CONTENT TEST LEARNING QUESTION WORKSHEET
========================================

Original tested checklist by Alfred.

Purpose
-------
Use this worksheet to design one interpretable short-form content comparison,
release one eligible candidate at a time, and keep production checks separate
from audience evidence. It is platform-neutral and must be adapted to current
first-party destination rules and metric definitions before a real release.

This is a blank decision aid, not evidence that a candidate is safe, uploaded,
public, observed, or successful. Completing it does not prove causality or make
an account, destination, metric, audience, or result exist.

Truthful evidence vocabulary
----------------------------
PASS         Identified evidence satisfies the declared check.
FAIL         Identified evidence contradicts the declared check.
BLOCKED      Required evidence is missing, stale, inaccessible, or conflicting.
NOT_TESTED   The check has not been performed.
SUPERSEDED   A changed candidate, question, destination, or rule invalidated it.
INSUFFICIENT The observation window closed without decision-grade evidence.
CONFLICTED   Material confounding or opposing signals prevent a narrow decision.
STOP         A truth, rights, privacy, safety, policy, or quality gate failed.

Do not turn BLOCKED, NOT_TESTED, SUPERSEDED, INSUFFICIENT, or CONFLICTED into
PASS. Do not use a local validation result as audience evidence.


A. TEST ENVELOPE
================

Worksheet identity:
Prepared at:
Prepared by role:
Editorial decision owner:
Decision due by:

Owned destination under review:
Destination ownership evidence:
Account or channel state:
Current destination rules checked at:
Rules or policy revision/reference:
Metric-definition reference:

Topic boundary:
Intended audience need:
Exact editorial decision this test may inform:
Decisions explicitly out of scope:

Assumptions:
-

Known limitations:
-

Stop conditions:
-

Test-envelope result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Evidence:
Reason:
Recheck trigger:


B. ONE LEARNING QUESTION
========================

Write one question that names one decision, one eligible population, one
intended difference, one primary observation, and one evidence window.

Learning question:

Editorial decision:
Eligible release population:
Single interpretable surface allowed to differ:
Controlled surfaces:
-

Primary observation:
Evidence window start rule:
Evidence window end rule:
Processing-delay policy:
Stop, hold, and supersession conditions:
-

Question review:
[ ] It asks about one editorial decision rather than which item is “best.”
[ ] It names one intended creative difference.
[ ] It does not vary a truth, rights, privacy, disclosure, accessibility,
    safety, or minimum-quality requirement.
[ ] Its primary observation is available from the named destination.
[ ] Its evidence window is declared before release.
[ ] It permits an insufficient or blocked result.
[ ] It does not imply universal audience preference or platform causality.

Learning-question result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Evidence:
Reason:


C. CANDIDATE IDENTITIES AND STATES
==================================

Never compare labels alone. Bind each candidate to immutable artifacts.

CANDIDATE A
Candidate ID:
Exact master filename:
Master digest and algorithm:
Script identity or digest:
Caption identity or digest:
Metadata identity or digest:
Visual-system revision:
Audio revision:
Duration:
Current state: LOCAL / PRIVATE / QUEUED / PUBLIC / PUBLICLY_VERIFIED
State evidence:
State observed at:

CANDIDATE B
Candidate ID:
Exact master filename:
Master digest and algorithm:
Script identity or digest:
Caption identity or digest:
Metadata identity or digest:
Visual-system revision:
Audio revision:
Duration:
Current state: LOCAL / PRIVATE / QUEUED / PUBLIC / PUBLICLY_VERIFIED
State evidence:
State observed at:

Identity review:
[ ] Every candidate has one exact master identity.
[ ] Scripts, captions, and metadata are bound to the intended master.
[ ] Local, private, queued, public, and publicly verified are distinct states.
[ ] A provider processing state is not treated as correct playback.
[ ] An obscure route is not treated as proof of private visibility.
[ ] No candidate is called observed before its evidence window completes.

Candidate-identity result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Reason:


D. RELEASE ELIGIBILITY — RUN FOR EACH CANDIDATE
===============================================

Candidate ID:
Eligibility review revision:
Reviewed at:
Reviewer role:

Truth and claim boundary
[ ] Hook, body, captions, and metadata express the same bounded claim.
[ ] No fabricated experience, customer, testimonial, metric, or result appears.
[ ] Uncertainty and limits are retained rather than edited into confidence.
Result: PASS / FAIL / BLOCKED / NOT_TESTED
Evidence:

Rights and provenance
[ ] Every media component is original or has a documented usable license.
[ ] Source, license, attribution, transformation, and retention notes exist.
[ ] No third-party watermark or copied interface is present.
Result: PASS / FAIL / BLOCKED / NOT_TESTED
Evidence:

Privacy and identity
[ ] No private information, identifying attribution, account data, contact
    detail, secret, credential, sensitive screenshot, or private path appears.
[ ] Metadata, filenames, frames, captions, audio, and attachments were checked.
Result: PASS / FAIL / BLOCKED / NOT_TESTED
Evidence:

Disclosure and accessibility
[ ] AI-assistant disclosure is accurate for the destination and context.
[ ] Captions are complete, synchronized, readable, and claim-equivalent.
[ ] Essential meaning is not conveyed only by color, audio, or rapid motion.
Result: PASS / FAIL / BLOCKED / NOT_TESTED
Evidence:

Picture, audio, and metadata
[ ] Picture was reviewed across the full provider-processed duration.
[ ] Speech, music, and effects were reviewed together at intended loudness.
[ ] Title, description, labels, thumbnail, and visibility setting are accurate.
[ ] Destination-specific policy and rights checks are current.
Result: PASS / FAIL / BLOCKED / NOT_TESTED
Evidence:

Overall eligibility: ELIGIBLE / INELIGIBLE / BLOCKED / NOT_TESTED / SUPERSEDED
Reason:
Unresolved items:
-

Safety rule: INELIGIBLE, BLOCKED, NOT_TESTED, or SUPERSEDED cannot be converted
into a test release merely to obtain a comparison.


E. INTENDED DIFFERENCE AND DEPENDENT EDITS
==========================================

Intended surface:
Variant A treatment:
Variant B treatment:
Why the difference is interpretable:

Necessary dependent edits:
- edit:
  candidate:
  reason:
  reviewer decision: ACCEPTABLE_PART_OF_TREATMENT / MATERIAL_CONFOUNDER / BLOCKED

Controlled-surface comparison
-----------------------------
Underlying topic and claim boundary: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Body explanation: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Call to action: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Factual sources: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Rights basis: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Visual system: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Audio intelligibility: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Caption completeness: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Duration: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Destination and visibility: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED
Release protocol: EQUIVALENT / DIFFERENT / BLOCKED / NOT_TESTED

Material confounders:
-

Difference review: PASS / FAIL / BLOCKED / NOT_TESTED / CONFLICTED
Reason:
Decision: PROCEED / REVISE / HOLD / STOP


F. RELEASE ORDER AND LADDER
===========================

First candidate selected:
Selection reason:
Later candidates held pending:
Maximum releases authorized by this record:
Minimum cadence or spacing rule:

For the selected candidate, record each transition separately.

1. LOCAL ELIGIBILITY
Candidate and digest:
Result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Evidence:

2. PRIVATE UPLOAD, IF THE DESTINATION OFFERS A VERIFIED PRIVATE STATE
Upload route or provider record:
Visibility control observed:
Result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Evidence:

3. PROVIDER-PROCESSED REVIEW
Processed picture:
Processed audio:
Processed captions:
Rights or policy state:
Metadata and visibility:
Result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Evidence:

4. PUBLIC RELEASE
Exact candidate digest released:
Release record:
Result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Evidence:

5. LOGGED-OUT PUBLIC VERIFICATION
Exact HTTPS route:
Status and served-media observation:
Visible disclosure:
Candidate identity comparison:
Observed at:
Result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Evidence:

6. OBSERVATION WINDOW
Window opened at:
Window closes at:
Provider processing allowance:
Result: OPEN / CLOSED / BLOCKED / NOT_STARTED / SUPERSEDED
Evidence:

Do not release a held later candidate merely because it already exists locally.


G. PREDECLARED OBSERVATION
==========================

Primary destination metric label exactly as displayed:
Available first-party definition:
Definition checked at:
Why this observation is closest to the learning question:

Unit of observation:
Population or eligibility rule:
Window start:
Window end:
Expected reporting delay:
Final, delayed, sampled, estimated, or unknown:

Decision threshold or comparison rule:
Why that rule is justified before observation:
Minimum evidence needed for any editorial inference:
Treatment of missing values: MISSING, never zero unless zero is observed
Treatment of sparse exposure:
Treatment of moderation or delivery anomalies:

Secondary observations, if any:
- label:
  purpose:
  must not replace the primary observation because:

Prohibited post-hoc substitutions:
-

Observation-plan result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Reason:


H. OBSERVATION RECORD
=====================

Complete only from authoritative destination evidence.

Candidate ID and released digest:
Verified public route:
Release observed at:
Evidence window:
Retrieval observed at:
Destination metric label:
Destination value: VALUE / MISSING / PROCESSING / UNAVAILABLE
Value and unit:
Definition or scope at retrieval:
Evidence record identity:

Processing anomalies:
-

Moderation or policy events:
-

Visibility or delivery anomalies:
-

Candidate or metadata changes after release:
-

Observation integrity:
[ ] Candidate identity matches the released master.
[ ] The route and visibility were verified rather than inferred from upload.
[ ] The declared window completed or remains accurately open.
[ ] Missing values remain missing.
[ ] Displayed counts from different destinations were not merged.
[ ] Followers are not represented as unique people across platforms.
[ ] No secondary metric replaced the predeclared primary metric after viewing.
[ ] Anomalies and definition changes are preserved.

Observation result: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Reason:


I. DECISION RECORD
==================

Choose exactly one narrow state:
[ ] ADOPT        Evidence supports trying the treatment for the declared decision.
[ ] RETAIN       Evidence supports keeping the current treatment.
[ ] INSUFFICIENT The completed evidence cannot distinguish the treatments.
[ ] CONFLICTED   Material confounding or opposing signals prevent a decision.
[ ] BLOCKED      Release, verification, or evidence retrieval did not complete.
[ ] SUPERSEDED   Candidate, destination, definition, or question changed.
[ ] STOP         A truth, rights, privacy, safety, policy, or quality issue failed.

Decision:
Decision owner:
Decision time:
Candidate and evidence identities used:
Reason:

Narrow action this decision permits:
Actions it does not permit:
Next recheck or release trigger:

Claim review:
[ ] The conclusion answers only the predeclared editorial question.
[ ] It does not claim universal preference, causality, conversion, or revenue.
[ ] It does not call a blocked release an underperforming release.
[ ] It does not call sparse, missing, or processing evidence a loss.
[ ] It does not generalize one destination's metric to another destination.
[ ] It preserves material confounders and dependent edits.
[ ] It does not weaken future eligibility gates.

Decision-record result: PASS / FAIL / BLOCKED / NOT_TESTED
Reason:


J. FAILURE-SHAPED REVIEW TESTS
==============================

Predeclare the expected state before exercising each relevant scenario. These
are review prompts, not claims that a destination or workflow passed.

1. Candidate digest changes after private review.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

2. Captions are complete for A but incomplete for B.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

3. The destination has no authoritative version of the intended metric.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

4. A is released on one platform and B on a different platform.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

5. Upload succeeds but logged-out public retrieval fails.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

6. The primary observation remains processing at the deadline.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

7. Exposure is too sparse for the predeclared decision rule.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

8. A dependent edit becomes a material confounder.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

9. A destination metric definition changes during the window.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

10. A rights, privacy, disclosure, accessibility, or safety gate fails.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

11. A later candidate is ready before the first observation window closes.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED

12. The displayed view count favors one candidate but the primary observation
is missing or conflicted.
Expected state and action:
Observed evidence:
Result: PASS / FAIL / BLOCKED / NOT_TESTED


K. COMPACT TWENTY-CHECK REVIEW
==============================

[ ] 1. The test names one exact editorial decision.
[ ] 2. One interpretable surface is allowed to vary.
[ ] 3. Controlled surfaces and dependent edits are explicit.
[ ] 4. Every candidate passed the same truth and quality gates.
[ ] 5. Every candidate has an immutable master identity.
[ ] 6. Local, private, queued, public, and verified states are distinct.
[ ] 7. One verified owned destination is used unless destination is the test.
[ ] 8. Cross-platform displayed metrics are not treated as interchangeable.
[ ] 9. The primary observation was chosen before release.
[ ] 10. The destination's exact metric label and definition are preserved.
[ ] 11. The evidence window and processing-delay policy are declared.
[ ] 12. Missing values remain missing rather than becoming zero.
[ ] 13. Stop, hold, and supersession conditions are named.
[ ] 14. One candidate releases first when later releases depend on its evidence.
[ ] 15. Provider-processed media is reviewed before public release.
[ ] 16. The exact public route and served media are verified after release.
[ ] 17. Moderation, delivery, visibility, and processing anomalies are recorded.
[ ] 18. INSUFFICIENT, CONFLICTED, BLOCKED, SUPERSEDED, and STOP remain available.
[ ] 19. The conclusion is no broader than the design and evidence.
[ ] 20. No safety, rights, truth, privacy, or quality rule became a variant.

Overall review: PASS / FAIL / BLOCKED / NOT_TESTED / SUPERSEDED
Reason:
Open items:
-
Owner:
Recheck trigger:


SOURCE, RIGHTS, AND CLAIM BOUNDARIES
===================================

This worksheet, its state vocabulary, learning-question record, eligibility
lanes, controlled-surface comparison, release ladder, observation record,
decision states, failure-shaped tests, and twenty-check review are original
text by Alfred. It contains no third-party media, customer
material, private account data, audience metric, or claimed publication result.

The worksheet is platform-neutral. Current first-party policy, metric,
disclosure, rights, privacy, visibility, processing, and account rules must be
checked before a real release. A completed worksheet can organize evidence; it
cannot prove that a creative difference caused an outcome, that unseen factors
were controlled, or that a result will repeat.