212212js/douyin-market-opportunity-skill

Douyin 抖音市场机会评审表

Create a new spreadsheet from this template. Keep the reference file unchanged.

Workflow

1. Read artifact-template.json and resolve its paths relative to this skill directory.
2. Confirm the current operating boundary before analysis. By default this workflow only evaluates the products the user has put in scope; do not introduce unrelated supply-chain products.
3. Collect or accept the fixed evidence pack below. Keep each source separate and label missing evidence rather than filling gaps with assumptions.
4. Load @spreadsheets and invoke its reference/template workflow with the retained file.
5. Clone or import the reference instead of replacing its visual system with generic defaults.
6. Update only the pages that are supported by the current inputs, render and verify the finished spreadsheet, then return the final artifact.

Source handoff and division of work

Use this default handoff unless the user explicitly changes it:

Source or task	Default owner	What is needed
抖音指数	Codex	The user supplies the live viewing URL/entry and leaves the required account logged in. Read the visible values, date range, search trend, demographic or related-word evidence, and keep a screenshot/source location for every number used.
27周榜单分析表	User	The user sends the newest file whenever it changes. Treat the newest copy as authoritative; do not reuse an old price-band or ranking conclusion if the table has changed.
市场分析判断	User	The user may send their own market judgment, meeting notes, or hypotheses. Preserve it as 我的初步想法 / 会上讨论, not as a proven fact.
差评与真实评价	User	The user supplies collected review excerpts, links, screenshots, or a review summary. Only use review claims that are visibly source-backed; write 无可验证差评 when no separable review evidence is supplied.
表格整合、证据校验、剪辑切口	Codex	Convert the above into a meeting-ready workbook, retaining sources and removing claims the current product cannot carry.

Fixed evidence pack for every keyword

For every future keyword, use the same four input packs. They are the minimum evidence for a complete market-opportunity review; do not replace them with generic AI conclusions.

Fixed input	Minimum fields to retain	Workbook destination	Decision it supports
关键词指数情况	Viewing location, date range, trend screenshot, visible search values, user portrait, related words/seasonality when available.	01运营：市场判断	Is there a real demand signal, who is searching, and when?
抖音爆款视频拆解	Original link, cover or original-video frame, title, play/engagement data, original first 3 seconds, structure, reusable action, and non-reusable claims.	06剪辑：案例与框架 and 04剪辑：框架参考	What opening, scene and proof action should be tested?
相关链接差评/好评	Link or source label, faithful user excerpt, positive/negative tag, shoe/use scene, pain or desired result.	02运营：真实痛点 and 05剪辑：素材库	What does the user truly tolerate, praise or reject?
关键词下链接竞争截图	Search/ranking screenshot, visible top links, price/offer form, sales or payment signal when available, and capture date.	01运营：市场判断	How crowded is the market, what form sells, and which price idea is only a meeting hypothesis?

Always collect the current product scope and any sample facts alongside the four packs. This is the boundary for 可直接测 / 样品验证后 / 不可承接; it is not a fifth market-evidence pack.

Required inputs for a run

Ask for only the missing items:

1. The Douyin Index viewing location (URL or a clear navigation path) and a logged-in browser session when the data is not public.
2. The latest keyword competition/ranking screenshot or 抖音榜单分析_27周.xlsx equivalent update.
3. The Feishu video table or original-video references with covers/links where available.
4. The related-link review collection, retaining separable good and bad review excerpts.
5. Current sellable product scope and any product facts already confirmed by sample, product page, or test.

When an input pack is missing, ask only for that pack and produce a clearly marked partial update instead of using made-up content.

Evidence and judgment rules

• Distinguish 数据事实 (index, ranking, product-page, review, screenshot) from 我的初步想法 and 会上待定.
• A price-band gap is not proof of opportunity. State it as a test hypothesis until demand, competition and product-fit evidence agree.
• A whole-shoe issue such as shoe mouth rubbing, outsole slip resistance, noise, size/last fit or height gain must not be repackaged as an insole or shoe-charm promise.
• A pain point can enter 04剪辑：框架参考 only if it can be shown in a video and the in-scope product can plausibly respond after real-product verification.
• Keep high-frequency but non-actionable or unserviceable problems in 07会后：用户不接受的问题 / red lines, not in the sellable creative list.
• Use direct review wording sparingly and faithfully. Do not invent buyer quotes, counts, testing results, or product functions.
• For any product claim such as soft, breathable, sweat-absorbing, non-slip, easy to install, or secure, mark it 需实物验证 until a product fact or test is provided.

Universal market presentation framework

This is the fixed display order for every future keyword. The product category can change; the decision order cannot. The first page must answer the meeting's core question: is this a testable market direction, and what evidence supports that judgment?

Display layer	It must answer	Required evidence	What it must not do
1. Market decision	Is the keyword/category worth entering, testing, pausing, or rejecting?	A concise conclusion plus the strongest index, ranking and competition evidence.	Do not begin with a long script, generic trend language, or a final SKU decision.
2. Demand profile	Who is looking, when, why, and in which use scenes?	Douyin Index trend, date range, user portrait, related words, seasonality where available.	Do not infer demand from a single high number without the time range and comparison context.
3. Market and competition	What sells now, at what price, in what product form, and how crowded is it?	Latest 27-week ranking table, relevant product/price screenshots, leading links or content examples.	Do not call an unoccupied price band a blue ocean by itself.
4. User voice	What is the buyer actually trying to solve, tolerate, praise or reject?	User-supplied, verifiable reviews, Q&A, screenshots or excerpts arranged by scene and frequency.	Do not turn seller copy or imagined reviews into user evidence.
5. Opportunity and product fit	Which user need can the current product genuinely carry?	A pain-point-to-product map, product facts/sample tests, and a clear 可直接测 / 样品验证后 / 不可承接 classification.	Do not borrow whole-product features or add unrelated products just to fill an opportunity.
6. Content validation	Which scene, opening and proof action should be tested first?	A target user, one pain point, product action, scene, original-video reference and a reference-only opening sample.	Do not present AI sample copy as final copy or promise unverified benefits.
7. Meeting decision	What needs the team to decide now, and what evidence still needs collecting?	Explicit 会上待定, owner, evidence gap and next validation action.	Do not pre-decide selection, SKU, price, link structure or bundle rules for the meeting.

Keyword-adaptive rule

• The retained workbook is a visual and decision-layout reference, not reusable market evidence.
• For a new keyword, replace every category-specific title, screenshot, user scene, price statement, product boundary and conclusion. Never carry forward 洞洞鞋 numbers, examples or product claims.
• Keep the page hierarchy and colors, but rename page titles when the new keyword or project needs clearer names.
• If a source layer is absent, leave a clearly marked evidence gap rather than copying an old category's content.

Workbook update map

Maintain the retained workbook's sheet logic unless the user explicitly asks to change it:

• 01运营：市场判断 — the market-decision, demand-profile and competition layers. Embed the corresponding index/ranking/product screenshots in this page.
• 02运营：真实痛点 — the user-voice layer: product/use form, scene, verified review pain and desired outcome. Do not pre-decide product capability here.
• 04剪辑：框架参考 — the product-fit and content-validation layers: actionable content cut points, target user/scene, faithful user wording, provisional product response and opening sample. Label all AI visuals/copy as reference only.
• 05剪辑：素材库 — preserve source material unless the user asks to edit it.
• 06剪辑：案例与框架 — original-video cover, opening/reference structure and what can or cannot be reused.
• 07会后：用户不接受的问题 — product red lines only, unless the user explicitly asks for a separate decision page. Product red lines are issues that could trigger rejection, complaints, returns or negative reviews; this is not a list of words that copywriters are forbidden to say.

Delivery checklist

1. State which evidence is newly read from Douyin Index and which inputs were supplied by the user.
2. Preserve evidence screenshots/cover images in the corresponding worksheet rather than leaving opaque labels like 图一.
3. Put every decision-relevant item into one of three buckets: 可直接测、样品验证后、不可承接/红线.
4. When a new 27-week table is supplied, state what changed from the prior version: demand, audience, price band, competition, ranking or the market conclusion.
5. Scan for unsupported whole-product claims and move or remove them before delivery.
6. For every video case, put the original cover/frame beside the written breakdown; if an attachment is absent, say so plainly rather than inserting a made-up visual.
7. Render every sheet, verify the final .xlsx opens and contains no formula errors, then deliver one meeting-ready workbook.

Fidelity

Preserve the visual hierarchy, formulas, dimensions, tables, charts, validation, conditional formatting, and frozen panes. For a new keyword, category-specific sheet names, titles and content must be replaced; do not preserve them merely because they exist in the retained reference.

User instructions control requested content and explicit deviations. The retained reference controls layout and formatting where the user has not requested a change.