Research and adapt a video case
Use one copy per case to separate popularity evidence, production knowledge, and a testable English-language adaptation. The goal is an original, teachable production method.
This is an offline reference and printable worksheet. It does not save typed responses. Edit the companion CSV in your spreadsheet app or print this page. The CSV contains blank response fields and guidance, with no automatic scoring or formulas.
1 Record what was observed
Before looking for winners, define the sampling rule: platform, language, topic, format, publication interval, inclusion criteria, and exclusions. Log ordinary performers and counterexamples as well as standout items. Manual public research is enough for this worksheet.
Comparison rule: prefer the same creator/channel, format, topic, and matched observation age or interval. Explain unavoidable differences. A cumulative lifetime count is not a seven-day count. Do not pool platform metrics without checking their definitions. A positive baseline with matching definitions permits a relative comparison; an absent or zero baseline makes the ratio unavailable.
Claim to write: “At [timestamp], the original item showed [raw metric]. Within [defined sample and window], it was [supported comparison] relative to [baseline].” Do not imply a platform-wide ranking, organic reach, or a causal explanation from this alone.
2 Apply the two gates
These are prerequisites, not points. A higher priority score cannot compensate for either gate. Evaluate the exact intended claim and use, rather than an entire creator or topic.
| Gate | Pass | Hold or fail |
|---|---|---|
| Evidence | The stated popularity claim has traceable original evidence, a timestamp, a defined sample/window and a defensible baseline. | Hold if a needed field or comparison is missing. Fail if the claim is contradicted or cannot be verified. Study may continue, but label popularity unverified. |
| Rights for intended use | Each intended asset/use has a recorded ownership, license, or other established basis with scope and restrictions checked. | Hold if permission or scope is unresolved. Fail if the intended use conflicts with a known restriction. Use independently created replacements where appropriate. |
Researching or linking an original is different from distributing its footage, music, script, voice, likeness, or files. Record sources and terms for your new assets too. Check commercial use, derivatives, attribution, distribution, territory/term, and redistribution where relevant. Do not assume an AI output or publicly viewable work is automatically clear for commercial reuse. If the basis remains uncertain, keep it out of publication until resolved.
3 Separate knowledge from interpretation
| Record | What belongs here | Evidence required |
|---|---|---|
| Observed | What is visible, audible, or directly documented | Source URL, timestamp or shot reference |
| Creator disclosed | The production steps the creator actually stated | Creator source, time reference, short paraphrase and disclosure limits |
| Reconstructed | One plausible route to a similar abstract effect | Assumptions, alternative routes, untested steps and later pilot result |
| Appeal hypothesis | A possible explanation for attention or engagement | Alternatives such as topic, timing, audience, distribution and what would weaken the explanation |
| Transfer hypothesis | Why a new English-language original might work for a defined audience | Context changes, target-viewer review and a falsifiable test |
4 Prioritize with six dimensions
These 1–5 anchors are a subjective planning aid, created for this course. They have not been validated to predict views, virality, or sales. Record evidence and counterevidence for every rating; do not infer evidence from a high score.
Missing is not zero. Leave a score blank when it cannot be judged and record the missing input. Calculate an equal-weight total of 6–30 only if both relevant gates pass and all six scores are present. Otherwise leave the total blank and show coverage, such as 4 of 6. Compare cases within one review batch using the same rubric version. No universal pass mark is proposed.
1 Audience need
Is there a specific viewer and a useful reason to watch?
- No identifiable viewer or need beyond general attention.
- A viewer segment is named, but the viewing situation and need remain vague.
- One viewer segment, viewing situation, and desired outcome are stated clearly.
- The proposed need is supported by relevant audience comments, interviews, or owned-channel evidence, with sources recorded.
- Several independent relevant signals support the need, and conflicting signals or boundaries are documented.
Score: ____ / 5 Evidence and counterevidence: __________________________________
Reviewer confidence and reason: ______________________________
2 Hook and payoff
Does the opening make a clear promise that the original ending delivers?
- The opening promise and ending do not match, or the promise is misleading.
- There is a curiosity gap, but the viewer cannot tell what will be resolved.
- The opening establishes a concrete question and the ending visibly answers it.
- A storyboard review confirms that the visual change and payoff are legible, including with sound off.
- A small, documented target-viewer comprehension review confirms the intended promise and payoff; misunderstandings are recorded.
Score: ____ / 5 Evidence and counterevidence: __________________________________
Reviewer confidence and reason: ______________________________
3 Localization fit
Can the mechanism work in a specific English-language audience context?
- The concept depends on a reference that does not transfer to the target audience.
- Several cultural, language, or context dependencies remain unexplained.
- A coherent English rewrite and locally understandable substitute context are specified.
- A fluent reviewer has checked wording, visual cues, units, subtitles, and cultural assumptions.
- Target-viewer comprehension review supports the adaptation; the new script, setting, and visuals are independently created.
Score: ____ / 5 Evidence and counterevidence: __________________________________
Reviewer confidence and reason: ______________________________
4 Production control
Can an original version be made reliably with available resources?
- No workable production route or resource cap has been identified.
- A route exists, but several critical steps, inputs, or tool capabilities remain untested.
- The shot plan, available tools, and resource cap are documented; a full original pilot is still unverified.
- An original pilot was completed inside the recorded cap, with failure recovery and actual costs logged.
- A repeat run of the original sample was completed with comparable quality inside the cap; versions and repeatable steps are recorded.
Score: ____ / 5 Evidence and counterevidence: __________________________________
Reviewer confidence and reason: ______________________________
5 Teaching value
Does the case teach a reusable decision and production method?
- The case is a one-off spectacle with no transferable method identified.
- It suggests a tip, but the steps or decision criteria are incomplete.
- A reusable method, before-and-after example, and step list can be taught.
- The lesson includes tradeoffs, failure diagnosis, and an editable original practice artifact.
- A learner has repeated the method with their own original sample and documented changes, failures, and recovery.
Score: ____ / 5 Evidence and counterevidence: __________________________________
Reviewer confidence and reason: ______________________________
6 Product relevance
Does an optional learning product naturally solve the next viewer problem?
- The proposed offer is unrelated to the video or would require misleading claims.
- A product connection exists, but the next viewer problem is unclear.
- One next problem and a relevant optional worksheet, template, or lesson are specified.
- The free video is complete on its own, and the proposed product clearly helps execute the next step.
- An original prototype has been reviewed or used by relevant learners; feedback supports usefulness and its limits are recorded. This is not proof of purchases.
Score: ____ / 5 Evidence and counterevidence: __________________________________
Reviewer confidence and reason: ______________________________
Priority total: ____ / 30 Coverage: ____ / 6
Decision and tradeoff: __________________________________________________
Disagreement between reviewers: __________________________________________
5 Design the original adaptation
Transfer hypothesis: For [audience], changing [context] while preserving [mechanism] will improve [predeclared metric] relative to [own-channel comparison], because [reason]. We would weaken or reject this hypothesis if [observable outcome].
6 Plan the production and repeat run
| Shot and time | Purpose and inputs | Action and tool role | Output and acceptance check | Fallback |
|---|---|---|---|---|
| 1 _____ | ________________ | ________________ | ________________ | ____________ |
| 2 _____ | ________________ | ________________ | ________________ | ____________ |
| 3 _____ | ________________ | ________________ | ________________ | ____________ |
| 4 _____ | ________________ | ________________ | ________________ | ____________ |
Use tool roles before choosing a vendor: capture or generate, composite, edit, caption, sound, export. Record model/application version and date, original input assets, prompt/settings where relevant, seed where supported, project dimensions, export settings, and output filenames. A seed alone does not establish reproducibility. Verify current usage terms and capabilities when selecting a tool.
Quality checks: subject and motion continuity, the intended scale cue, readable on-screen text, subtitle timing, sound, export, and truthful presentation of generated/composited scenes. Repeat a key shot from your own documented sample. Record changes in tool versions and the difference between repeated output and the original pilot.
| Attempt | Failure observed | Suspected cause | Change and result | Time and cost |
|---|---|---|---|---|
| _____ | ________________ | ________________ | ________________ | ____________ |
| _____ | ________________ | ________________ | ________________ | ____________ |
Resource and rights record
Set a labor-time cap and incremental-cost cap before production. Record planned and actual research, generation retries, editing, review, storage and export. Keep currency explicit. Track existing-subscription allocation separately from new cash spending; this worksheet neither sets a price nor authorizes a purchase or subscription.
7 Test within your own channel
- Predeclare the question. Choose one primary outcome, a meaningful difference, the comparison baseline, the observation window, and a minimum usable sample rule before seeing results. Derive these from your channel's volume and the decision you need to make; this worksheet supplies no invented industry benchmarks.
- Make two original variants of one topic. Change one main opening variable. Keep channel, intended audience, duration, payoff, description, and any optional CTA as comparable as possible. Do not repost other people's material or repeatedly upload near-duplicates to chase a result.
- Choose a design honestly. Use a suitable platform-supported randomized comparison if available. Otherwise use a planned, balanced sequential comparison and record order and exposure differences. It gives directional evidence, not clean causal identification.
- Observe the same age/window. Preserve source exports or snapshots, timestamps, counts, denominator definitions and exclusions. Record time/day, recommendation traffic, audience mix and overlap, external sharing, events, offer changes, and platform changes.
- Keep the stopping rule. At the predeclared end, report an inconclusive result if data are insufficient. If continuing is justified, write a new plan before new observations. Stop earlier for rights, safety, or serious factual issues. Do not keep extending the window until one variant wins.
- Decide using the whole path. A stronger click rate with weaker retention is a tradeoff to investigate. View counts alone do not establish product demand. Keep an optional product test separate and only run it if the offer and publication have been authorized.
| Measure | Definition to record | Important limitation |
|---|---|---|
| CTR | Clicks attributable to eligible impressions ÷ those eligible impressions, in the same window | Use the platform's exact population; not applicable when the relevant surface has no comparable impression denominator |
| Retention at a set time | Viewers or sessions still watching at the predeclared moment ÷ eligible starts, under the platform's definition | Keep start, replay and cohort definitions consistent; do not mix raw counts from different populations |
| Completion | Completed views ÷ eligible starts where supported | Length and replays may change meaning; preserve platform definitions |
| Purchase conversion | Attributed completed purchases ÷ matching qualified product-page visits | Specify unique users or sessions, attribution method/window and refund status. Use not applicable if no approved offer is live |
For each variant, record raw numerators and denominators alongside rates. Missing denominators produce an unavailable rate, not zero. Report uncertainty or appropriate intervals when useful; zero observed purchases with known exposure does not prove zero demand. Do not label a result statistically established without an adequate design and analysis.
8 Worked concept for practice
Fictional exercise · No real source case · No popularity or sales data
A storm inside a paper cup
An original miniature scene uses an unexpected-scale reveal. A close-up suggests a storm over a tiny boat; a pullback reveals the cup, handmade paper props, and compositing setup. The closing before-and-after explains one practical technique. The familiar structure is a starting point for your own expression, not a request to imitate a specific creator.
| Illustrative timing | Original scene | Purpose |
|---|---|---|
| 0–3 seconds | Paper cup with a miniature cloud. A: “A storm fits inside this cup.” B: “How big is this storm?” | Two proposed opening wordings for the same visual and payoff |
| 3–9 seconds | Handmade paper boat in an original water animation | Hold the scale question without making a factual disaster claim |
| 9–15 seconds | Pull back to the tabletop set | Resolve the scale question |
| 15–22 seconds | Original plate beside the composite | Show one production choice |
| 22–26 seconds | Concise method recap; optional worksheet invitation | Deliver a complete free lesson and an optional next step |
Proposed route, not completed work: sketch → photograph or film your own cup and paper boat → make or properly license cloud/rain layers → composite masks and scale → edit sound and captions → review on a phone-sized display. Possible failure: unclear scale. First repair to test: add an original tabletop context cue to the reveal. Log whether it works; do not assume it will.
Transfer hypothesis: the scale question may be understandable without culturally specific references. A target-viewer comprehension review can test that assumption. The evidence gate for “popular case” remains on hold because this is fictional; it may be used as a labeled design exercise. Rights for any final pilot remain pending until actual assets are checked. No scores or performance outcomes have been invented.
9 Finish the learning record
Attach your evidence log, English script, original shot list, project/version record, failure log, cost record and test result. Explain one decision you changed and why. Preserve the distinction between “we observed,” “the creator disclosed,” “we reconstructed,” and “we hypothesize.” Revisit source freshness and rights before reuse.
Scope: manual research, original production and authorized organic testing. No reposting, automated scraping, spam, paid ads, external contact, pricing commitment or subscription commitment is included. No promise of virality or earnings is made.