Twelve-stage operating workflow
Every stage ends in an authorized decision, not another document
The twelve stages run in a fixed order because later conclusions depend on earlier ones. Each stage states the question it answers, what it consumes, what it produces, who owns it, what can stop it, and which human must approve before the project advances.
Product definition
Stage map
The order is fixed because the dependencies are real
Classification constrains the regulatory pathway, the pathway sets the evidence burden, and the evidence burden drives cost, timeline, and capital. Running them out of order is how projects discover fatal problems late.
Twelve-stage development map
Stages run in a fixed order because later conclusions depend on earlier ones. Select a stage to read its question, inputs, outputs, evidence status, owner, gate, dependencies, and next action.
Stage 04HOLDEvidence: Partial
Market, adoption, pricing, and reimbursement
Who buys it, who is paid for it, and through which payment mechanism?
- Inputs
- Target setting, care pathway, competitors, buyer and user roles, coding and payment landscape, jurisdiction.
- Outputs
- Beachhead segment, payment pathway options, adoption model, and the health-economic evidence plan implied by them.
- Owner
- Market, competition, reimbursement, and health-economics roles.
- Decision gate
- GO, CONDITIONAL GO tied to a coverage route, or PIVOT on segment, claim, or geography.
- Depends on
- Stage 3 gate output.
- Next action
- Validate the payment route with a payer-facing conversation or coding analysis before locking claims.
Text equivalent
- Secure intake and normalized project record — Do we hold the rights to this material, and can it become one reliable project record? Owner: Intake, provenance, and privacy roles. Gate: GO once scope, manifest, and access policy are accepted; ABSTAIN when rights or provenance cannot be established.
- Product and jurisdiction classification — What kind of product is this, and which authority governs it where you intend to sell? Owner: Classification and regulatory roles. Gate: CONDITIONAL GO with the claim clarifications required; ABSTAIN if intended use remains contradictory.
- Clinical need and scientific plausibility — Is the clinical problem worth solving, and is the proposed mechanism scientifically plausible? Owner: Clinical need, scientific, and safety roles. Gate: GO, HOLD pending named evidence, PIVOT on indication or population, or STOP on an unmitigable safety finding.
- Market, adoption, pricing, and reimbursement — Who buys it, who is paid for it, and through which payment mechanism? Owner: Market, competition, reimbursement, and health-economics roles. Gate: GO, CONDITIONAL GO tied to a coverage route, or PIVOT on segment, claim, or geography.
- Patent landscape and preliminary freedom-to-operate hypotheses — Is there protectable space, and which third-party rights need qualified legal review? Owner: Intellectual-property analysis role, escalating to qualified counsel. Gate: CONDITIONAL GO subject to counsel review; HOLD or STOP on a material unresolved conflict.
- Regulatory pathway and scenario mapping — Which submission pathway is defensible, and what evidence does each scenario demand? Owner: Regulatory strategy, precedent, and standards roles. Gate: GO on a selected scenario, PIVOT on claims to reach a viable pathway, or ABSTAIN where precedent is genuinely absent.
- Engineering and design controls — Can the intended use become a safe, verifiable design under design controls? Owner: Systems, software, human-factors, and cybersecurity engineering roles. Gate: GO, CONDITIONAL GO on feasibility work, or PIVOT on architecture or intended use.
- Manufacturing, supply chain, DFM, and quality readiness — Can this be made consistently, compliantly, and at a cost the market supports? Owner: Quality, supplier, and manufacturing roles. Gate: CONDITIONAL GO with the named quality and supply actions; HOLD where cost of goods breaks the commercial model.
- Preclinical, clinical, and verification-and-validation evidence strategy — Which studies and tests actually answer the questions regulators, payers, and clinicians will ask? Owner: Preclinical, clinical, and statistical roles. Gate: GO on an integrated plan, or HOLD until endpoint feasibility is resolved.
- Milestone capital, probabilistic cost and timeline, and risk-adjusted valuation — What will the next milestone cost, how long will it take, and what is the venture plausibly worth? Owner: Simulation, valuation, and capital roles. Gate: GO, CONDITIONAL GO on scope reduction, or PIVOT to a narrower first milestone.
- Investor, partner, and vendor matching — Who is genuinely able to finance or execute the next milestone, and may they see this project? Owner: Investor, partner, and conflict-screening roles. Gate: GO only after project qualification and explicit disclosure approval; otherwise HOLD.
- Continuous audit, outcome capture, and model calibration — What has changed, which prior decisions are now invalid, and how well were we calibrated? Owner: Monitoring, calibration, and adjudication roles. Gate: Re-issued gate results on every affected decision rather than a new standalone report.
| Stage | Question answered | Domain owner | Decision gate |
|---|---|---|---|
| 1. Secure intake and normalized project record | Do we hold the rights to this material, and can it become one reliable project record? | Intake, provenance, and privacy roles. | GO once scope, manifest, and access policy are accepted; ABSTAIN when rights or provenance cannot be established. |
| 2. Product and jurisdiction classification | What kind of product is this, and which authority governs it where you intend to sell? | Classification and regulatory roles. | CONDITIONAL GO with the claim clarifications required; ABSTAIN if intended use remains contradictory. |
| 3. Clinical need and scientific plausibility | Is the clinical problem worth solving, and is the proposed mechanism scientifically plausible? | Clinical need, scientific, and safety roles. | GO, HOLD pending named evidence, PIVOT on indication or population, or STOP on an unmitigable safety finding. |
| 4. Market, adoption, pricing, and reimbursement | Who buys it, who is paid for it, and through which payment mechanism? | Market, competition, reimbursement, and health-economics roles. | GO, CONDITIONAL GO tied to a coverage route, or PIVOT on segment, claim, or geography. |
| 5. Patent landscape and preliminary freedom-to-operate hypotheses | Is there protectable space, and which third-party rights need qualified legal review? | Intellectual-property analysis role, escalating to qualified counsel. | CONDITIONAL GO subject to counsel review; HOLD or STOP on a material unresolved conflict. |
| 6. Regulatory pathway and scenario mapping | Which submission pathway is defensible, and what evidence does each scenario demand? | Regulatory strategy, precedent, and standards roles. | GO on a selected scenario, PIVOT on claims to reach a viable pathway, or ABSTAIN where precedent is genuinely absent. |
| 7. Engineering and design controls | Can the intended use become a safe, verifiable design under design controls? | Systems, software, human-factors, and cybersecurity engineering roles. | GO, CONDITIONAL GO on feasibility work, or PIVOT on architecture or intended use. |
| 8. Manufacturing, supply chain, DFM, and quality readiness | Can this be made consistently, compliantly, and at a cost the market supports? | Quality, supplier, and manufacturing roles. | CONDITIONAL GO with the named quality and supply actions; HOLD where cost of goods breaks the commercial model. |
| 9. Preclinical, clinical, and verification-and-validation evidence strategy | Which studies and tests actually answer the questions regulators, payers, and clinicians will ask? | Preclinical, clinical, and statistical roles. | GO on an integrated plan, or HOLD until endpoint feasibility is resolved. |
| 10. Milestone capital, probabilistic cost and timeline, and risk-adjusted valuation | What will the next milestone cost, how long will it take, and what is the venture plausibly worth? | Simulation, valuation, and capital roles. | GO, CONDITIONAL GO on scope reduction, or PIVOT to a narrower first milestone. |
| 11. Investor, partner, and vendor matching | Who is genuinely able to finance or execute the next milestone, and may they see this project? | Investor, partner, and conflict-screening roles. | GO only after project qualification and explicit disclosure approval; otherwise HOLD. |
| 12. Continuous audit, outcome capture, and model calibration | What has changed, which prior decisions are now invalid, and how well were we calibrated? | Monitoring, calibration, and adjudication roles. | Re-issued gate results on every affected decision rather than a new standalone report. |
Stage detail
What each stage consumes, produces, and requires
Stage 01
Secure intake and normalized project record
Do we hold the rights to this material, and can it become one reliable project record?
- Inputs
- Uploaded documents, guided intake answers, data-rights and consent terms, tenant and project identity.
- Analysis
- Virus scanning, sandboxed parsing, hashing and receipting, extraction into a normalized schema, deduplication against material already held.
- Evidence
- File manifest with hashes, extraction spans linked to source pages, signed acceptance of purpose and retention terms.
- Outputs
- Normalized project record covering product identity, intended use, population, users, mechanism, jurisdiction, stage, and decision sought.
- Domain owner
- Intake, provenance, and privacy roles.
- Decision gate
- GO once scope, manifest, and access policy are accepted; ABSTAIN when rights or provenance cannot be established.
- Next action
- Confirm the extracted critical facts before any downstream stage reads them.
- Required human approval
- Project owner accepts the record; privacy reviewer approves any restricted data class.
Possible blockers
- Unclear or absent data rights
- Unusable or unreadable source material
- Personal or patient data submitted outside an authorized path
Stage 02
Product and jurisdiction classification
What kind of product is this, and which authority governs it where you intend to sell?
- Inputs
- Confirmed intended use, indication, claims, mechanism, user, environment, and target jurisdictions.
- Analysis
- Comparison against classification rules and precedent, separation of device, SaMD, diagnostic, combination-product, and drug-adjacent characteristics, scenario construction where claims are ambiguous.
- Evidence
- Cited classification rules, precedent products, and the specific claim language that drives each scenario.
- Outputs
- Ranked classification scenarios with the claim changes that would move the product between them.
- Domain owner
- Classification and regulatory roles.
- Decision gate
- CONDITIONAL GO with the claim clarifications required; ABSTAIN if intended use remains contradictory.
- Next action
- Fix the claim language, or carry both scenarios forward with their separate evidence burdens.
- Required human approval
- Regulatory reviewer confirms the scenario set before downstream planning relies on it.
Possible blockers
- Claims too vague to classify
- Intended use inconsistent across submitted material
- Jurisdiction outside current supported scope
Stage 03
Clinical need and scientific plausibility
Is the clinical problem worth solving, and is the proposed mechanism scientifically plausible?
- Inputs
- Clinical problem statement, mechanism description, prior literature, bench or early data, standard of care.
- Analysis
- Assessment of unmet need against current care, mechanism plausibility, plausible harm modes, and the strength and directness of any cited literature.
- Evidence
- Literature with authority, directness, independence, recency, and applicability recorded separately, plus explicit contradiction flags.
- Outputs
- Benefit-risk hypothesis, articulated clinical value, and the ranked evidence gaps that would change it.
- Domain owner
- Clinical need, scientific, and safety roles.
- Decision gate
- GO, HOLD pending named evidence, PIVOT on indication or population, or STOP on an unmitigable safety finding.
- Next action
- Run the highest-value study or bench test identified as most likely to change this decision.
- Required human approval
- Clinical reviewer signs the benefit-risk hypothesis; safety is a hard gate and cannot be overridden by score.
Possible blockers
- Mechanism contradicted by credible evidence
- Foreseeable harm without a mitigation route
- No meaningful improvement over standard of care
Stage 04
Market, adoption, pricing, and reimbursement
Who buys it, who is paid for it, and through which payment mechanism?
- Inputs
- Target setting, care pathway, competitors, buyer and user roles, coding and payment landscape, jurisdiction.
- Analysis
- Beachhead definition, adoption-friction modelling, competitive substitution, coding and coverage route assessment, and the economic evidence a payer would require.
- Evidence
- Cited coding and coverage material, comparator products, and published economic evidence, each with its applicability recorded.
- Outputs
- Beachhead segment, payment pathway options, adoption model, and the health-economic evidence plan implied by them.
- Domain owner
- Market, competition, reimbursement, and health-economics roles.
- Decision gate
- GO, CONDITIONAL GO tied to a coverage route, or PIVOT on segment, claim, or geography.
- Next action
- Validate the payment route with a payer-facing conversation or coding analysis before locking claims.
- Required human approval
- Commercial reviewer approves the beachhead and payment assumption set.
Possible blockers
- No viable payment mechanism in the target jurisdiction
- Buyer economics that cannot absorb the price
- Evidence expectations that exceed the capital plan
Stage 05
Patent landscape and preliminary freedom-to-operate hypotheses
Is there protectable space, and which third-party rights need qualified legal review?
- Inputs
- Technical disclosure, claim concepts, product architecture, target jurisdictions, filing history.
- Analysis
- Landscape mapping, white-space identification, and construction of preliminary freedom-to-operate hypotheses with a structured red-team pass against them.
- Evidence
- Cited patent families and publications, with the specific claim elements that create each hypothesis.
- Outputs
- White-space map, potential conflicts, and a prioritized question list written for counsel.
- Domain owner
- Intellectual-property analysis role, escalating to qualified counsel.
- Decision gate
- CONDITIONAL GO subject to counsel review; HOLD or STOP on a material unresolved conflict.
- Next action
- Instruct counsel with the prepared question list rather than an open-ended search.
- Required human approval
- Qualified legal counsel owns any freedom-to-operate opinion. Decidian never issues one.
Possible blockers
- Dominant blocking right with no design-around
- Disclosure events that may have compromised novelty
- Unclear ownership or inventorship
Stage 06
Regulatory pathway and scenario mapping
Which submission pathway is defensible, and what evidence does each scenario demand?
- Inputs
- Classification scenarios, benefit-risk hypothesis, claims, standards landscape, precedent submissions.
- Analysis
- Pathway scenario construction, predicate and precedent analysis, applicable-standards mapping, evidence-burden comparison, and agency-interaction planning.
- Evidence
- Cited guidance, recognized standards, and precedent products, with jurisdiction and recency recorded.
- Outputs
- Pathway scenarios with requirement sets, testing implications, timelines, and a proposed agency-interaction plan.
- Domain owner
- Regulatory strategy, precedent, and standards roles.
- Decision gate
- GO on a selected scenario, PIVOT on claims to reach a viable pathway, or ABSTAIN where precedent is genuinely absent.
- Next action
- Prepare the pre-submission package or narrow claims to the defensible scenario.
- Required human approval
- Regulatory professional of record approves the pathway. Regulatory feasibility is a hard gate.
Possible blockers
- No defensible pathway for the current claims
- Evidence burden incompatible with the capital plan
- Claims that trigger an unintended jurisdictional category
Stage 07
Engineering and design controls
Can the intended use become a safe, verifiable design under design controls?
- Inputs
- Intended use, user and use environment, claims, standards set, risk landscape, existing prototypes.
- Analysis
- Translation of user needs into design inputs, architecture and interface definition, risk analysis, human-factors planning, software lifecycle and cybersecurity threat modelling.
- Evidence
- Traceability from user need to design input to verification method, with standards references attached.
- Outputs
- Design inputs, architecture, risk file structure, human-factors plan, and a verification-and-validation plan.
- Domain owner
- Systems, software, human-factors, and cybersecurity engineering roles.
- Decision gate
- GO, CONDITIONAL GO on feasibility work, or PIVOT on architecture or intended use.
- Next action
- Execute the identified feasibility test before committing to detailed design.
- Required human approval
- Engineering lead approves design inputs; all builds and testing are performed by people, not the platform.
Possible blockers
- Requirement that no known technology satisfies
- Use-related risk without an effective control
- Cybersecurity exposure incompatible with intended use
Stage 08
Manufacturing, supply chain, DFM, and quality readiness
Can this be made consistently, compliantly, and at a cost the market supports?
- Inputs
- Design architecture, materials and components, volume assumptions, target cost, quality-system status.
- Analysis
- Design-for-manufacture review, process and validation planning, supplier and single-source risk assessment, quality-system gap analysis against the applicable standard.
- Evidence
- Component and supplier data, process capability assumptions, and quality-system documentation status.
- Outputs
- DFM findings, quality-system gap list, validation and transfer plan, and a cost-of-goods range.
- Domain owner
- Quality, supplier, and manufacturing roles.
- Decision gate
- CONDITIONAL GO with the named quality and supply actions; HOLD where cost of goods breaks the commercial model.
- Next action
- Close the highest-risk quality-system or supply gap before design freeze.
- Required human approval
- Quality authority approves the plan and retains release authority throughout.
Possible blockers
- Critical single-source component
- Process incapable of the required tolerance
- Quality system absent where the pathway requires it
Stage 09
Preclinical, clinical, and verification-and-validation evidence strategy
Which studies and tests actually answer the questions regulators, payers, and clinicians will ask?
- Inputs
- Pathway requirements, payer evidence expectations, risk file, design inputs, prior data.
- Analysis
- Integrated evidence planning across bench, preclinical, clinical, and human-factors work, with endpoint selection, statistical framing, and sequencing to retire the largest uncertainty first.
- Evidence
- Standards and guidance references, comparable study designs, and existing internal data with its applicability recorded.
- Outputs
- Integrated evidence plan, protocol concepts, sequencing, and cost and duration ranges per study.
- Domain owner
- Preclinical, clinical, and statistical roles.
- Decision gate
- GO on an integrated plan, or HOLD until endpoint feasibility is resolved.
- Next action
- Commission the sequenced first study rather than the largest one.
- Required human approval
- Clinical and regulatory reviewers approve the plan; investigators retain all study conduct and interpretation.
Possible blockers
- Endpoint that cannot be measured within the plan
- Study cost or duration that exceeds available capital
- Evidence needs that conflict between regulator and payer
Stage 10
Milestone capital, probabilistic cost and timeline, and risk-adjusted valuation
What will the next milestone cost, how long will it take, and what is the venture plausibly worth?
- Inputs
- Evidence plan, engineering and manufacturing plans, regulatory scenario, market model, current burn and team.
- Analysis
- Probabilistic simulation with correlated delays, milestone-based capital structuring, and risk-adjusted valuation across scenarios.
- Evidence
- Cost and duration inputs traced to plans and comparables, with assumptions and sensitivity drivers listed.
- Outputs
- P50 and P80 cost and schedule ranges, milestone capital plan, use of funds, and a valuation range by scenario.
- Domain owner
- Simulation, valuation, and capital roles.
- Decision gate
- GO, CONDITIONAL GO on scope reduction, or PIVOT to a narrower first milestone.
- Next action
- Restructure the raise around the milestone that most reduces risk per dollar.
- Required human approval
- Founder or board approves the capital plan. Valuation output is analysis, never investment advice.
Possible blockers
- Capital requirement far beyond plausible raise
- Timeline that expires the IP or market window
- Uncertainty too wide to structure a milestone
Stage 11
Investor, partner, and vendor matching
Who is genuinely able to finance or execute the next milestone, and may they see this project?
- Inputs
- Qualified project record, disclosure approvals, investor mandates, partner qualification records, conflict data.
- Analysis
- Requirement-driven matching against mandates and verified partner capability, conflict screening, and construction of a consented disclosure package.
- Evidence
- Transparent match reasons tied to specific project attributes and stated mandate criteria.
- Outputs
- Ranked, permissioned matches with the reasoning shown to both sides.
- Domain owner
- Investor, partner, and conflict-screening roles.
- Decision gate
- GO only after project qualification and explicit disclosure approval; otherwise HOLD.
- Next action
- Release the consented package to the shortlist and record any nonbinding interest.
- Required human approval
- Project owner authorizes every disclosure and can revoke access at any time.
Possible blockers
- Disclosure not approved by the project
- Unresolved conflict of interest
- Project not yet qualified for release
Stage 12
Continuous audit, outcome capture, and model calibration
What has changed, which prior decisions are now invalid, and how well were we calibrated?
- Inputs
- New evidence, milestone results, regulatory and standards changes, competitor and market events, realized costs and timelines.
- Analysis
- Dependency-driven invalidation, re-evaluation of affected decisions, and scoring of prior forecasts against observed outcomes.
- Evidence
- Change records linked to the decisions they affect, plus realized-versus-forecast comparisons.
- Outputs
- Alerts, invalidated decisions, updated conclusions, and recalibrated forecast models by product type and stage.
- Domain owner
- Monitoring, calibration, and adjudication roles.
- Decision gate
- Re-issued gate results on every affected decision rather than a new standalone report.
- Next action
- Review invalidated decisions with the accountable reviewer.
- Required human approval
- Reviewers re-authorize any decision whose basis has materially changed.
Possible blockers
- Stale project record with no reporting
- Outcomes withheld, preventing calibration
Find out which stage your project is actually at
Most programmes are further back than their deck suggests. We identify the stage you are at and the gate output missing before the next one.