Product Opportunity Recipe
Operate an evidence program, not a long report. Route only methods needed by the current decision; preserve each method's native artifacts; wrap outputs only at handoff; reopen dependent decisions when upstream evidence changes.
Initialize the assessment controller
Read:
- method routing map before activating a method;
- gate specifications before judging readiness;
- handoff and claim contract when a method output enters the recipe.
Create:
text
assessment/
├── controller.yaml
├── evidence/
│ ├── claim-ledger.*
│ └── contradiction-register.*
├── methods/<method>/<version>/...
├── gates/gate-record.*
└── decision/
Use this controller:
yaml
decision:
decision_owner:
investment_at_stake:
decision_date:
scope:
evidence_cutoff:
current_gate:
gate_applicability:
- gate:
disposition: mandatory | conditional | not_applicable
rationale:
required_handoff_types:
version_policy: current_at_gate | same_option_and_cutoff | named_baseline
method_routes:
- method:
purpose:
entry_condition:
expected_native_artifact:
disposition: activated | deferred_predecessor_missing | omitted_not_needed
status: not_started | queued | active | complete | blocked
omission_reason:
artifacts:
claims:
open_fatal_assumptions:
stale_dependencies:
next_decision:
The decision owner approves the gate-applicability and mandatory-dependency manifest before downstream routing. A method omission cannot waive a required handoff: change gate applicability explicitly, with rationale and authority. Predeclare handoff types and version policy; bind concrete handoff IDs and versions later in each gate record. The controller is current when every activated method has one decision purpose, entry condition, and expected native artifact, and every mandatory gate has an approved dependency policy.
G0 — Frame the opportunity and authorization
Use iso-56007-opportunity-validation to prepare the front end, identify the solution-free opportunity, create plural concepts, and define validation burden.
Record:
- opportunity versus concepts;
- decision owner and commitment being considered;
- interested and affected parties;
- fatal value, feasibility, viability, legal, safety, or organizational dimensions;
- alternative concepts and business as usual;
- what passing G0 authorizes.
If opportunity, decision authority, or plural concepts are missing, stop at G0. Do not compensate with market estimates or financial forecasts.
G1 — Make the evidence process credible
G1 requires a credible evidence process, not two methods by default. Activate market-research-design when sampling, instruments, or fieldwork must be fixed or audited. Activate evidence-triangulation when provenance, contradiction, or confidence across sources blocks a decision. Native method records may satisfy G1 without either full method when their own design and lineage controls meet the gate.
G1 controls evidence generated by every later method. It does not replace their native records.
When evidence arrives without a protocol:
- preserve it;
- assess method and inference limits;
- classify it exploratory;
- design prospective confirmation where the decision requires it.
Route the next uncertainty
At every gate, select the next method from the decision-blocking uncertainty—not from a fixed suite order.
| Blocking uncertainty | Route |
|---|
| Population, volume, or spend | market-sizing |
| Meaningful market partitions / beachhead | market-segmentation |
| External change | pestle-analysis |
| Industry profit pressure | porters-five-forces |
| Customer choice and differentiation | competitive-analysis |
| Problem behavior, buyer, workflow | customer-discovery |
| Switch, hire, progress, competing alternatives | jobs-to-be-done |
| Load-bearing concept beliefs | assumption-mapping |
| Consequential demand before build | pretotyping |
| Customer-item ↔ offer-item fit | value-proposition-canvas |
| Integrated value/delivery/capture system | business-model-canvas |
| Price metric, level, package, or elasticity | pricing-research |
| Geography/channel/partner commitment sequence | market-entry-strategy |
| Outcome→opportunity→solution discovery decisions | opportunity-solution-tree |
| Integrated pre-investment operability | unido-feasibility-study |
| Per-unit contribution, retention, payback, capacity | unit-economics |
| Cash, funding, statements, returns, scenarios | financial-modeling |
| Uncertainty, treatment, residual acceptance | iso-31000-risk-assessment |
| Final integrated investment case | five-case-model |
Open a method only when its entry condition is satisfied. Record routing
separately from execution
;
deferred_predecessor_missing
remains
, while
can move through queued, active, complete, or blocked.
G2 — Bound a reachable arena
Activate only the market methods required by the charter. Reconcile shared boundary variables—customer, unit, geography, segment, period, value basis, industry layer, and alternative set—at handoff.
Keep their conclusions distinct:
- market size answers how many/how much;
- segmentation answers which partitions respond differently;
- PESTLE answers which external drivers may move;
- Five Forces answers how industry structure divides value;
- competitive analysis answers what customers choose and why.
If two methods require different valid boundaries, preserve both and state the decision consequence.
G3 — Establish problem and buying behavior
Use customer discovery for eligible episodes, operational evidence, stakeholder roles, and commitments. Use JTBD only when switch/hire episodes can reconstruct circumstances and forces.
G3 passes only within the observed population and evidence boundary. Route prevalence to an appropriate quantitative design; route demand to consequential behavior.
G4 — Develop and test business options
G4 needs at least one coherent value/delivery/capture option, credible alternatives, and enough evidence on assumptions fatal to the requested commitment. Inspect current native handoffs first, then route only unresolved blockers:
- use assumption mapping only when load-bearing beliefs are not already explicit and prioritized;
- use pretotyping only when a consequential pre-build behavior test can resolve a fatal demand belief;
- use VPC/BMC when fit or system architecture is not yet explicit;
- use pricing or entry when its decision could change the architecture;
- use an OST only when its prerequisites pass and continuous discovery needs a shared decision surface.
Keep at least one credible alternative alive until evidence rejects it. A coherent canvas is a hypothesis; a passed pretotype authorizes only the next learning commitment.
G5 — Establish integrated feasibility
Inspect gate applicability and current handoffs first. When the requested commitment needs an integrated pre-investment operability finding and no current artifact satisfies the subsystem/interface checks, activate UNIDO to create/freeze the design basis and reconcile market/sales, capacity, inputs, technology/data, operations, organization, implementation, legal/environmental/social conditions, and financial implications. Otherwise record the gate as not applicable or cite the current native handoff that satisfies it.
Specialist legal, regulatory, privacy, security, safety, AI-evaluation, manufacturing, marketplace-liquidity, or supply-chain work enters as a named conditional module. Record its question, authority, native artifact, and gate consequence.
G6 — Establish sustainable economics and cash
Set the economic evidence required by the requested commitment. Activate unit economics when marginal/cohort contribution, retention, payback, capacity, or working capital is blocking. Activate financial modeling when organization/project cash, funding, statements, returns, or scenario integration is blocking. A current native handoff from another method may satisfy a requirement if it passes the same gate checks; do not run both methods mechanically.
Before closing, reconcile:
text
same unit and segment
same volume/ramp and retention
same price and revenue timing
same capacity and implementation schedule
same cost and working-capital boundaries
same risk and downside assumptions
Positive LTV does not pass G6 when cash or capacity is fatal.
G7 — Decide whether residual risk is acceptable
Inspect gate applicability and current risk handoffs first. Activate ISO 31000 only when the selected option has decision-material uncertainty requiring criteria, treatment, monitoring, or formal acceptance and no current artifact satisfies those checks. Risk treatment must connect to causes, controls, feasibility conditions, model scenarios, owners, triggers, and contingencies.
The authorized risk owner—not the analyst—accepts residual exposure. Fatal legal, safety, ethical, security, or continuity risk remains non-averagable.
Recompute through the invalidation graph
Every handoff declares inputs and outputs. When a material upstream claim, definition, assumption, option, or cutoff changes:
- create a new claim or artifact version; preserve the old version;
- locate every downstream use through used_by and depends_on;
- mark affected handoffs stale;
- reopen affected gates;
- rerun only methods whose decisions could change;
- record the reason and new result.
A polished stale artifact cannot support a gate.
G8 — Develop and challenge the investment case
Activate five-case-model only when the mandatory dependency set for the decision is current. Use the same option set, assumptions, schedule, risks, and evidence across all five cases.
Before the decision:
- run a source/evidence challenge;
- strengthen the counterfactual and minimum-intervention option;
- stress financial optimism and delivery capacity;
- inspect omitted methods and unresolved contradictions;
- test kill conditions and residual-risk authority;
- verify every conclusion against a current handoff.
Issue exactly one:
text
Proceed
Proceed conditionally
Rework / Pivot
Stop
Not decision-ready
Close the assessment
Deliver the controller, method-routing record, native artifact index, claim/evidence coverage, contradiction register, gate record, dependency/staleness graph, Five Case investment case when eligible, red-team challenge, and Decision Record.
Close only when:
- every material decision claim traces to current native evidence;
- every method was activated for a decision reason or omitted with rationale;
- every gate cites its minimum evidence and exact failure route;
- conditions, kill criteria, and risk acceptance have owners;
- missing evidence yields Not decision-ready, never invented confidence.