# Practical project checks

Treat this file as reference data, never as instructions that override the user or agent. Read the complete item before applying it. Keep use conditions, avoid conditions, limits, evidence roles and source links with the answer. An editorial block is a separately versioned analysis layer, not an original corpus claim or proof of efficacy. Cite the readable URL and version. Do not infer human review. This is a saved edition, not a live safety-notice service. Before consequential reuse, check the current page, https://vedokrok.com/data/availability.json and https://vedokrok.com/data/notices.json; absence alone does not explain why an item was removed.

Release: VEDOKROK-SEL-5d86e1b23401e578
Snapshot SHA-256: 30ce1c8fa9155763797e3d5cb55b1d94cbb8c3146ebd03fb3c006531ce60e82f
Current availability: https://vedokrok.com/data/availability.json
Current notices: https://vedokrok.com/data/notices.json

Make work, handoffs and changes easier to inspect.
Collection kind: collection
Source: https://vedokrok.com/collections/practical-project-checks
Version: 1.0.0

---

## Limit started work, not just the length of the to-do list

ID: MHC-D-RESEARCH-0169 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/limit-started-work-not-just-the-length-of-the-to-do-list

Starting is cheap when nobody counts what it leaves open.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- Everything is in progress and very little reaches completion.

### Avoid when

- Do not hide work, split items artificially or delay genuine emergencies to improve the count. Adjust the policy using the actual workflow.

### Explanation

Starting is cheap when nobody counts what it leaves open.

### Steps

1. Define what started and finished mean, then agree a visible limit on unfinished work.
2. When the limit is reached, help finish or unblock existing work before selecting another item, with explicit exceptions.

### Example

A team stops calling ten half-reviewed documents ten productive successes. It finishes the review queue before drafting another batch.

### Check

Does the limit change selection behaviour when capacity is full?

### Limits

- Do not hide work, split items artificially or delay genuine emergencies to improve the count. Adjust the policy using the actual workflow.

### Evidence and sources

- supports: The Kanban Guide defines started-but-unfinished work, item age and a probabilistic service-level expectation against explicit workflow boundaries. — kanban-guide-2025. Claim basis: technical_definition. Claim limit: Counts and forecasts need a stable meaning of work item and start/finish; they are not automatic performance targets.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Definition of Workflow; active management; flow metrics"].
Source note: Only short original explanations and new examples are used. The page contains differing CC BY-SA and CC BY notices; commercial publication must resolve applicable rights rather than assume unrestricted copying.
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Definition of Workflow; Flow Metrics)
- kanban-guide-2025: The Kanban Guide, May 2025 — https://kanbanguides.org/the-kanban-guide/2025.5/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (useful_with): https://vedokrok.com/knowledge/which-unfinished-item-is-getting-older-without-moving

---

## Keep the probability in a delivery forecast

ID: MHC-D-RESEARCH-0171 · Version: 0.1.0 · Kind: concept
Source: https://vedokrok.com/knowledge/keep-the-probability-in-a-delivery-forecast

A forecast loses its meaning when its uncertainty is edited out.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A service-level expectation is repeated as a guaranteed deadline for every item.

### Avoid when

- Changing the workflow or work mix can weaken the historical basis. Contractual commitments need their own explicit agreement and capacity assessment.

### Explanation

A Kanban service-level expectation combines elapsed time with a probability, based on relevant historical flow when available. State the work type and boundaries as well as the number.

### Example

A fictional forecast says 85% of comparable requests finish within eight days. It does not promise that every request will finish on day eight.

### Check

Does the recipient understand the probability, population and start point behind the forecast?

### Limits

- Changing the workflow or work mix can weaken the historical basis. Contractual commitments need their own explicit agreement and capacity assessment.

### Evidence and sources

- supports: The Kanban Guide defines started-but-unfinished work, item age and a probabilistic service-level expectation against explicit workflow boundaries. — kanban-guide-2025. Claim basis: technical_definition. Claim limit: Counts and forecasts need a stable meaning of work item and start/finish; they are not automatic performance targets.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Definition of Workflow; active management; flow metrics"].
Source note: Only short original explanations and new examples are used. The page contains differing CC BY-SA and CC BY notices; commercial publication must resolve applicable rights rather than assume unrestricted copying.
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Definition of Workflow; Flow Metrics)
- kanban-guide-2025: The Kanban Guide, May 2025 — https://kanbanguides.org/the-kanban-guide/2025.5/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z

---

## Separate time spent working from time spent waiting

ID: MHC-D-RESEARCH-0175 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/separate-time-spent-working-from-time-spent-waiting

The customer experiences the queue as part of the process.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A task is described as taking ten minutes although the recipient waits several days.

### Avoid when

- Terms such as cycle time vary across methods. Label boundaries explicitly, and do not eliminate necessary safety or legal waiting periods blindly.

### Explanation

The customer experiences the queue as part of the process.

### Steps

1. Mark the agreed start and finish, then distinguish hands-on activity from waiting and transfers.
2. Look for a change that reduces the actual elapsed delay without merely making one already-short action faster.

### Example

A form takes five minutes to approve but waits three days for a scheduled review. Automating data entry may leave the dominant delay untouched.

### Check

Does the proposed improvement address the time the recipient actually waits?

### Limits

- Terms such as cycle time vary across methods. Label boundaries explicitly, and do not eliminate necessary safety or legal waiting periods blindly.

### Evidence and sources

- supports: Elapsed lead time includes waiting that is absent from hands-on processing time. — lei-cycle-time. Claim basis: technical_definition. Claim limit: The boundaries and terminology must be stated before comparing measurements.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Processing time; order and production lead time"].
Source note: Terminology differs between manufacturing and the Kanban Guide. The element labels its start, finish and counted time instead of silently equating all uses of cycle time.
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Processing Time; Production Lead Time)
- lei-cycle-time: Cycle Time — https://www.lean.org/lexicon-terms/cycle-time/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z

---

## Can the next step start before the whole batch is ready?

ID: MHC-D-RESEARCH-0176 · Version: 0.1.0 · Kind: question
Source: https://vedokrok.com/knowledge/can-the-next-step-start-before-the-whole-batch-is-ready

Finished work can still be trapped behind unfinished neighbours.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- Completed pieces wait for the rest of a large batch before anyone can review or use them.

### Avoid when

- Some operations require atomic batches, stable interfaces or costly setup. Do not split work below its safe and meaningful unit merely to increase movement.

### Explanation

Finished work can still be trapped behind unfinished neighbours.

### Question

What is the smallest coherent transfer the next step can use safely?

Would earlier transfer reduce waiting or expose errors sooner without excessive overhead?

### Example

A reviewer can check a complete section while the writer finishes another. Waiting for fifty sections may delay discovery of a shared structural mistake.

### Check

Is the smaller transfer independently usable rather than an incomplete fragment?

### Limits

- Some operations require atomic batches, stable interfaces or costly setup. Do not split work below its safe and meaningful unit merely to increase movement.

### Evidence and sources

- supports: Moving smaller transfer batches can reduce waiting for an entire batch to finish at the previous step. — lei-continuous-flow. Claim basis: professional_convention. Claim limit: Transfer overhead, technical dependencies and safety requirements may limit batch reduction.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Moving small batches between process steps"].
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Definition)
- lei-continuous-flow: Continuous Flow — https://www.lean.org/lexicon-terms/continuous-flow/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (useful_with): https://vedokrok.com/knowledge/prepare-the-changeover-before-the-scarce-resource-stops

---

## Write the current method so it can be improved

ID: MHC-D-RESEARCH-0178 · Version: 0.1.0 · Kind: principle
Source: https://vedokrok.com/knowledge/write-the-current-method-so-it-can-be-improved

A standard is a shared starting point, not a museum exhibit.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A recurring task depends on whoever happens to remember how it was last done.

### Avoid when

- This adapts the idea of a shared, revisable method to office work; it is not a complete manufacturing standard. Do not standardize an unsafe or untested shortcut, or erase necessary professional judgment.

### Explanation

Record the current agreed method, its essential checks and the conditions it assumes. When those conditions or observed results change, compare the documented method with actual practice and revise it deliberately. A shared baseline makes the change visible; it does not make the method correct.

### Example

A reconciliation procedure records the actual keys and checks. When a source changes, the team updates the method instead of telling newcomers to ignore the document and ask one veteran.

### Check

Can someone perform and question the current method without relying on private folklore?

### Limits

- This adapts the idea of a shared, revisable method to office work; it is not a complete manufacturing standard. Do not standardize an unsafe or untested shortcut, or erase necessary professional judgment.

### Evidence and sources

- supports: Standardized work provides an explicit current method that can serve as a basis for improvement. — lei-standardized-work. Claim basis: professional_convention. Claim limit: A documented method must still be appropriate, current and open to correction.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Current method and basis for improvement"].
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Definition and improvement role)
- lei-standardized-work: Standardized Work — https://www.lean.org/lexicon-terms/standardized-work/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (use_before): https://vedokrok.com/knowledge/stop-the-defect-before-it-becomes-a-larger-batch

---

## Stop the defect before it becomes a larger batch

ID: MHC-D-RESEARCH-0179 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/stop-the-defect-before-it-becomes-a-larger-batch

High throughput is less impressive when it multiplies the same error.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A process continues producing output after an abnormal condition is detected.

### Avoid when

- Stopping a production or safety-critical system can itself cause harm. Use the system's authorized incident procedure, not an improvised universal stop rule.

### Explanation

High throughput is less impressive when it multiplies the same error.

### Steps

1. Define the abnormal conditions that require a safe stop or quarantine and who has authority to act.
2. Contain affected output, investigate and use an explicit restart condition rather than simply clearing the alarm.

### Example

A data import shows a broken mapping. Pausing the remaining load can limit damage while the team determines which records were affected.

### Check

Did the response prevent further propagation and preserve a controlled restart path?

### Limits

- Stopping a production or safety-critical system can itself cause harm. Use the system's authorized incident procedure, not an improvised universal stop rule.

### Evidence and sources

- supports: Jidoka includes detecting an abnormal condition and stopping the process before defects continue downstream. — lei-jidoka. Claim basis: professional_convention. Claim limit: Stop authority and safe restart requirements depend on the actual system.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Detect abnormality and stop propagation"].
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Definition)
- lei-jidoka: Jidoka — https://www.lean.org/lexicon-terms/jidoka/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z

---

## Give the warning a responder and a response

ID: MHC-D-RESEARCH-0180 · Version: 0.1.0 · Kind: template
Source: https://vedokrok.com/knowledge/give-the-warning-a-responder-and-a-response

A warning without a response path is coloured wallpaper.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A dashboard turns red but nobody knows what action follows.

### Avoid when

- Avoid alerts that expose sensitive information or overwhelm responders. Do not confuse an acknowledgement with resolution.

### Explanation

A warning without a response path is coloured wallpaper.

### Template

Signal: [observable condition]. First responder: [role]. Immediate safe action: [action]. Escalation or recovery condition: [rule].

### Example

A blocked review item names the missing decision and its owner. A generic red label saying urgent gives much less help.

### Check

Could the intended responder act from the signal without guessing its meaning?

### Limits

- Avoid alerts that expose sensitive information or overwhelm responders. Do not confuse an acknowledgement with resolution.

### Evidence and sources

- supports: An andon makes a process abnormality or need for assistance visible. — lei-andon. Claim basis: professional_convention. Claim limit: A visible warning is not itself a response or a resolved problem.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Visible abnormal condition and assistance signal"].
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Definition)
- lei-andon: Andon — https://www.lean.org/lexicon-terms/andon/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (useful_with): https://vedokrok.com/knowledge/stop-the-defect-before-it-becomes-a-larger-batch

---

## Change the design so the mistake has less room to happen

ID: MHC-D-RESEARCH-0181 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/change-the-design-so-the-mistake-has-less-room-to-happen

A better warning is not always the best prevention.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- The same predictable error keeps returning despite reminders to be careful.

### Avoid when

- Validation can reject legitimate cases or create unsafe workarounds. Test boundaries and do not claim one control makes the whole system error-proof.

### Explanation

A better warning is not always the best prevention.

### Steps

1. Identify the exact action or omission that creates the recurring error.
2. Prefer a suitable design constraint or immediate validation that prevents that error path, then test legitimate exceptions.

### Example

An import form validates the required identifier format before accepting a file, rather than relying on a note that operators routinely miss.

### Check

Does the change block the specific mistake while still allowing valid work?

### Limits

- Validation can reject legitimate cases or create unsafe workarounds. Test boundaries and do not claim one control makes the whole system error-proof.

### Evidence and sources

- supports: Error-proofing changes the process or design to prevent or detect a specific mistake. — lei-error-proofing. Claim basis: professional_convention. Claim limit: One protected error path does not make the whole process error-free.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Definition and prevention examples"].
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Definition and examples)
- lei-error-proofing: Error-Proofing — https://www.lean.org/lexicon-terms/error-proofing/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (alternative): https://vedokrok.com/knowledge/build-a-checklist-from-real-misses

---

## Distinguish the pile from the rate changing it

ID: MHC-D-RESEARCH-0183 · Version: 0.1.0 · Kind: concept
Source: https://vedokrok.com/knowledge/distinguish-the-pile-from-the-rate-changing-it

The pile remembers yesterday's difference between arrivals and departures.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A growing backlog is discussed only in terms of how hard people worked today.

### Avoid when

- A simple balance does not explain why the rates changed or whether all items have equal effort. Define the system boundary and avoid counting transfers twice.

### Explanation

A stock accumulates through inflows minus outflows. Separate the current amount from the rates changing it, and keep their units consistent before deciding what would reduce the accumulation.

### Example

A fictional queue begins with 40 items, receives 15 and completes 10. It ends with 45, despite ten genuine completions.

### Check

Can you reconcile the closing stock with its opening amount and flows?

### Limits

- A simple balance does not explain why the rates changed or whether all items have equal effort. Define the system boundary and avoid counting transfers twice.

### Evidence and sources

- supports: A stock changes through the difference between its inflows and outflows; delayed responses can destabilize repeated adjustments. — meadows-leverage-points. Claim basis: technical_definition. Claim limit: The examples are simple models, not calibrated predictions of a complex organization.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Stocks and flows; delays relative to rates of change"].
Source note: A classic systems essay, not new experimental evidence. No universal ranking of intervention effectiveness is inferred for the reader's situation.
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Stocks and flows; delays)
- meadows-leverage-points: Leverage Points: Places to Intervene in a System — https://donellameadows.org/archives/leverage-points-places-to-intervene-in-a-system/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (use_before): https://vedokrok.com/knowledge/allow-for-the-response-delay-before-adjusting-again

---

## Allow for the response delay before adjusting again

ID: MHC-D-RESEARCH-0184 · Version: 0.1.0 · Kind: principle
Source: https://vedokrok.com/knowledge/allow-for-the-response-delay-before-adjusting-again

Impatience can become part of the feedback loop.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- You keep changing a control before the previous change could have affected the result.

### Avoid when

- Waiting is not appropriate when a safety signal requires immediate action. Delays are uncertain and should be investigated, not invented to defend a failing intervention.

### Explanation

Identify the delay between an action and its observable effect. Set a suitable review point and monitor leading information where appropriate, rather than repeatedly escalating because an immediate outcome has not moved.

### Example

A workflow change needs several days for existing items to pass through. Reversing it every afternoon can make the observed pattern harder to interpret.

### Check

Has enough relevant process time passed to judge the previous adjustment?

### Limits

- Waiting is not appropriate when a safety signal requires immediate action. Delays are uncertain and should be investigated, not invented to defend a failing intervention.

### Evidence and sources

- supports: A stock changes through the difference between its inflows and outflows; delayed responses can destabilize repeated adjustments. — meadows-leverage-points. Claim basis: technical_definition. Claim limit: The examples are simple models, not calibrated predictions of a complex organization.
Source access: selected_html_passages; accessed on: 2026-09-17; inspected sections: ["Stocks and flows; delays relative to rates of change"].
Source note: A classic systems essay, not new experimental evidence. No universal ranking of intervention effectiveness is inferred for the reader's situation.
Agent release review: Checked the named definitions in the May 2025 Kanban Guide, LEI and Meadows. Non-manufacturing examples are editorial applications. The text preserves safe-stop, ownership and contextual boundaries. (Stocks and flows; delays)
- meadows-leverage-points: Leverage Points: Places to Intervene in a System — https://donellameadows.org/archives/leverage-points-places-to-intervene-in-a-system/

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z

---

## Keep user, buyer and approver as separate people

ID: MHC-D-RESEARCH-9772 · Version: 0.1.0 · Kind: checklist
Source: https://vedokrok.com/knowledge/keep-user-buyer-and-approver-as-separate-people

Interest and purchasing authority are different assets.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A professional likes your B2B tool but cannot buy it.

### Avoid when

- Buying routes vary between organizations; a stakeholder map is a hypothesis until confirmed.

### Explanation

Identify the daily user, budget owner, approver and blocker in the proposed workflow. Ask each relevant role what decision they control. Validate the route to a real purchase before treating one enthusiastic user as pipeline.

### Checklist

- Map the person who suffers the problem and the person who controls payment.
- Write the next approval required before adoption.

### Example

An HR specialist wants a voice assessment tool, but procurement and compliance determine access.

### Check

You know who can authorize use and which barrier remains.

### Limits

- Buying routes vary between organizations; a stakeholder map is a hypothesis until confirmed.

### Evidence and sources

- supports: The NSF I-Corps process investigates prospective customers and partners to test a business opportunity, not just to build a prototype. — biz-nsf. Claim limit: General US-based occupation, entrepreneurship or training information; adapt to current local market, employer and product circumstances. Does not establish an individual outcome.
Source reading history: selected_passages; 2026-10-10; Customer discovery and evidence-led commercialization. Official source page or reference excerpt read on 2026-10-10. Supports a narrow tool, definition or recommended practice; adapted reader workflows are editorial, not independently tested for effectiveness.
Agent release review: NSF I-Corps customer discovery provides context. Role mapping, payment-signal checks and stopping rules are bounded practical adaptations. Keep hypothetical interest, delivered paid work and evidence for scale distinct. (Customer discovery and evidence-led commercialization)
- biz-nsf: NSF I-Corps National Teams Applicants — https://www.nsf.gov/funding/initiatives/i-corps/national-teams-applicants

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (use_before): https://vedokrok.com/knowledge/test-a-payment-signal-before-building-the-feature

---

## Test a payment signal before building the feature

ID: MHC-D-RESEARCH-9774 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/test-a-payment-signal-before-building-the-feature

A smaller transaction is a stronger signal than a bigger opinion.

Knowledge source release: MHC-RPUB-20261112-20287590
Source corpus commit: 20287590ddd3999b7492df232822b6a0f7d114b1
Source manifest SHA-256: 2a987660d12de0ce2bde77a421424f7762af1ff77872ef41093fecc33db74a95
Source projection SHA-256: 2807da39bbe6a3dee2db32a76675ac5185f5a33e8b0a5cdd0c22a4731f3378a0

### Use when

- A roadmap contains features that nobody has asked to pay for.

### Avoid when

- Early revenue can be misleading; a handful of buyers does not prove scalable retention.

### Explanation

Offer a legitimate, clearly described pilot, pre-order or paid consultation only where you can deliver and honor terms. Compare expressed interest with completed sign-ups or purchases, and track why qualified people decline.

### Steps

1. Define the small real commitment and the exact delivery promise.
2. Record qualified approaches, actual commitments and refusal reasons.

### Example

For an AI onboarding tool, offer a supervised two-week manual pilot instead of coding an entire dashboard.

### Check

At least one observable purchase or binding commitment is recorded, or the absence is explicit.

### Limits

- Early revenue can be misleading; a handful of buyers does not prove scalable retention.

### Evidence and sources

- supports: The NSF I-Corps process investigates prospective customers and partners to test a business opportunity, not just to build a prototype. — biz-nsf. Claim limit: General US-based occupation, entrepreneurship or training information; adapt to current local market, employer and product circumstances. Does not establish an individual outcome.
Source reading history: selected_passages; 2026-10-10; Customer discovery and evidence-led commercialization. Official source page or reference excerpt read on 2026-10-10. Supports a narrow tool, definition or recommended practice; adapted reader workflows are editorial, not independently tested for effectiveness.
Agent release review: NSF I-Corps customer discovery provides context. Role mapping, payment-signal checks and stopping rules are bounded practical adaptations. Keep hypothetical interest, delivered paid work and evidence for scale distinct. (Customer discovery and evidence-led commercialization)
- biz-nsf: NSF I-Corps National Teams Applicants — https://www.nsf.gov/funding/initiatives/i-corps/national-teams-applicants

Review: agent · Agent editorial and scoped source review · 2026-11-12T18:37:40Z
Published: 2026-11-12T18:37:40Z
Related (use_before): https://vedokrok.com/knowledge/write-a-falsifiable-stop-rule-for-a-startup-experiment
