| Outcome confidence | Moderate (0.6) |
| Check again in (days) | 90 |
| What success looks like (optional) | No Sev-1 from deployment model before next architecture review. |
Office lease
Draft
| Field | Value |
|---|---|
| Decision type | Business strategy |
| Title | Austin HQ Lease Renewal vs Relocation |
Click Start.
1
Goal
| UI field | Value |
|---|---|
| What are you deciding? | Select office arrangement for 45-person HQ from October 2026. |
| Background (optional) | Current lease ends Sept 30. Must seat 50 with hybrid ratio 60/40. Max all-in rent $42k/month is non-negotiable. |
| Decide by | 2026-09-01 |
| Decision type | business |
| Governance level | Standard |
2
Options
Click Add option twice (4 total).
| # | Name | Trust in scores | Notes (optional) |
|---|---|---|---|
| 1 | Renew Block 180 | High (0.8) | Current building |
| 2 | Domain Northside | Moderate (0.6) | Premium campus; quotes ~$48k |
| 3 | East Austin flex campus | Moderate (0.6) | Mid-cost flex space |
| 4 | Full remote + quarterly offsites | Low (0.4) | Lowest rent; culture tradeoff |
3
Constraints
Click Next (or Add constraint if you have must-haves).
4
Criteria
Click Add criterion once (3 criteria). Score Domain Northside lower on Cost to reflect over-budget rent.
| Name | Notes (optional) | Weight |
|---|---|---|
| Employee experience | Hybrid workspace quality | 0.40 |
| Cost | All-in monthly cost within $42k cap | 0.35 |
| Recruiting signal | Ability to attract talent | 0.25 |
5
Scores
| Option | Employee experience | Cost | Recruiting signal |
|---|---|---|---|
| Renew Block 180 | 82 | 88 | 80 |
| Domain Northside | 90 | 35 | 92 |
| East Austin flex campus | 78 | 80 | 75 |
| Full remote + quarterly offsites | 65 | 95 | 60 |
6
Assumptions
| UI field | Value |
|---|---|
| Key assumption | Hybrid attendance stays near 60% office. |
| What would prove you wrong? | Quarterly occupancy survey drops below 45%. |
| Best evidence | Broker proposals May 2026 |
| Top risk | Remote-first hurts junior hiring — Severity 3 Likelihood 2 |
7
Review
| UI field | Value |
|---|---|
| Decision window | 1 year |
| Outcome confidence | Moderate (0.6) |
| Check again in (days) | 180 |
| What success looks like (optional) | Lease signed or remote policy published by Oct 1. |
Acquire vs build vs partner
Draft
| Field | Value |
|---|---|
| Decision type | Engineering / build vs buy |
| Title | Acquire vs Build vs Partner — Customer Support AI |
Click Start.
1
Goal
| UI field | Value |
|---|---|
| What are you deciding? | How we deliver Tier-1 support automation by Q1 2027. |
| Background (optional) | Board wants measurable deflection in 9 months. Legal reviewing acquisition path. |
| Decide by | 2026-10-01 |
| Decision type | engineering |
| Governance level | Enterprise |
2
Options
Click Add option once (3 total).
| # | Name | Trust in scores | Notes (optional) |
|---|---|---|---|
| 1 | Acquire Startup X | Moderate (0.6) | $8M; strong tech; integration risk |
| 2 | Build in-house | Moderate (0.6) | 12-month roadmap |
| 3 | Partner with Zendesk AI | High (0.8) | Faster; less differentiation |
3
Constraints
Click Next (or Add constraint if you have must-haves).
4
Criteria
| Name | Notes (optional) | Weight |
|---|---|---|
| Strategic fit | Alignment with product strategy | 0.30 |
| Time to market | Speed to Tier-1 automation | 0.30 |
| Risk-adjusted ROI | Return after integration risk | 0.40 |
5
Scores
| Option | Strategic fit | Time to market | Risk-adjusted ROI |
|---|---|---|---|
| Acquire Startup X | 92 | 88 | 85 |
| Build in-house | 75 | 45 | 90 |
| Partner with Zendesk AI | 70 | 95 | 72 |
6
Assumptions
| UI field | Value |
|---|---|
| Key assumption | Startup X IP is clean for acquisition. |
| What would prove you wrong? | Diligence fails on IP chain of title. |
| Best evidence | LOI term sheet draft |
| Top risk | Acquire Startup X — integration failure — Severity 5 Likelihood 4 |
7
Review
| UI field | Value |
|---|---|
| Decision window | Custom… → 9_months |
| Outcome confidence | Moderate (0.6) |
| Check again in (days) | 60 |
| What success looks like (optional) | 25% ticket deflection in pilot queue. |
API gateway
Draft
| Field | Value |
|---|---|
| Decision type | Architecture / technical |
| Title | API Gateway Pattern Selection |
Click Start.
1
Goal
| UI field | Value |
|---|---|
| What are you deciding? | Standardize edge/API layer for 12 microservices. |
| Background (optional) | Need rate limiting, mTLS, observability, gradual Kong migration off legacy nginx proxy. |
| Decide by | 2026-08-15 |
| Decision type | architecture |
| Governance level | Consulting |
2
Options
Click Add option three times (5 total).
| # | Name | Trust in scores | Notes (optional) |
|---|---|---|---|
| 1 | Kong Gateway Enterprise | High (0.8) | mTLS + plugins; license cost |
| 2 | AWS API Gateway + Lambda authorizers | High (0.8) | Managed; vendor lock-in |
| 3 | Envoy + custom control plane | Moderate (0.6) | Latency win; ops burden |
| 4 | Apigee X | High (0.8) | Enterprise; long migration |
| 5 | Keep nginx + Lua | Very high (1.0) | Status quo; tech debt |
3
Constraints
Click Next (or Add constraint if you have must-haves).
4
Criteria
Click Add criterion three times (5 total).
| Name | Notes (optional) | Weight |
|---|---|---|
| Security & mTLS | Edge security posture | 0.25 |
| Operability | Day-2 ops burden | 0.25 |
| Latency overhead | Added request latency | 0.20 |
| License / TCO (3 yr) | Three-year cost | 0.20 |
| Team familiarity | Existing skills | 0.10 |
5
Scores
Fill all 25 cells. Sample rows:
| Option | Security & mTLS | Operability | Latency overhead | License / TCO | Team familiarity |
|---|---|---|---|---|---|
| Kong Gateway Enterprise | 88 | 82 | 85 | 70 | 75 |
| AWS API Gateway + Lambda authorizers | 85 | 88 | 82 | 78 | 80 |
| Envoy + custom control plane | 90 | 55 | 92 | 85 | 50 |
| Apigee X | 86 | 60 | 80 | 55 | 65 |
| Keep nginx + Lua | 70 | 92 | 88 | 95 | 90 |
Why? (optional) example: Envoy / Operability → Small SRE team cannot run control plane.
6
Assumptions
| UI field | Value |
|---|---|
| Key assumption | Kong contract renews at current tier. |
| What would prove you wrong? | Renewal quote increases more than 20%. |
| Best evidence | Architecture review notes June 2026 |
| Top risk | Envoy path delays migration — Severity 4 Likelihood 3 |
7
Review
| UI field | Value |
|---|---|
| Decision window | 6 months |
| Outcome confidence | High (0.8) |
| Check again in (days) | 90 |
| What success looks like (optional) | One production service on chosen gateway with SLO met. |
Hire FTE vs contractor
Draft
| Field | Value |
|---|---|
| Decision type | Career / job |
| Title | Staff Senior Backend Engineer — H2 2026 |
Click Start.
1
Goal
| UI field | Value |
|---|---|
| What are you deciding? | Fill one senior backend slot by September 1. |
| Background (optional) | Critical path on billing rewrite. Budget $180k loaded FTE or $120/hr contract cap 6 months. |
| Decide by | 2026-07-01 |
| Decision type | career |
| Governance level | Standard |
2
Options
Use the two default rows (no add/remove).
| # | Name | Trust in scores | Notes (optional) |
|---|---|---|---|
| 1 | Full-time hire | Moderate (0.6) | Long-term capacity; slower start |
| 2 | 6-month senior contractor | High (0.8) | Faster start; cap $120/hr |
3
Constraints
Click Next (or Add constraint if you have must-haves).
4
Criteria
Use the two default criteria rows.
| Name | Notes (optional) | Weight |
|---|---|---|
| Delivery by Nov 1 | Billing rewrite on schedule | 0.60 |
| Cost within budget | Loaded FTE vs contract cap | 0.40 |
5
Scores
| Option | Delivery by Nov 1 | Cost within budget |
|---|---|---|
| Full-time hire | 65 | 70 |
| 6-month senior contractor | 88 | 75 |
6
Assumptions
| UI field | Value |
|---|---|
| Key assumption | Billing rewrite scope stays frozen. |
| What would prove you wrong? | Scope change adds more than 2 weeks slip. |
| Best evidence | Engineering roadmap v3 |
| Top risk | Contractor knowledge loss at month 6 — Severity 2 Likelihood 4 |
7
Review
| UI field | Value |
|---|---|
| Decision window | 60 days |
| Outcome confidence | High (0.8) |
| Check again in (days) | 60 |
| What success looks like (optional) | Billing MVP milestone green by Nov 1. |
Portfolio pick at ceiling
Draft
| Field | Value |
|---|---|
| Decision type | Business strategy |
| Title | H2 2026 Engineering Portfolio — Pick Top Initiative |
Click Start.
1
Goal
| UI field | Value |
|---|---|
| What are you deciding? | Select one primary bet for H2 given fixed capacity of 3 squads. |
| Background (optional) | Seven proposals from product council; leadership wants a single board narrative. |
| Decide by | 2026-07-20 |
| Decision type | business |
| Governance level | Enterprise |
2
Options
Click Add option five times (7 total). You may see Shortlist is long.
| # | Name | Trust in scores | Notes (optional) |
|---|---|---|---|
| 1 | Mobile app v2 | Moderate (0.6) | Consumer growth bet |
| 2 | Enterprise SSO | High (0.8) | Revenue unlock for F500 |
| 3 | AI copilot beta | Moderate (0.6) | Differentiation; uncertain ROI |
| 4 | Billing reliability | High (0.8) | Reduces churn risk |
| 5 | Self-serve onboarding | High (0.8) | PLG conversion |
| 6 | Partner API program | High (0.8) | Ecosystem play |
| 7 | Data platform modernization | High (0.8) | Enables all squads |
3
Constraints
Click Next (or Add constraint if you have must-haves).
4
Criteria
Click Add criterion twice (4 total).
| Name | Notes (optional) | Weight |
|---|---|---|
| Revenue impact | Near-term revenue lift | 0.30 |
| Strategic alignment | Board narrative fit | 0.25 |
| Execution confidence | Squad can deliver in H2 | 0.25 |
| Risk mitigation | Downside if initiative fails | 0.20 |
5
Scores
Fill all 28 score cards (7 options × 4 criteria). Sample:
| Option | Revenue impact | Strategic alignment | Execution confidence | Risk mitigation |
|---|---|---|---|---|
| Mobile app v2 | 78 | 72 | 80 | 70 |
| Enterprise SSO | 85 | 88 | 82 | 75 |
| AI copilot beta | 88 | 90 | 55 | 60 |
| Billing reliability | 82 | 78 | 90 | 85 |
| Self-serve onboarding | 80 | 75 | 85 | 72 |
| Partner API program | 76 | 80 | 78 | 70 |
| Data platform modernization | 84 | 86 | 70 | 80 |
6
Assumptions
| UI field | Value |
|---|---|
| Key assumption | Board accepts a single-H2 narrative. |
| What would prove you wrong? | Board requests parallel bet approval. |
| Best evidence | Product council ranking June 2026 |
| Top risk | AI copilot distracts squads — Severity 3 Likelihood 3 |
7
Review
| UI field | Value |
|---|---|
| Decision window | 6 months |
| Outcome confidence | Moderate (0.6) |
| Check again in (days) | 90 |
| What success looks like (optional) | One initiative staffed with committed OKRs by Aug 1. |
Spectrum cheat sheet
| Example | Options | Teaches |
|---|---|---|
| 1 Vendor analytics | 4 | Full Genius; clear Accept |
| 2 Cloud deployment | 3 | Close call / weak margin |
| 3 Office lease | 4 | Must-haves via criteria scores |
| 4 Acquire vs build | 3 | Risk in Top risk field |
| 5 API gateway | 5 | Ideal shortlist + Why? (optional) notes |
| 6 Hire FTE vs contractor | 2 | Fast Standard-tier path |
| 7 Portfolio pick | 7 | Max options + Shortlist is long |
Quick start (under 20 minutes)
- Pick Example 6 (fastest) or Example 1 (richest).
- Open
/decisions/newand this guide side by side. - Copy each step table into the matching Genius screen.
- Save draft anytime; resume from Home → Your drafts.
- Run score → review Scorecard → open Brief page → Run critique → Brief → Accept (or Override with rationale).
Common mistakes
| Mistake | What happens |
|---|---|
| Only 1 criterion | Validation fails — use Add criterion (minimum 2) |
| 8+ options | Cannot add — max 7 |
| Missing What would prove you wrong? | Weaker assumption check at finalize |
| Skipping Trust in scores | Cannot Run score until each option has confidence set |
| Unscored criteria cards | Run score blocked — enter Score (0–100) on every option × criterion |
| Weights ≠ 1.00 | Run score blocked — watch Total pill |