Alexa for Shopping · ACO

24 Hacks — The AI-Era Amazon Playbook

ACO 24 Hacks 中英對照手冊
本手冊以 Amazon US 帳號為前提；Alexa、AMC、Brand Tailored Promotion 等功能以美國站最完整，其他站點陸續開通。工具名與平台名保留英文。
版本 2026-07-16　·　24 hacks　·　英上中下對照
1
HACK 1 OF 24

Make Alexa Rewrite Your Listing For You

讓 Alexa 幫你重寫 listing

THE PAYOFF ／ 為什麼值得做

STEP BY STEP ／ 逐步操作
  1. 01On your live PDP, ask Alexa 5 questions: What's this for? What do people like? What don't they like? What are they buying instead? Why over alternatives?在你上架中的 PDP 上，問 Alexa 5 個問題：這是拿來做什麼的？大家喜歡它哪裡？不喜歡哪裡？他們改買了什麼替代品？為什麼要選它而不選其他？
  2. 02Paste the 5 answers + your title, bullets, images, and A+ screenshots into ChatGPT. (GoFullPage Chrome Extension)把這 5 個答案，加上你的標題、五點描述、圖片、A+ 截圖，一起貼進 ChatGPT。（截整頁用 GoFullPage Chrome 外掛）
  3. 03Prompt: 'Optimize this listing for AI, SEO, and conversions. First, give me 7 more Alexa questions to ask. Then rewrite the title, bullets, A+, and gallery plan.'提示詞：「為 AI、SEO 與轉換率優化這個 listing。先給我 7 個可以再問 Alexa 的問題，然後重寫標題、五點、A+ 與圖庫規劃。」
  4. 04Run the new Alexa questions, paste answers back, get the rewritten listing.把新的 Alexa 問題再問一輪，把答案貼回去，拿到重寫好的 listing。
  5. 05Fix image #2 first — it helps conversion better than any A+ module.先修第 2 張圖——它對轉換率的幫助勝過任何 A+ 模組。
2
HACK 2 OF 24

The $2 CPC Cart-Abandoner Audience Nobody Runs

沒人在跑的 $2 CPC 購物車棄置受眾

THE PAYOFF ／ 為什麼值得做

AMC lets you retarget shoppers who added to cart but didn't buy. That's the lowest-hanging fruit on Amazon. One seller sees average CPCs drop from $8 to $2 in supplements with conversion spiking. All sellers have AMC access now. Use it.

AMC 讓你能對「加了購物車卻沒結帳」的買家再行銷，這是 Amazon 上最好摘的果子。一位業者在保健品類看到平均 CPC 從 $8 掉到 $2，轉換率同時飆升。現在所有賣家都有 AMC 權限了，用起來。

STEP BY STEP ／ 逐步操作
  1. 01Ads → Measurement & Reporting → AMC → Create Audience → Instructional Query → Cart Abandoners template.廣告後台 → Measurement & Reporting → AMC → Create Audience（建立受眾）→ Instructional Query → 選「Cart Abandoners（購物車棄置）」範本。
  2. 02Look-back: 7 days. Auto-adjust date: ON. Name it [ASIN]_Cart_7d. Paste your ASINs. Build (takes ~15 min).回溯期設 7 天，Auto-adjust date（自動調整日期）開啟，命名為 [ASIN]_Cart_7d，貼上你的 ASIN，按 Build（約 15 分鐘完成）。
  3. 03Create a Brand Tailored Promotion that delivers a coupon when the cart-abandoner returns.建立一個 Brand Tailored Promotion（品牌訂製促銷），在棄置買家回訪時發給他優惠券。
  4. 04Build 2 campaigns: Sponsored Display retargeting + a Sponsored Products exact-match on your top keyword.開 2 個廣告活動：Sponsored Display 再行銷，加上一個鎖你主力關鍵字的 Sponsored Products 完全比對。
  5. 05Bid rule: keyword bid −40% below suggested, audience modifier +200%, ALL placements 0%. Forces spend to the audience.出價規則：關鍵字出價比建議值低 40%、受眾加成 +200%、所有版位 0%——把預算逼到這群受眾身上。
  6. 06Repeat the flow with 'Repeat Buyer not on Subscribe & Save' (attach S&S coupon) and 'Competitor Brand Searcher' (paste competitor brands).同一套流程再跑兩組：「回購但未加入 Subscribe & Save 的買家」（附 S&S 優惠券），以及「搜尋競品品牌的人」（貼上競品品牌名）。
3
HACK 3 OF 24

Build AEO-Optimized Landing Pages with AI (No Shopify Needed)

用 AI 打造 AEO 優化的 landing page（不用 Shopify）

AEO＝Answer Engine Optimization，為「AI 問答引擎」而做的頁面優化

THE PAYOFF ／ 為什麼值得做

Use Claude (or ChatGPT/Codex) to build standalone HTML landing pages with full schema markup — no designer, no developer, no Shopify store. The $20 Claude plan is enough. AI crawlers can read them, and you show up in AI answers within 1–2 weeks.

用 Claude（或 ChatGPT/Codex）做出帶完整 schema 標記的獨立 HTML landing page——不用設計師、不用工程師、不用 Shopify。$20 的 Claude 方案就夠。AI 爬蟲讀得到，1–2 週內就能在 AI 的回答裡曝光。

STEP BY STEP ／ 逐步操作
  1. 01Feed Claude 5–10 landing pages whose style you like, plus your Amazon listing link, logo, brand colors, and images. Talk in plain English and iterate the same way ("make the buttons orange instead of red").餵給 Claude 5–10 個你喜歡風格的 landing page，加上你的 Amazon listing 連結、logo、品牌色與圖片。用白話跟它講，改也一樣用白話（「把按鈕從紅色改成橘色」）。
  2. 02Rule it: no JavaScript to display content (JS only for CSS/headers) so AI crawlers can read the page. Add schema/JSON-LD in the header — copy the format from amzmarketer.com/etta (view page source), then tell Claude to replicate it.下規則：內容不准用 JavaScript 呈現（JS 只用於 CSS/header），這樣 AI 爬蟲才讀得到。在 header 加 schema/JSON-LD——參考 amzmarketer.com/etta（看網頁原始碼），叫 Claude 照抄格式。
  3. 03Structure: first 70 words carry your key answer/claim (AI reads the top and bottom, skips the middle). Add 5–15 objection-killing FAQs. One product, one goal, one CTA per page — never mix products.結構：前 70 字放你最關鍵的答案/主張（AI 讀頭尾、跳過中間）。加 5–15 條「打消疑慮型」FAQ。一頁只放一個產品、一個目標、一個 CTA——絕不混產品。
  4. 04Name the HTML file after the product/solution (e.g. /magnesium-balm-for-runners.html). Multiple angles = multiple pages — same product, different avatar (runners vs. menopausal women). Five products = five pages minimum.HTML 檔名用產品/解方命名（例：/magnesium-balm-for-runners.html）。多切角＝多頁——同一產品、不同受眾（跑者 vs. 更年期女性）各一頁。五個產品至少五頁。
  5. 05Host it cheap: Cloudflare domain ~$10/yr, central hosting on Bluehost/HostGator, or Unbounce if you want done-for-you. One domain can hold all your pages (separate keyword-match domains rank slightly better but are extra work).便宜上架：Cloudflare 網域約 $10/年、集中主機用 Bluehost/HostGator，懶得弄就用 Unbounce（全代管）。一個網域可放全部頁面（用關鍵字對應的獨立網域排名略好，但更費工）。
  6. 06Get indexed by AI: submit your URL directly to Google and Bing (search "submit website to Google"). ChatGPT and Claude piggyback on those crawlers — expect AI visibility in 1–2 weeks. (No manual submit form for ChatGPT; it's product feeds only.)讓 AI 收錄：把網址直接提交給 Google 和 Bing（搜「submit website to Google」）。ChatGPT 和 Claude 是搭這兩家爬蟲的順風車——1–2 週內就會被看見。（ChatGPT 沒有手動提交表單，只吃 product feed。）
4
HACK 4 OF 24

Make Your Website AI-Crawlable In One Click

一鍵讓你的網站能被 AI 爬取

THE PAYOFF ／ 為什麼值得做

Perplexity, ChatGPT, and Alexa crawl llms.txt, schema, and FAQ files, and almost no Amazon brand has them. One seller's GPT generates the full bundle from a single ASIN. FTP it. You're indexed.

Perplexity、ChatGPT 和 Alexa 會爬 llms.txt、schema 與 FAQ 檔，而幾乎沒有 Amazon 品牌準備這些。一位業者的 GPT 只要一個 ASIN 就能生出整包，FTP 上傳，你就被收錄了。

已更正：標準檔名為 llms.txt（雙 L，2024 由 Answer.AI 的 Jeremy Howard 提出，MS/OpenAI 爬蟲已抓取）；原投影片誤植 LMS.txt，此處已統一。

STEP BY STEP ／ 逐步操作
  1. 01Scan QR code below and open Leo's 'GEO Bundle Builder for E-commerce Products' custom GPT in ChatGPT.掃下方 QR code，在 ChatGPT 打開 這個自訂 GPT「GEO Bundle Builder for E-commerce Products」。
  2. 02Paste your Amazon product URL (or multiple ASINs at once for a multi-product bundle).貼上你的 Amazon 產品網址（要做多產品包就一次貼多個 ASIN）。
  3. 03Paste your DTC site URL when prompted.系統詢問時，貼上你的 DTC 官網網址。
  4. 04Download the zip — it includes index.html, llms.txt, product.json schema, faq.html, product-specs.pdf, sitemap.xml.下載 zip——裡面有 index.html、llms.txt、product.json schema、faq.html、product-specs.pdf、sitemap.xml。
  5. 05Open your FTP client (or Shopify file uploader) and drag every file into your site root.打開 FTP 工具（或 Shopify 檔案上傳器），把每個檔案拖進你的網站根目錄。
  6. 06For multi-product catalogs, also drop category LLMs.txt files (e.g., /supplements/llms.txt) to multiply citation surface area.若是多產品目錄，再放上分類的 llms.txt（例：/supplements/llms.txt），擴大被引用的覆蓋面。

佐證：步驟與 PAYOFF 現皆統一為 llms.txt（雙 L）。

5
HACK 5 OF 24

Replace Your PDF Lead Magnet With A Brand App

把 PDF 名單磁鐵換成一支品牌小工具（brand app）

THE PAYOFF ／ 為什麼值得做

20-page PDFs are dead. They're homework, not gifts. A 'brand app' (a single-page tool that solves one tiny problem in under 60 seconds) creates a real surprise-and-delight moment. One seller built one in an afternoon.

20 頁的 PDF 已經沒用了——那是作業，不是禮物。「brand app」（一頁式小工具，60 秒內解決一個小問題）才能製造真正的驚喜時刻。一位業者一個下午就做出一支。

STEP BY STEP ／ 逐步操作
  1. 01In Claude or ChatGPT: 'Act as a senior PM. Interview me about my brand to write a PRD for a single-page app I'll build in Replit. Must solve one specific problem in under 60 seconds.' Save as PRD.md.在 Claude 或 ChatGPT 輸入：「扮演資深 PM，訪談我的品牌，幫我寫一份 PRD，用來做一支我要在 Replit 上建的一頁式 app，必須在 60 秒內解決一個具體問題。」存成 PRD.md。
  2. 02Screenshot your homepage. Prompt: 'Generate a brand-guidelines.md — colors, fonts, tone, voice.' Save both files.截你的首頁。提示詞：「產生一份 brand-guidelines.md——顏色、字體、語氣、聲音。」兩個檔都存下來。
  3. 03Open Replit → new app → drag in PRD.md + brand-guidelines.md → prompt: 'Build this app per the PRD using these brand guidelines. Use Claude for any generated text.'打開 Replit → 新 app → 把 PRD.md 和 brand-guidelines.md 拖進去 → 提示詞：「依這份 PRD、套用這份品牌指南做出這支 app，任何要生成的文字都用 Claude。」
  4. 04Refine in plain English ('make the button our exact blue,' 'add confetti when it finishes,' 'gate the result behind email').用白話持續微調（「把按鈕改成我們品牌那個藍」「完成時放彩帶」「結果要留 email 才給看」）。
  5. 05In Shopify, create a new page with an iframe pointing to your Replit URL — your app lives at yourbrand.com/[appname].在 Shopify 建一個新頁面，用 iframe 指向你的 Replit URL——你的 app 就掛在 yourbrand.com/[appname]。
  6. 06Email capture goes BEFORE the result is revealed. Earned email list, no coupon scraping.email 收集要放在「揭曉結果之前」——這樣拿到的是真名單，不是來刮優惠券的。
6
HACK 6 OF 24

The Blur Test That Picks Your Winning Main Image

用「模糊測試」挑出會贏的主圖

THE PAYOFF ／ 為什麼值得做

Sellers post 1–2 main image variants and pray. One seller generates 10–12, runs them through SERP, then blurs the whole page. Whichever still pops at low detail wins the click. Lifted one client 23% in 3 weeks; another added $12K in 3 weeks.

賣家通常只放 1–2 張主圖然後求神拜佛。一位業者一次生 10–12 張，擺進 SERP，再把整頁模糊化——在低細節下還能跳出來的那張，就是能贏得點擊的。一個客戶 3 週提升 23%，另一個 3 週多賺 $12K。

STEP BY STEP ／ 逐步操作
  1. 01Generate 12 main image variations (Nano Banana, ChatGPT 5.5 or ListingOptimization.ai). Arrange as a 3×4 collage.生 12 張主圖變體（Nano Banana、ChatGPT 5.5 或 ListingOptimization.ai），排成 3×4 拼貼。
  2. 02Mark the 1–2 standout elements (hang tag, splash, prop). Build 3 fresh composites using only those elements.標出最跳眼的 1–2 個元素（吊牌、潑濺特效、道具），只用這些元素重組出 3 張新圖。
  3. 03Drop the 3 finalists into a SERP simulator (PickFu SERP, ProductPinion, ListingOptimization.ai, or a Photoshop mock) alongside your top 8 organic competitors.把 3 張決賽圖放進 SERP 模擬器（PickFu SERP、ProductPinion、ListingOptimization.ai，或用 Photoshop 手拼），跟你前 8 名自然排名對手擺在一起。
  4. 04Apply a heavy Gaussian blur across the whole simulated SERP so text disappears, only shape and contrast remain.對整個模擬 SERP 套上重度高斯模糊，讓文字消失、只剩形狀與對比。
  5. 05The image that still grabs your eye = winner. Shuffle competitor positions and re-blur to rule out position bias.在模糊下還能抓住你視線的那張＝贏家。把對手位置洗牌再模糊一次，排除位置偏誤。
  6. 06Test top 2 in PickFu for qualitative feedback. Graduate to Manage Your Experiments for live A/B.把前 2 名丟 PickFu 收質化回饋，再進 Amazon 的 Manage Your Experiments 做實站 A/B。
7
HACK 7 OF 24

A $0.40-Per-Second Veo 3.1 Product Video

每秒 $0.40 的 Veo 3.1 產品影片

THE PAYOFF ／ 為什麼值得做

A $300–$500 3D product video used to take 3–7 days. One seller ships a Sponsored Brand–compliant 3D video in under an hour for ~$2.40. His launch test hit 0.87% CTR — strong for SBV.

一支要價 $300–$500 的 3D 產品影片以前得做 3–7 天。一位業者現在 1 小時內、約 $2.40 就能產出一支符合 Sponsored Brands 規範的 3D 影片。他的上市測試打到 0.87% CTR——以 SBV（Sponsored Brands Video）來說很強。

STEP BY STEP ／ 逐步操作
  1. 01Pull your white-background Amazon main image as the anchor.拿你 Amazon 白底主圖當錨定圖（anchor）。
  2. 02In Gemini: 'Write a Veo 3.1 prompt for a 6-second 3D hero shot using the [sugar-sacrifice] framework. Tech specs: 1080p min, 16:9, 6–45s, no Amazon logo, no end card, brand in first 3 sec, audio at -14 LUFS.'在 Gemini 輸入：「用 [sugar-sacrifice] 框架，寫一段 Veo 3.1 提示詞，做一支 6 秒的 3D 主視覺鏡頭。技術規格：最低 1080p、16:9、長度 6–45 秒、不放 Amazon logo、不放結尾卡、品牌要在前 3 秒出現、音量 -14 LUFS。」
  3. 03aistudio.google.com → Veo Studio → Product Shot → upload anchor image, paste prompt, choose Veo 3.1, generate.aistudio.google.com → Veo Studio → Product Shot → 上傳錨定圖、貼提示詞、選 Veo 3.1，生成。
  4. 04Download the 6-sec clip.下載這段 6 秒影片。
  5. 05In Canva, add a 2-sec intro (Nano Banana: 'epic cinematic wide shot of [product] in deep space, 8K') + 2-sec outro with your logo.在 Canva 加 2 秒開場（Nano Banana 提示詞：「epic cinematic wide shot of [product] in deep space, 8K」）＋ 2 秒帶你 logo 的結尾。
  6. 06Upload to Amazon as a Sponsored Brand Video. Total cost: ~$2.40. Total time after first run: ~30 minutes.上傳到 Amazon 當 Sponsored Brands Video。總成本約 $2.40，熟練後總耗時約 30 分鐘。
8
HACK 8 OF 24

Get Backlinks From Spotify, Apple, and Google Podcasts

從 Spotify、Apple、Google Podcasts 拿反向連結

THE PAYOFF ／ 為什麼值得做

Spotify, Apple, and Google Podcasts carry sky-high domain authority and link straight back to your ASIN. One seller's team builds podcasts from scientific papers in Notebook LM, uploads to all three, and ranks organically in 2–3 weeks. 20 min per podcast.

Spotify、Apple、Google Podcasts 的網域權重極高，而且能直接連回你的 ASIN。一位業者的團隊用 NotebookLM 把科學論文做成 podcast，一次上架三家，2–3 週內自然排名。每支 podcast 只花 20 分鐘。

STEP BY STEP ／ 逐步操作
  1. 01Find 3–5 PubMed papers, journal articles, or news stories in your category.找 3–5 篇你品類相關的 PubMed 論文、期刊文章或新聞。
  2. 02Notebook LM (free) → new notebook → upload all PDFs/URLs as sources.NotebookLM（免費）→ 新 notebook → 把所有 PDF/網址當來源上傳。
  3. 03Click Generate → Audio Overview (Deep Dive) → 10-min two-host podcast auto-built.按 Generate → Audio Overview（Deep Dive）→ 自動生出一支 10 分鐘的雙主持人 podcast。
  4. 04Download the MP3. In show notes, paste a short summary plus your Amazon listing or DTC URL as the 'more info' link.下載 MP3。在 show notes 貼一段簡短摘要，並把你的 Amazon listing 或 DTC 網址放成「more info」連結。
  5. 05Upload to Spotify for Podcasters, Apple Podcasts Connect and submit RSS to Google Podcasts.上傳到 Spotify for Podcasters、Apple Podcasts Connect，並把 RSS 提交給 Google Podcasts。
  6. 06Repeat 20–30 times with different source papers. Within 2–3 weeks the backlinks index.換不同來源論文重複做 20–30 次。2–3 週內這些反向連結就會被收錄。
  7. Bonusgenerate the long-form video, chop into 15–30s clips with FFmpeg, post to TikTok / Meta Reels / Youtube Shorts.順便生出長版影片，用 FFmpeg 切成 15–30 秒短片，發到 TikTok / Meta Reels / YouTube Shorts。
9
HACK 9 OF 24

The 30-Second Meta Setting That Stops You Paying For Existing Customers

一個 30 秒的 Meta 設定，讓你不再為既有客戶付冤枉錢

THE PAYOFF ／ 為什麼值得做

Meta will let you tell its algorithm who you already own. If you don't claim them in Account Settings, you keep paying premium CPMs to re-find your own buyers. One seller has 'yet to see anyone do it correctly.' 30 seconds. Real CPM drop.

Meta 讓你能主動告訴它的演算法「哪些人已經是你的客戶」。如果你沒在 Account Settings 裡標記他們，你就會一直付高價 CPM 去重新找回自己的買家。一位業者說她「還沒看過有人設定對」。30 秒的事，CPM 實打實下降。

STEP BY STEP ／ 逐步操作
  1. 01business.facebook.com → menu → Ads Manager → Advertising Settings.business.facebook.com → 選單 → Ads Manager → Advertising Settings。
  2. 02Engaged Audiences → Create Custom Audience → Source: Meta Pixel → All Events → set retention to 180 days (default is 30). Save.Engaged Audiences → Create Custom Audience → 來源選 Meta Pixel → All Events → 保留天數設 180 天（預設是 30），存檔。
  3. 03Repeat for Lead Forms — change retention from 30 → 90 days (Meta's max).對 Lead Forms 重複一次——保留天數從 30 改成 90 天（Meta 上限）。
  4. 04Existing Customers → Create Audience → Customer List → upload a CSV of ALL paying customers.Existing Customers → Create Audience → Customer List → 上傳「所有付費客戶」的 CSV。
  5. 05Upload two more lists: high-value (subscribers, repeat buyers) and low-value (users who never converted in 400 days).再上傳兩份名單：高價值（訂閱者、回購者）與低價值（400 天內從未轉換的人）。
  6. 06Add a Value Rule: 'pay up to 200% more for high-value, 50% less for low-value.' Meta optimizes auction prices in your favor instead of guessing.加一條 Value Rule：「高價值最多多付 200%、低價值少付 50%。」讓 Meta 依此優化競價，而不是靠瞎猜。
10
HACK 10 OF 24

The SQP Filter That Finds A Product Idea In 10 Minutes

10 分鐘用 SQP 篩出一個產品點子

THE PAYOFF ／ 為什麼值得做

Use Amazon's own Brand Analytics + Search Query Performance to find keywords where shoppers click but conversion is broken. Demand exists but nobody has the right product. One seller found 'coffin makeup shelf' this way. Another seller found 'meat resting blanket' that's now $600K/month.

用 Amazon 自家的 Brand Analytics + Search Query Performance（SQP），找出「買家有點擊、但轉換卻壞掉」的關鍵字——代表需求存在、卻沒人有對的產品。一位業者就是這樣挖到「coffin makeup shelf（棺材造型化妝架）」。另一個賣家挖到「meat resting blanket」，現在月營收 $600K。

STEP BY STEP ／ 逐步操作
  1. 01Open Helium 10 Cerebro or Amazon Product Opportunity Explorer (free in Seller Central).打開 Helium 10 Cerebro，或 Amazon Product Opportunity Explorer（Seller Central 內免費）。
  2. 02Apply these filters: search volume >5,000, click share >50%, conversion share <10%, title density <5.套這些篩選：搜尋量 >5,000、click share >50%、conversion share <10%、title density <5。
  3. 03In the top 10 products, at least 5 must have <150 reviews AND <$5,000/month. This proves weak incumbents.前 10 名產品裡，至少 5 個要「評論 <150 且 月營收 <$5,000」——這證明現有玩家很弱。
  4. 04ChatGPT prompt: 'Read these product reviews and Q&A. What attribute is everyone complaining is missing or wrong?' That gap is your angle. (hint: Helium 10 is rolling out an MCP that can do this fast)ChatGPT 提示詞：「讀這些產品評論和 Q&A，大家一致抱怨缺了或做錯了哪個屬性？」那個缺口就是你的切角。（提示：Helium 10 正在推一個能快速做這件事的 MCP）
  5. 05Cross-check Etsy and Pinterest for the same keyword. A different material or style trending there but absent on Amazon = your launch.拿同一個關鍵字去 Etsy 和 Pinterest 交叉比對。在那邊正夯、Amazon 卻沒有的不同材質或風格＝你要上的品。
  6. 06Source the differentiated variant. Index inside a week.把這個差異化版本找工廠備貨，一週內完成上架收錄。
11
HACK 11 OF 24

Get An Article Cited By Every LLM In 24 Hours

24 小時內讓一篇文章被每個 LLM 引用

THE PAYOFF ／ 為什麼值得做

One seller took a dead article ranking nowhere and got it indexed across every LLM and ranked for dozens of keywords in 24 hours. Press releases are still the fastest white-hat backlink path, and AEO formatting makes the content citeable.

一位業者把一篇排名掛零的死文章，在 24 小時內做到被每個 LLM 收錄、並在數十個關鍵字排名。新聞稿（press release）至今仍是最快的白帽反向連結路徑，而 AEO 格式讓內容變得「可被引用」。

STEP BY STEP ／ 逐步操作
  1. 01Rewrite the first 60 words to contain a one-sentence direct topic answer. LLMs read beginning + end, not the middle.重寫前 60 字，塞進一句直接回答主題的話。LLM 讀頭尾、不讀中間。
  2. 02Add an FAQ section at the bottom. This is where LLMs grab citations.在文末加一段 FAQ——這就是 LLM 抓引用的地方。
  3. 03Strip all hedging ('we're the best at...'). Replace every claim with cited data: '[Study by X] found 90% reduction.' LLMs reject unsupported claims.砍掉所有模糊話術（「我們最擅長…」）。每個主張都換成有出處的數據：「[X 的研究]發現降低 90%。」LLM 會排斥沒有佐證的主張。
  4. 04Push the article via a PR distribution service (pressreleaseai.com, eReleases, EIN, Newswire) for $200–$400. Generates ~500 white-hat backlinks within hours class='zh'>用新聞稿發布服務（pressreleaseai.com、eReleases、EIN、Newswire）推出去，$200–$400，幾小時內生出約 500 條白帽反向連結 class='step'>05Syndicate to Medium, Claude Artifacts (make public), and a public Grok chat. All are high-DA, indexed in 24–48 hrs.再轉貼到 Medium、Claude Artifacts（設為公開）與一個公開的 Grok chat——全都是高網域權重、24–48 小時內被收錄。
  5. 06Submit the press-release URL to Google Search Console for force-indexing. Most rankings appear within a day.把新聞稿網址提交到 Google Search Console 強制收錄。大部分排名一天內就會出現。
12
HACK 12 OF 24

The Sponsored Brand Video Combo Nobody Is Running

沒人在跑的 Sponsored Brands Video 組合技

THE PAYOFF ／ 為什麼值得做

One seller spends $5M/month on Amazon ads. SB Video targeting competitor ASINs places your video right above the reviews on competitor pages. Most sellers don't realize the first video frame is the thumbnail. Took one product 0 → $200K/month.

一位業者每月花 $5M 在 Amazon 廣告上。用 SB Video 鎖定競品 ASIN，能把你的影片直接放在對手頁面「評論的正上方」。多數賣家不知道：影片的第一格畫面就是縮圖。這招把一個產品從 0 做到月營收 $200K。

STEP BY STEP ／ 逐步操作
  1. 01Create a new Sponsored Brand Video campaign in the Ads Console.在廣告後台開一個新的 Sponsored Brands Video 活動。
  2. 02Targeting type: Product targeting (NOT keywords). Paste in 20–50 competitor ASINs ranking for your money keywords.鎖定類型選 Product targeting（不是關鍵字）。貼進 20–50 個在你主力關鍵字有排名的競品 ASIN。
  3. 03Pick videos where the FIRST FRAME sells the product on its own. It's the static thumbnail when autoplay doesn't trigger.選那種「第一格畫面本身就能賣產品」的影片——沒觸發自動播放時，那格就是靜態縮圖。
  4. 04Run 3–5 different videos as A/B variations against the same ASINs.針對同一批 ASIN，用 3–5 支不同影片跑 A/B。
  5. 05Add the Audience Bid Modifier: 'high likelihood to purchase based on recent shopping activity' → +20% to start.加上 Audience Bid Modifier：「依近期購物行為判斷、高購買可能性」→ 起步 +20%。
  6. 06Ratchet to +30–40% if it pulls. Accounts running this run roughly half the ACoS of accounts that don't.有拉力就往上調到 +30–40%。有跑這招的帳號，ACoS 大約是沒跑的一半。
13
HACK 13 OF 24

The Free Tool That Hands You Your FAQs And Angles

直接餵你 FAQ 與頁面切角的免費工具

原 hack 13（借 1P vendor＋eCapital 美國信用額度）因涉及美國制度、對台灣受眾不適用，已替換為此通用版

THE PAYOFF ／ 為什麼值得做

AnswerThePublic (One seller's tool) shows exactly what people type into AI models, Google, social, and shopping — pure gold for FAQ questions and page angles. The free tier gives ~2 runs. Your best commercial-intent question becomes your page title and first-70-words answer.

AnswerThePublic（一位業者的工具）能精準呈現人們在 AI 模型、Google、社群與購物平台實際輸入的字——挖 FAQ 問題與頁面切角的金礦。免費版約 2 次查詢。你最好的「商業意圖」問題，就拿去當頁面標題與前 70 字的答案。

STEP BY STEP ／ 逐步操作
  1. 01Go to AnswerThePublic.com. Enter your URL (best) or your main keyword. Free tier ≈ 2 runs, so pick carefully.到 AnswerThePublic.com，輸入你的網址（最準）或主要關鍵字。免費版約 2 次，慎選。
  2. 02It maps what people type into AI models, Google, social, and shopping. Screenshot or export the full report.它會攤開人們在 AI 模型、Google、社群與購物端輸入的字。把整份報告截圖或匯出。
  3. 03Focus ONLY on the 'C' (commercial-intent) flagged queries. Skip the purely informational ones.只鎖定標記「C」（商業意圖）的查詢，跳過純資訊型的。
  4. 04Your single best AI-visibility question becomes your page TITLE and your first-70-words answer (ties to hack 3).把最有 AI 曝光潛力的那一題，拿去當頁面「標題」與前 70 字的答案（接 hack 3）。
  5. 05Feed the screenshot into ChatGPT/Claude: 'Turn these into 10 FAQ Q&As and 3 landing-page angles.'把截圖丟給 ChatGPT/Claude：「把這些變成 10 組 FAQ 問答與 3 個 landing page 切角。」
  6. 06Drop the FAQs into your PDP Q&A, A+, and your llms.txt/FAQ file (ties to hacks 4 & 21).把 FAQ 放進你的 PDP Q&A、A+，以及 llms.txt/FAQ 檔（接 hack 4、21）。
14
HACK 14 OF 24

A Free Custom GPT That Spits Out A Differentiated Product

一個免費的自訂 GPT，直接吐出差異化產品

THE PAYOFF ／ 為什麼值得做

One seller (two successful exits) built a free GPT that ingests Helium 10 data and outputs a level-3 differentiated product using the '80% match, 20% twist' rule. Skips the manual analysis. He launched a wooden wall-mounted dog bowl from this exact flow.

一位業者（兩次成功出場）做了一個免費 GPT：吃進 Helium 10 資料，用「80% 相似、20% 變化」法則吐出一個第三級差異化產品，省掉手動分析。他就是用這套流程推出了「壁掛式木製狗碗」。

STEP BY STEP ／ 逐步操作
  1. 01In Helium 10 X-Ray (or Atlas / Jungle Scout), export the top 30 products for a 3–4 word search term (e.g., 'scented candles men').在 Helium 10 X-Ray（或 Atlas / Jungle Scout）匯出某個 3–4 字搜尋詞的前 30 名產品（例：「scented candles men」）。
  2. 02Open Andrew's free Custom GPT (search ChatGPT for 'Optimize Hero' / 一位業者product differentiation).打開 這個免費自訂 GPT（在 ChatGPT 搜「Optimize Hero」／一位業者product differentiation）。
  3. 03Upload the CSV. Prompt: 'Generate 10 product ideas that are 80% aligned with the top features and 20% differentiated. Identify the dominant scent, size, and positioning — then twist one variable.'上傳 CSV。提示詞：「生成 10 個產品點子，與熱賣特徵 80% 對齊、20% 差異化。找出主導的香味、尺寸與定位——然後只扭轉其中一個變數。」
  4. 04Take the top 3 ideas to Nano Banana (in Gemini). Generate 30–40 product images with packaging variations.把前 3 名點子帶到 Nano Banana（在 Gemini 裡），生 30–40 張含不同包裝的產品圖。
  5. 05Run images through PickFu, ProductPinion or OptimizeHero.ai for human votes + comments.把圖丟 PickFu、ProductPinion 或 OptimizeHero.ai 收真人投票＋留言。
  6. 06Feed comments back into Nano Banana ('more pink packaging, female model, bright smile') and re-generate.把留言餵回 Nano Banana（「包裝更粉、換女模、笑容更亮」）重新生成。

結語：Winners outperform losers 10×／贏家的表現是輸家的 10 倍。

15
HACK 15 OF 24

The Label Hack That Added 70% Sales In 2 Weeks

2 週增加 70% 銷售的「標籤」小技巧

THE PAYOFF ／ 為什麼值得做

Add (or enlarge) a USP label on the product itself in your main image, even if no such label exists on the physical product. Click share jumped from 9.34% to 17%. Another test: 70% extra sales, +$6,700 in 2 weeks. One seller says Dude Wipes, Angry Orange, and Gorilla Grip all do it.

在主圖的產品本體上加上（或放大）一個 USP 賣點標籤——就算實體產品上根本沒有這個標籤也行。click share 從 9.34% 跳到 17%；另一次測試：多 70% 銷售、2 週 +$6,700。一位業者說 Dude Wipes、Angry Orange、Gorilla Grip 都在這樣做。

STEP BY STEP ／ 逐步操作
  1. 01Take your current main image. Use ChatGPT or your designer to add an oversized label/sticker on the product itself.拿你現在的主圖，用 ChatGPT 或你的設計師，在產品本體上加一個放大的標籤/貼紙。
  2. 02The label says ONE of: a) your top-volume keyword (e.g., 'Made in USA Pack of 10'), b) your single biggest USP, or c) a quantity indicator.標籤只寫三選一：a) 你搜尋量最大的關鍵字（例：「Made in USA Pack of 10」）、b) 你最大的單一賣點，或 c) 一個數量標示。
  3. 03Create 3 variations: safe/compliant, medium-risk, pushed.做 3 個版本：安全合規、中度風險、激進。
  4. 04Upload all 3 to ProductPinion or PickFu. Run a 100-vote test with 4 competitor main images side-by-side (SERP sim).三個版本都上傳 ProductPinion 或 PickFu，跟 4 張競品主圖並排跑 100 票測試（模擬 SERP）。
  5. 05Take the winner to Amazon Manage Your Experiments. A/B for 2 weeks.把贏家拿進 Amazon Manage Your Experiments，做 2 週 A/B。
  6. 06If it loses, isolate one variable at a time (bottle change, hangtag, color) and re-test. Most 'losses' are won by finding which element actually failed.如果輸了，一次只隔離一個變數（換瓶身、吊牌、顏色）重測。多數「敗仗」是靠找出「真正失分的那個元素」翻盤的。
16
HACK 16 OF 24

Capture Creator Emails Before They Ghost You

在創作者神隱之前，先抓到他們的 email

THE PAYOFF ／ 為什麼值得做

One brand does $250K/month GMV on TikTok Shop. The single mechanic that made it work: pull the creator's phone, email, and shipping OFF TikTok DMs and onto an email list so they stop ghosting, all with a usage-rights consent hack baked in.

一個品牌在 TikTok Shop 每月做 $250K GMV。真正讓它跑起來的關鍵只有一招：把創作者的電話、email、寄送資訊「從 TikTok 私訊裡拉出來」放進 email 名單，他們就不會再神隱——而且順手把「使用授權同意」也一併收好。

STEP BY STEP ／ 逐步操作
  1. 01Auto-send the moment you approve a creator sample: 'You've been selected for a boosted-ads opportunity. Please fill this form to qualify.'一核准創作者的樣品就自動寄出：「你被選中參加加碼投放（boosted-ads）機會，請填這份表單以取得資格。」
  2. 02Form collects: email, phone, shipping, plus 'Are you excited to post about our products? Yes/No.' When they click Yes, NEXT field is a usage-rights agreement. 95% click yes. Now you can run their videos on Meta legally.表單收集：email、電話、寄送資訊，再加一題「你期待幫我們產品發文嗎？Yes/No」。他們按 Yes 後，下一欄就是使用授權同意書。95% 會按 Yes——你就能合法把他們的影片拿去 Meta 投放。
  3. 03Drop them into Beehiiv. Day 0: brand story. Day 1: WhatsApp group invite. Day 2: creative brief with proven hooks. Day 3: 5-min training playlist. Day 4–30: bingo challenge ('10 videos using 3 of our hooks = $20 gift card at $50 GMV').把他們丟進 Beehiiv。Day 0：品牌故事；Day 1：WhatsApp 群組邀請；Day 2：附驗證過鉤子的 creative brief；Day 3：5 分鐘教學播放清單；Day 4–30：賓果挑戰（「用我們 3 個鉤子拍 10 支影片、達 $50 GMV＝$20 禮物卡」）。
  4. 04Identify ~7 creators who post but don't go viral. Light a small ad budget behind their videos so views jump from 200 to 5,000. They get a dopamine hit, so they keep posting.找出約 7 個「有發文但沒爆」的創作者，在他們影片後面點一小筆廣告預算，讓觀看從 200 跳到 5,000。他們嚐到多巴胺，就會一直發。
  5. 05Every Wednesday 8:15pm, host a live audit call with creators. Repeat the 5-stage funnel: Unactivated → Activated → Emerging → Performer → Evangelist.每週三晚上 8:15，跟創作者開一場直播檢討會。反覆推這個 5 階段漏斗：未啟動 → 已啟動 → 崛起中 → 表現者 → 傳教士。
17
HACK 17 OF 24

Amazon's 7-Second AI Video Beats Pro Footage

Amazon 的 7 秒 AI 影片，打贏專業拍攝

THE PAYOFF ／ 為什麼值得做

Sponsored Brand Video gets an order of magnitude more impressions than any other ad format. In split tests, Amazon's FREE 7-second AI video has beaten professionally-produced footage. One seller show us how to build it in under 10 minutes with zero budget.

Sponsored Brands Video 的曝光量比任何其他廣告格式高一個數量級。在分組測試中，Amazon 免費的 7 秒 AI 影片打贏了專業製作的影片。一位業者教我們如何 0 預算、10 分鐘內做出來。

STEP BY STEP ／ 逐步操作
  1. 01Seller Central → Campaign Manager → Create Campaign → Sponsored Brands → Video format.Seller Central → Campaign Manager → Create Campaign → Sponsored Brands → 選 Video 格式。
  2. 02In the creative builder, click 'Generate video with AI'. It pulls from your existing listing images automatically.在素材編輯器點「Generate video with AI」，它會自動抓你現有的 listing 圖片。
  3. 03Pick ONE keyword cluster you already rank well for in Sponsored Products (e.g., your travel use-case keywords). Don't reuse a broad list.挑一個你在 Sponsored Products 已經排名不錯的關鍵字群（例如你的旅行使用情境關鍵字）。不要沿用一份大而雜的清單。
  4. 04Tailor the auto-generated 7-second text overlay to that exact intent (e.g., 'Fits any carry-on' for the travel cluster).把自動生成的 7 秒文字疊層，改成貼合那個意圖（例如旅行群就寫「Fits any carry-on」）。
  5. 05Duplicate for 2 more clusters, each with its own custom-text AI video.複製到另外 2 個關鍵字群，每群都配自己客製文字的 AI 影片。
  6. 06Launch all 3 at the same daily budget. After 14 days, scale the lowest-ACoS winner. Kill the rest.三支用相同日預算一起開跑。14 天後，把 ACoS 最低的贏家放大，其餘關掉。
18
HACK 18 OF 24

Pull Your SP Video Hooks Straight From POE

直接從 POE 抓出你的 SP 影片鉤子

POE＝Product Opportunity Explorer（Seller Central 的商機探測器）

THE PAYOFF ／ 為什麼值得做

Sponsored Product Video now lets you stack 3–5 short videos per ASIN, each with its own feature title. One seller says pulling those titles from Product Opportunity Explorer means the first frame matches the exact customer objection. CTR lift on day one.

Sponsored Products Video 現在讓你每個 ASIN 疊 3–5 支短影片，每支各有自己的功能標題。一位業者說：把這些標題直接從 Product Opportunity Explorer 抓出來，第一格畫面就能精準對到顧客的疑慮，第一天就見 CTR 提升。

STEP BY STEP ／ 逐步操作
  1. 01Seller Central → Brand Analytics → Product Opportunity Explorer → search your top niche keyword.Seller Central → Brand Analytics → Product Opportunity Explorer → 搜你最主要的利基關鍵字。
  2. 02Click into the niche → Customer Review Insights → sort by top pain points and desired features.點進該利基 → Customer Review Insights → 依「主要痛點」與「渴望的功能」排序。
  3. 03Pick the top 3 pain points (e.g., 'adjustable fit,' 'fits in shoes,' 'doesn't slide').挑前 3 個痛點（例：「可調鬆緊」「能塞進鞋裡」「不會滑」）。
  4. 04For each: open Google Flow (labs.google/flow). Upload your hero image as frame 1, demo image as frame 2. Prompt: 'Transition image 1 → image 2 showing [pain point solved].'每個痛點：打開 Google Flow（labs.google/flow），把主視覺圖當第 1 格、示範圖當第 2 格。提示詞：「Transition image 1 → image 2 showing [痛點被解決]。」
  5. 05In your SP Video campaign, upload all 3. Name each video's feature title EXACTLY as the pain point.在你的 SP Video 活動裡上傳這 3 支，每支影片的功能標題「一字不差」用那個痛點命名。
  6. 06Let them rotate. Cut the worst performer at day 14.讓它們輪播，第 14 天砍掉表現最差的那支。
19
HACK 19 OF 24

The Dormant AMC Audience That Wakes Up On Sale Days

只在特賣日「甦醒」的休眠 AMC 受眾

THE PAYOFF ／ 為什麼值得做

Build an AMC audience of people who only buy your category when a coupon or lightning deal is live. Hold it dormant 364 days. Activate it the morning of a deal. One seller says you're spending only on shoppers primed to convert.

用 AMC 建一群「只在有優惠券或 Lightning Deal 時才買你品類」的受眾。讓它休眠 364 天，等特賣當天早上再啟動。一位業者說：這樣你的預算只花在「已經準備好轉換」的買家身上。

STEP BY STEP ／ 逐步操作
  1. 01Seller Central → Ads Console → Measurement & Reporting → Amazon Marketing Cloud.Seller Central → Ads Console → Measurement & Reporting → Amazon Marketing Cloud。
  2. 02Audience builder → 'Generate SQL with AI' → prompt: 'Build an audience of shoppers who purchased in my category ONLY when a discount badge, coupon, or lightning deal was active in the last 12 months.'Audience builder → 「Generate SQL with AI」→ 提示詞：「建立一個受眾：過去 12 個月內，只在有折扣標章、優惠券或 Lightning Deal 時才買我這個品類的買家。」
  3. 03Save the audience. It sits unused. AMC just stores the segment.儲存受眾，放著不用。AMC 只是幫你存下這個分群。
  4. 04The morning your next coupon or Lightning Deal goes live: open your best Sponsored Brands campaign → attach this audience with a +50% bid modifier.下一檔優惠券或 Lightning Deal 上線的當天早上：打開你最強的 Sponsored Brands 活動 → 掛上這個受眾，加成 +50%。
  5. 05Deactivate the audience the moment the deal ends.特賣一結束就立刻停用這個受眾。
  6. 06Compare conversion rate during the promo window vs. your default audience. Big delta.比較「促銷期間 vs. 你的預設受眾」的轉換率——差距會很大。
20
HACK 20 OF 24

The 60-Minute Schema Fix That Gets You Cited By ChatGPT

60 分鐘的 schema 修正，讓 ChatGPT 引用你

THE PAYOFF ／ 為什麼值得做

Across hundreds of LLM citations analyzed by One seller's team, the #1 predictor of being cited by ChatGPT/Perplexity/Gemini wasn't backlinks or content. It was validated schema markup. No schema = LLMs skip your site. This is 2004-era Google: easy now, defensible later.

一位業者的團隊分析了數百筆 LLM 引用，發現被 ChatGPT/Perplexity/Gemini 引用的第一預測因子，不是反向連結、也不是內容，而是「通過驗證的 schema 標記」。沒有 schema＝LLM 直接跳過你的網站。這就像 2004 年的 Google：現在做很簡單，之後就是護城河。

STEP BY STEP ／ 逐步操作
  1. 01schema.org/validator → paste your homepage, product page, and About page URLs.到 schema.org/validator → 貼上你的首頁、產品頁、About 頁網址。
  2. 02If errors or 'no schema detected,' ChatGPT prompt: 'Generate complete JSON-LD schema for a [product/brand] page at [URL]. Include Organization, Product, FAQPage, and Review schema. Return only the script tag for the <head>.'如果報錯或「no schema detected」，用 ChatGPT 提示詞：「為 [URL] 的 [產品/品牌] 頁面產生完整 JSON-LD schema，包含 Organization、Product、FAQPage 與 Review schema，只回傳要放進 <head> 的 script 標籤。」
  3. 03Paste the <script type='application/ld+json'> block into <head>. Shopify: Online Store → Themes → Edit code → theme.liquid. WordPress: 'Insert Headers and Footers' plugin.把 <script type='application/ld+json'> 區塊貼進 <head>。Shopify：Online Store → Themes → Edit code → theme.liquid；WordPress：用「Insert Headers and Footers」外掛。
  4. 04Re-run the validator to confirm green.重跑驗證器，確認全綠通過。
  5. 05Repeat for your top 3 product pages and a dedicated FAQ page.對你前 3 個產品頁，以及一個獨立的 FAQ 頁，重複同樣動作。
  6. 06Add an audio version of your top FAQ (voice schema is now weighted for the 55% of 20-somethings using voice search via Alexa Plus / ChatGPT voice).幫你最重要的 FAQ 加一個語音版本（現在 voice schema 有加權——55% 的 20 幾歲族群用 Alexa Plus / ChatGPT 語音在做語音搜尋）。
21
HACK 21 OF 24

Find The Alexa Shopping Questions Your Listing Doesn't Answer

找出你 listing 沒回答到的 Alexa 購物問題

THE PAYOFF ／ 為什麼值得做

Alexa Shopping is driving $12B+ in incremental revenue and growing. Winners aren't the ones with the best titles, they're the ones whose listings answer the use cases Alexa gets asked about. In her example, One seller found 'coconut oil for pet skin' and 'for teeth whitening,' all questions zero top listings addressed.

Alexa Shopping 正帶來 $12B+ 的增量營收，還在成長。贏家不是標題最漂亮的，而是「listing 回答得了 Alexa 被問到的使用情境」的那些人。一位業者的例子裡，她挖到「coconut oil for pet skin（椰子油護寵物皮膚）」和「for teeth whitening（美白牙齒）」——這些問題，前段班的 listing 沒有一個回答到。

STEP BY STEP ／ 逐步操作
  1. 01ChatGPT: 'What are the top 20 non-obvious use cases people ask AI shopping assistants about for [your product category]?'ChatGPT：「針對 [你的產品品類]，人們最常問 AI 購物助理、但不那麼直覺的前 20 個使用情境是什麼？」
  2. 02Cross-check on Amazon mobile: ask Alexa 'What can I use [product] for besides [obvious use]?' Screenshot every answer.到 Amazon 手機 App 交叉驗證：問 Alexa「[產品] 除了 [明顯用途] 以外還能用來做什麼？」把每個答案截圖。
  3. 03For each non-obvious use case (e.g., 'for pets,' 'for car detailing'), check the top 5 listings — if none mention it, that's your gap.每個非直覺用途（例：「for pets」「for car detailing」）都去看前 5 名 listing——如果沒人提到，那就是你的缺口。
  4. 04Add a Q&A in your A+ Premium answering it directly: 'Q: Can I use this for my dog's skin? A: Yes, [exact use].'在你的 A+ Premium 加一則直接回答的 Q&A：「Q：這能用在我家狗的皮膚上嗎？A：可以，[具體用法]。」
  5. 05Add the same to your Customer Q&A section and post the question from a buyer account, answer from your brand account within 24 hrs.同樣內容加到 Customer Q&A：用買家帳號發問、24 小時內用你的品牌帳號回答。
  6. 06Update backend search terms with the long-tail phrases (e.g., 'coconut oil for pets'), 250 bytes max.用這些長尾詞更新後台 search terms（例：「coconut oil for pets」），上限 250 bytes。
22
HACK 22 OF 24

The Email Teardown: Plain Text Beats Pretty

Email 拆解：純文字勝過美編

原 hack 22（美國 4 帳戶記帳系統）因台灣受眾不適用已替換為此通用版

THE PAYOFF ／ 為什麼值得做

Everyone believes designed HTML emails convert better. They don't. Plain text with maybe one image lands in the inbox more often, scans faster, and feels human. Layer send-time timing on top and opens lift 200–300%.

大家都以為美編過的 HTML email 轉換比較好，其實不然。純文字（最多配一張圖）更容易進到收件匣、掃讀更快、也更像真人寫的。再疊上「寄送時機」技術，開信率能拉高 200–300%。

STEP BY STEP ／ 逐步操作
  1. 01Rewrite your welcome email as plain text + at most one image. Designed HTML gets filtered to Promotions; plain text lands in the primary inbox.把你的歡迎信改成純文字＋最多一張圖。美編 HTML 常被歸到促銷夾，純文字才進主收件匣。
  2. 02Payoff at the top: coupon code FIRST, then a reason to scroll ('scroll down for a bonus gift'). Unsubscribe stays small at the very bottom.好處前置：優惠碼「先給」，再給一個往下滑的理由（「往下滑有加碼好禮」）。退訂連結放最底部、字小。
  3. 03Kill all seller-speak: no 'our goal is…', no 'product launch'. Every line is about the customer and their benefit, in their language.砍掉所有賣家腔：不要「我們的目標是…」、不要「新品上市」。每一句都寫顧客和他的好處，用顧客的語言。
  4. 04A 'gift' they were already promised isn't a gift. Stack a real surprise on top (extra 10% today, or a share-with-a-friend coupon).「早就答應要給的」不算禮物。要在上面再疊一個真驚喜（今天加碼 10%，或一張「分享給朋友」的優惠券）。
  5. 05Humanize AI-written copy with Walter AI to a score under 17 (passes AI detection). For replenishment products, add usage-timed trigger emails (e.g., day 25 of a 30-day supply).用 Walter AI 把 AI 寫的文案「人味化」到分數低於 17（能過 AI 偵測）。補貨型產品要加「按用量計時」的觸發信（例：30 天用量的第 25 天）。
  6. 06Flat ~2% click rate? Could be bots or security scanners. Connect AWeber's MCP to Claude and ask it to analyze unique clicks and drop-off by flow.點擊率老是卡在約 2%？可能是機器人或資安掃描器造成的。把 AWeber 的 MCP 接到 Claude，請它按各流程分析「不重複點擊」與流失點。
23
HACK 23 OF 24

The Untapped-Angle Method (No New Product Needed)

未開發切角法（不用開新品）

THE PAYOFF ／ 為什麼值得做

Stop fighting for saturated keywords. Find a product where the top 10 sellers have NO long-tail angle in their title, then own it. One seller added 'hand sanitizer party favors' to an existing title, ranked #1 the next week, rode the wave with zero new product development.

別再跟人搶已經打爛的關鍵字。找一個「前 10 名賣家的標題裡都沒有任何長尾切角」的產品，然後把那個切角占下來。有個賣家在既有標題加上「hand sanitizer party favors（乾洗手小禮）」，隔週就排到第 1，完全沒開發新品就吃到這波。

STEP BY STEP ／ 逐步操作
  1. 01Run Helium 10 Cerebro on your main niche keyword (or Brand Analytics SQP if no H10).對你的主要利基關鍵字跑 Helium 10 Cerebro（沒有 H10 就用 Brand Analytics SQP）。
  2. 02Pull every long-tail variation with search volume 2,000–10,000: NOT the 100K head terms.抓出所有搜尋量 2,000–10,000 的長尾變體——不要那些 10 萬量級的大詞。
  3. 03For each: search Amazon directly. Of the top 10 listings, how many have that long-tail phrase in their title? (Helium 10 Title Density auto-checks this.)每個詞都直接在 Amazon 搜：前 10 名 listing 裡，有幾個標題含這個長尾詞？（Helium 10 Title Density 會自動幫你算。）
  4. 04Find one where <2 of the top 10 use the phrase AND total results <500.找一個「前 10 名裡用這詞的 <2 個，而且總結果數 <500」的。
  5. 05Cross-check VOC.AI's demand-to-supply ratio for that keyword. Pursue only if low.用 VOC.AI 交叉查這個關鍵字的「需求／供給比」，只有在比值低時才追。
  6. 06Add the phrase to your existing title TODAY (and to bullet #1). New launch? Source the variant that matches the angle. Index check in 7 days.今天就把這個詞加進你現有的標題（以及第 1 條 bullet）。要推新品？就找符合這切角的變體備貨。7 天後檢查是否被收錄。
24
HACK 24 OF 24

Pre-Build 20 Creatives Before You Touch An Affiliate

在接觸任何聯盟客之前，先備好 20 支素材

THE PAYOFF ／ 為什麼值得做

99% of brands chase TikTok Shop affiliates first and get 0.1% response rates because they have no proof of conversion. Reverse the order: pre-build content, test in GMV Max, then approach affiliates with a winning script. Affiliates in days, not months.

99% 的品牌都先去追 TikTok Shop 聯盟客，卻只拿到 0.1% 的回覆率——因為他們拿不出轉換的證據。把順序反過來：先備好內容、在 GMV Max 裡測、然後帶著一支已經會贏的腳本去找聯盟客。招募聯盟客用「天」算，不是用「月」算。

STEP BY STEP ／ 逐步操作
  1. 01Pull KaloData (or Shop+) → export the top 50 organic videos by GMV in your category in the last 30 days.打開 KaloData（或 Shop+）→ 匯出你品類過去 30 天 GMV 前 50 名的自然流量影片。
  2. 02List the 5 most common hooks (first 3 sec), 5 most common settings, 5 most common demo angles.列出最常見的 5 種鉤子（前 3 秒）、5 種場景、5 種示範角度。
  3. 03Flipbook approach: shoot or AI-generate 5 hooks × 5 demos = 25 short-form variants, all 15–30 sec, all featuring your product.用「翻頁書」組合法：拍攝或 AI 生成 5 種鉤子 × 5 種示範＝25 支短影片變體，全部 15–30 秒、全部帶你的產品。
  4. 04Upload all 25 as shoppable videos in TikTok Shop Seller Center. Enable GMV Max at $50/day split across them.把 25 支都上傳到 TikTok Shop Seller Center 當可購物影片，開 GMV Max、每天 $50 分攤在這些影片上。
  5. 05After 7 days, identify the 2–3 variants with the highest 15-sec view rate AND add-to-cart rate.7 天後，挑出「15 秒觀看率 + 加購率」都最高的 2–3 支。
  6. 06Write a 4-line creator brief (NOT a 10-page PDF) using only the winning hooks: 'Say [hook]. Show [demo]. Use [CTA].' Send to 50 niche affiliates with the proof: 'Our ads using this script convert at X%.'寫一份 4 行的 creator brief（不是 10 頁 PDF），只用會贏的鉤子：「說 [鉤子]。示範 [demo]。用 [CTA]。」帶著證據寄給 50 個利基聯盟客：「我們用這支腳本的廣告轉換率是 X%。」