貓咪品牌創立劇本：從籌備到海外擴張Cat Brand Launch Playbook: From Preparation to Overseas Expansion

摘要Summary

從 2026 年 11 月簽約開始，用 12 個月兼職籌備，2027 年 11 月 Rj 離職全職、募資上線，2028 年下半年開始海外擴張。前 12 個月估計需要金主投入約新台幣 550 萬元，到第 18 個月累計約 950 萬元；海外擴張另需一輪募資。Starting with signed agreements in November 2026, the founders prepare part-time for 12 months; in November 2027 Rj leaves his job to go full-time and the crowdfunding campaign goes live; overseas expansion begins in the second half of 2028. The first 12 months need roughly NT$5.5 million from the investor, rising to about NT$9.5 million by month 18; overseas expansion needs a further funding round.

這是一份規劃劇本，不是承諾。金額有一部分來自查證過的行情，一部分是粗估，表格中都有標示；粗估的部分要在詢價後更新。This is a planning playbook, not a commitment. Some figures come from verified market rates and some are rough estimates, marked as such in the tables; the estimates should be updated once quotes come in.

三道關卡決定是否往下走，資金也跟著關卡分批撥付Three gates decide whether to go on, and funding is released gate by gate:

1. 關卡一（2027 年 3 月底）：需求成立。 飼主是否每天主動打開、願意付費。Gate 1 (end of March 2027): demand is real. Owners open the app daily on their own and are willing to pay.
2. 關卡二（2027 年 7 月底）：原型可用。 在真實家庭穩定運作，單件成本確定。Gate 2 (end of July 2027): the prototype works. It runs reliably in real homes and the unit cost is known.
3. 關卡三（2027 年 10 月底）：離職條件。 指標達標，金主承諾下一筆資金與 Rj 至少 12 至 18 個月的薪資。Gate 3 (end of October 2027): conditions to resign. Targets are met, and the investor commits the next tranche plus at least 12 to 18 months of Rj's salary.

這份劇本有三個關鍵判斷Three key judgements behind this playbook:

• 台灣寵物科技的群眾募資規模很小。 嘖嘖上找得到的寵物智慧產品，募得金額多在新台幣 12 萬到 120 萬元之間，所以台灣募資的任務是驗證與累積第一批用戶，不是主要資金來源。Pet-tech crowdfunding in Taiwan is small. Smart pet products on zeczec mostly raised NT$120,000 to 1.2 million, so a Taiwan campaign is for validation and first users, not the main source of capital.
• 海外才是規模。 美國有約 5,300 萬個養貓家庭，日本有約 884 萬隻貓，都是台灣的數倍到數十倍，這是金主最能發揮的地方。Scale lies overseas. The US has about 53 million cat-owning households and Japan about 8.84 million cats, many times Taiwan's market. This is where the investor can contribute most.
• 最大的人力缺口是技術負責人。 產品價值主要在軟體與影像處理，這個人最好在第一個月就開始接觸。The biggest staffing gap is a technical lead. Most of the product's value is in software and video processing, so the search should start in month one.

創立路線圖Launch Roadmap

每道關卡都是一個「繼續、調整或停止」的決定，金主的下一筆資金也在通過後才撥付。Rj 的離職放在關卡三之後，而不是日期到了就走。The roadmap runs in six phases: Phase 0, agreements and set-up (November 2026); Phase 1, demand validation (December 2026 to March 2027); Phase 2, prototyping and product (April to July 2027); Phase 3, beta and crowdfunding preparation (August to October 2027); Phase 4, full-time, campaign launch and shipping (November 2027 to April 2028); Phase 5, overseas expansion (from May 2028). Each gate is a decision to continue, adjust or stop, and the investor's next tranche is released only after it is passed. Rj resigns after Gate 3, not simply when a date arrives.

台灣市場分析Taiwan Market Analysis

台灣適合當驗證市場，但規模有限。需求在成長，寵物科技的付費習慣卻還小，所以台灣的任務是磨出產品、累積口碑與數據，再帶去海外。Taiwan is a good test market, but a limited one. Demand is growing while the habit of paying for pet tech is still small, so Taiwan's job is to refine the product and build reputation and data to take overseas.

需求面：貓在增加，支出在增加Demand: More Cats, More Spending

成長主要來自都會多貓家庭，飼主以單身族與住在集合住宅的族群為多，就是我們的目標客群。Growth comes mainly from urban multi-cat households, often single people living in apartments, which is exactly our target customer.

供給面：台灣寵物科技募資的真實規模Supply: The Real Size of Pet-Tech Crowdfunding

嘖嘖平台整體很大：2026 年 5 月累計集資突破新台幣 200 億元，2025 年全年約 35 億元。但寵物智慧產品的案例都很小：The zeczec platform as a whole is large: cumulative funding passed NT$20 billion in May 2026, with about NT$3.5 billion raised in 2025. But smart pet products have raised little:

對照之下，整體家電類的成功案例可以很大，例如窗之魔手洗窗機募得約新台幣 2,020 萬元。這代表平台有購買力，是寵物科技還沒有出現打中情緒的產品。我們的「情緒價值」訴求或許有機會突破，但預算不應建立在突破的假設上。By contrast, home-appliance campaigns can be very large; a window-cleaning robot raised about NT$20.2 million. The platform has buying power, but pet tech has not yet produced a product that hits an emotional nerve. Our emotional-value pitch may break through, but the budget should not assume it will.

通路與價格帶Channels and Price Range

• 線上 Online： momo 上貓用品與狗用品銷量約 6:4，貓用品成長較快（引自統合筆記）。On momo, cat and dog products sell about 60:40, with cat products growing faster (from the combined notes).
• 實體 Retail： 東森寵物約 109 家、寵物公園 100 家以上，總部通常要求試賣成績（引自統合筆記）。ETtoday Pets has about 109 stores and Pet Park over 100; headquarters usually want trial-sale results (from the combined notes).
• 價格帶 Price range： 競品第一年含訂閱約 119 至 370 美元；嘖嘖上寵物項圈的早鳥價約新台幣 1,099 至 5,900 元。我們的「項圈加鏡頭」組合，募資早鳥價可能落在新台幣 4,000 至 6,000 元，待成本確定後訂價。Competitors cost about US$119 to 370 in year one including subscription; early-bird prices for pet collars on zeczec ran NT$1,099 to 5,900. Our collar-plus-camera set might sit at NT$4,000 to 6,000 early-bird, to be priced once costs are known.

逐月劇本Month-by-Month Playbook

第 1 個月對應 2026 年 11 月。每個階段寫清楚三方各做什麼：創辦團隊（Rj 與伴侶，內部分工由兩人自行決定）、金主、外部人員，以及這個階段結束時要交付什麼。Month 1 is November 2026. Each phase sets out what three parties do: the founding team (Rj and his partner, who divide the work between themselves), the investor and outside contributors, plus what the phase must deliver.

第 0 階段：談定與設立（2026 年 11 月，第 1 個月）Phase 0: Agreements and Set-up (November 2026, month 1)

• 創辦團隊 Founding team： 確認 Rj 在 ChargeSPOT 勞動契約中兼職、競業、智慧財產與禁止招攬的條款；與金主談定股東協議；建立專案用的個人帳號與文件空間；開始接觸技術負責人候選；從人脈列出 20 個候選多貓家庭與 5 位候選 KOL。Review the side-work, non-compete, IP and non-solicitation terms in Rj's ChargeSPOT contract; agree the shareholder agreement with the investor; set up personal accounts and file space for the project; start approaching technical-lead candidates; list 20 candidate multi-cat households and 5 candidate KOLs from personal networks.
• 金主 Investor： 簽股東協議、撥付第一筆資金；介紹律師與會計師；開始接觸深圳代工廠人脈。Sign the shareholder agreement and release the first tranche; introduce a lawyer and an accountant; start reaching out to Shenzhen manufacturers.
• 外部 Outside： 律師處理公司設立與勞務出資條款；會計師處理勞務出資作價與稅務。A lawyer handles incorporation and labour-contribution terms; an accountant handles the valuation and tax of the labour contribution.
• 交付 Deliverables： 簽完的股東協議、三道關卡的指標門檻、第一階段預算。A signed shareholder agreement, thresholds for all three gates, and the Phase 1 budget.

第 1 階段：需求驗證（2026 年 12 月至 2027 年 3 月，第 2 至 5 個月）Phase 1: Demand Validation (December 2026 to March 2027, months 2 to 5)

這個階段對應統合筆記裡的 16 週測試。This phase is the 16-week test in the combined notes.

• 創辦團隊 Founding team： 設計驗證方法與指標儀表板；安裝市售鏡頭、確認能否取得影像；用開發板測試感測器；維繫 10 至 15 個試用家庭；每天人工挑片、剪片、配音，放進共享相簿模擬情緒存摺；做使用訪談；每月向金主報告。Design the validation method and metrics dashboard; install retail cameras and check whether video can be accessed; test sensors on dev boards; look after 10 to 15 test households; pick, cut and voice clips by hand every day into shared albums to simulate the emotional savings account; run user interviews; report to the investor monthly.
• 金主 Investor： 每月一次檢視數字；協助引薦 KOL 與代工廠。Review the numbers monthly; help with KOL and manufacturer introductions.
• 外部 Outside： 技術顧問（兼職）、插畫師替每隻試用貓畫分身畫稿、獸醫或貓行為顧問審視判讀方式。A part-time technical adviser; an illustrator drawing a twin sketch for each test cat; a vet or cat-behaviour adviser reviewing how behaviour is interpreted.
• 交付 Deliverables： 驗證報告，包括每天主動打開次數、看完片段比例、分身像不像、付費意願、感測器準確度。A validation report covering daily opens, share of clips watched, twin likeness, willingness to pay and sensor accuracy.
• 決策 Decision： 關卡一。達標就撥第二筆資金；未達標就調整方向或停止。Gate 1. If met, the second tranche is released; if not, adjust direction or stop.

第 2 階段：打樣與產品（2027 年 4 至 7 月，第 6 至 9 個月）Phase 2: Prototyping and Product (April to July 2027, months 6 to 9)

• 創辦團隊 Founding team： 定義功能優先順序；管理代工廠、韌體外包與設計師；控成本與進度；和品牌設計師一起定品牌口吻與內容風格；開始經營社群帳號與等待名單。Set feature priorities; manage the manufacturer, firmware contractor and designers; control cost and schedule; settle brand voice and content style with the brand designer; start running social accounts and a waiting list.
• 金主 Investor： 撥付第二筆資金；協助談代工條件。Release the second tranche; help negotiate manufacturing terms.
• 外部 Outside： 技術負責人轉全職、App 工程師加入、韌體外包、動畫設計師、品牌設計師、代工廠。The technical lead goes full-time; an app engineer joins; firmware contractor, animator, brand designer and manufacturer.
• 交付 Deliverables： 50 個可以配戴的感測項圈原型；App 第一版，包含情緒存摺、區域、簡單版分身；品牌名稱與識別；代工廠完整報價。50 wearable sensor-collar prototypes; app version one with the savings account, zones and a simple twin; brand name and identity; a full manufacturer quote.
• 決策 Decision： 關卡二。Gate 2.

第 3 階段：試用與募資準備（2027 年 8 至 10 月，第 10 至 12 個月）Phase 3: Beta and Crowdfunding Preparation (August to October 2027, months 10 to 12)

• 創辦團隊 Founding team： 跑 30 至 50 戶試用；送 NCC 認證；寫募資提案；募資預熱，包括累積等待名單、安排 KOL 試用影片、管理社群；Rj 準備 ChargeSPOT 的交接文件。Run a 30 to 50 household beta; submit for NCC certification; write the crowdfunding proposal; warm up the campaign by growing the waiting list, arranging KOL trial videos and managing community; Rj prepares handover documents for ChargeSPOT.
• 金主 Investor： 撥付第三筆資金；介紹海外通路與投資人，為第二年預備。Release the third tranche; introduce overseas channels and investors ahead of year two.
• 外部 Outside： 募資影片與頁面製作、認證實驗室。Crowdfunding video and page production; certification lab.
• 交付 Deliverables： 試用數據、NCC 證書、募資頁面與影片、至少數百人的等待名單。Beta data, NCC certificate, campaign page and video, and a waiting list of at least several hundred people.
• 決策 Decision： 關卡三，也就是離職決定。Gate 3, the decision to resign.

第 4 階段：全職、募資上線、出貨（2027 年 11 月至 2028 年 4 月，第 13 至 18 個月）Phase 4: Full-time, Campaign Launch and Shipping (November 2027 to April 2028, months 13 to 18)

• 創辦團隊 Founding team： Rj 依契約提前告知並離職、就任總經理；嘖嘖募資上線；募資期間的社群與客服；首批生產與出貨；出貨後的用戶經營。Rj gives notice as his contract requires, resigns and becomes general manager; the zeczec campaign goes live; community and customer service during the campaign; first production run and shipping; user engagement after delivery.
• 金主 Investor： 撥付第四筆資金、開始支付 Rj 薪資；主導第二輪募資與海外合作。Release the fourth tranche and start paying Rj's salary; lead the second funding round and overseas partnerships.
• 外部 Outside： 社群與客服人員、物流商。Community and customer-service hire; logistics provider.
• 交付 Deliverables： 募資結果、首批出貨、訂閱轉換率、電商上架。Campaign results, first shipments, subscription conversion, and marketplace listings.

第 5 階段：海外擴張（2028 年 5 月起）Phase 5: Overseas Expansion (from May 2028)

詳見「海外擴張」一節。See the Overseas Expansion section.

團隊配置Team

第一年只有技術負責人與 App 工程師是全職，其他都用專案或兼職方式，把固定成本壓低。「台灣行情」欄是月薪區間，取自 GetCareers 依主計總處 2025 年受僱員工薪資調查整理的估算；實際薪資依年資與談判而定。In year one only the technical lead and the app engineer are full-time; everyone else works on a project or part-time basis to keep fixed costs low. The market-rate column shows monthly salary ranges estimated by GetCareers from Taiwan's 2025 government earnings survey; actual pay depends on seniority and negotiation.

技術負責人是整個劇本最關鍵的一個人。 如果找得到能共同創業的人，給股份比給高薪更合理，也能減少第一年的現金支出。他的股份會和創辦團隊的勞務出資共用「不超過一半」的法規上限，要先和金主談好。The technical lead is the single most important person in this playbook. If a true co-founder can be found, equity makes more sense than a high salary and cuts year-one cash costs. Their shares would share the legal cap on labour-contributed shares (no more than half) with the founding team's, so this needs agreeing with the investor first.

外部夥伴：Sam 的可能角色External Partner: Possible Roles for Sam

Sam 的經歷有三塊對這份劇本特別有用：在 ChargeSPOT 擔任營運主管，熟悉硬體共享服務的營運；以 uspace 共同創辦人暨營運長的身分完成過一次成功的創業與出場；以及程式交易背景帶來的資料與模型能力，加上短暫的台積電經驗。他目前以家庭為重，對大量時間投入很敏感，但非常樂意投入少量時間、分享人脈。所以下面的角色都設計成「時間少、槓桿高」，可以只選一個，也可以組合。Three parts of Sam's background are especially useful here: running a shared-hardware service as Head of Operations at ChargeSPOT; a successful venture and exit as co-founder and COO of uspace; and data and modelling skills from algorithmic trading, plus a short stint at TSMC. His focus is family and he is wary of large time commitments, but he is very willing to give small amounts of time and share his network. Every role below is therefore designed for low time and high leverage, and they can be taken singly or combined.

角色選項Role Options

三種組合Three Packages

• 輕量 Light： 策略顧問加人脈引薦，每月約 1 至 2 小時。Strategic advisor plus connector, about 1 to 2 hours a month.
• 標準 Standard： 輕量組合再加關卡審查與營運把關。一年多出約三個半天和幾次會議，是時間和價值比例最好的組合。Light plus gate reviews and operations review. Adds about three half-days and a few meetings a year, the best ratio of value to time.
• 投資型 Investor： 標準組合再加天使投資，顧問股份之外另以現金取得股份，最能對應他想參與一筆可觀報酬的意願。Standard plus an angel investment, with cash shares on top of advisor shares; this best matches his interest in a meaningful financial upside.

顧問股份的參考級距：創業顧問常用的 FAST 範本，在點子階段建議一般顧問 0.25%、加上協助找人的策略顧問 0.5%、再加上人脈與專案參與的專家顧問 1%；公司進入新創階段後分別降為 0.2%、0.4%、0.8%。以 Sam 的經歷與可提供的人脈，標準組合大約落在專家級。實際比例要和金主一起談。Advisor equity benchmark: the widely used FAST template suggests, at idea stage, 0.25% for a standard advisor, 0.5% for a strategic advisor who also helps recruit, and 1% for an expert advisor who adds contacts and projects; at startup stage these fall to 0.2%, 0.4% and 0.8%. Given Sam's background and network, the Standard package sits around the expert level. The actual figure should be agreed with the investor.

不建議的角色Roles We Do Not Recommend

掛名共同創辦人、董事、監察人或兼任營運長。這些角色需要的時間與法律責任超出他想要的投入，也會讓他在 ChargeSPOT 的職位更容易引起利益衝突的疑慮。若之後他的投入意願改變，可以再談。Named co-founder, director, supervisor or part-time COO. These carry more time and legal responsibility than he wants, and make a conflict of interest with his ChargeSPOT role more likely to be questioned. They can be revisited if his appetite changes.

開始前要先處理的事To Settle Before Starting

1. ChargeSPOT 的契約與觀感 ChargeSPOT contracts and optics： 營運主管和 BD 經理兩位現職主管同時參與外部創業，比一個人更敏感。兩人都先確認各自聘僱契約裡的兼職、競業與智財條款，不使用公司時間與資源；在 Rj 離職前，Sam 的參與以不對外具名的顧問形式為宜。Two serving managers in one outside venture draws more attention than one. Both should check the side-work, non-compete and IP clauses in their employment contracts and use no company time or resources; until Rj leaves, Sam is best involved as an unnamed advisor.
2. 股份從哪裡來 Where the shares come from： 顧問股份會稀釋所有股東，要先和金主談好是保留一個顧問股份池，還是從創辦團隊的持股撥出。Advisor shares dilute every shareholder, so agree with the investor whether to keep an advisor pool or carve them out of the founding team's stake.
3. 法律與稅務形式 Legal and tax form： 閉鎖性公司的勞務出資是在章程與發起時規劃。公司設立後才加入的顧問能否以勞務出資，或改用認購新股並綁定給予條件，以及怎麼計稅，請會計師確認。Service-based equity in a closely held company is planned in the articles at incorporation. Whether an advisor joining later can contribute services for shares, or should instead subscribe for new shares tied to vesting conditions, and how this is taxed, needs an accountant's confirmation.
4. 書面協議 Written agreement： 用顧問合約寫清楚期待的投入、保密、智財歸屬、給予條件，以及離開時的處理方式，FAST 範本可以當起點。An advisor agreement should set out expected involvement, confidentiality, IP ownership, vesting, and what happens if he steps away; the FAST template is a good starting point.
5. 先問 Sam 本人 Ask Sam first： 以上是選項，不是預設。建議帶著這張表和他聊，由他選想要的角色與投入。These are options, not assumptions. Take this table to him and let him choose the roles and level of involvement he wants.

來源 Sources：Holloway：常見的新創顧問股份級距（FAST）、Promise Legal：FAST 式顧問合約範本、cpacorey：閉鎖性公司與勞務出資上限

預算與資金Budget and Funding

前 12 個月估計支出約新台幣 498 萬元，第 13 至 18 個月再支出約 632 萬元，其中一部分由募資收入支應。以下金額都是新台幣；美元處以 1 美元約 31 元粗估。「粗估」代表還沒有查到可靠行情，要在詢價後更新。Spending is estimated at about TWD 4.98 million in the first 12 months and a further TWD 6.32 million in months 13 to 18, part of it covered by crowdfunding revenue. All amounts are in TWD; USD items are converted at roughly 31 TWD per USD. "Rough estimate" means no reliable market rate has been found yet; update once quotes arrive.

各階段支出Spending by Phase

幾個有依據的數字Figures with a source:

• 鏡頭 Camera： Tapo C210 定價約 65 美元。The Tapo C210 lists at about USD 65.
• 鋁模 Aluminium mould： 單穴鋁模約 2,800 至 9,000 美元，適合 500 至 5,000 組的產量。A single-cavity aluminium mould costs about USD 2,800 to 9,000 and suits runs of 500 to 5,000.
• NCC： 藍牙測試費約 3 萬至 6 萬元，審驗費每件 10,300 元。NCC: Bluetooth testing costs about TWD 30,000 to 60,000, plus TWD 10,300 per application.
• 藍牙聯盟產品認證 Bluetooth SIG qualification： 2026 年 3 月起每項產品 6,000 美元（副會員）或 12,000 美元（採用會員）。From March 2026, USD 6,000 per product for associate members or USD 12,000 for adopter members.
• 首批生產 First production run： 項圈元件約 5.5 至 12 美元，加上電路板、外殼與組裝；鏡頭以公版或成品採購。每組約 2,200 元是粗估，要等代工報價。Collar parts about USD 5.50 to 12 plus board, housing and assembly; the camera sourced as a reference board or finished unit. About TWD 2,200 per set is a rough estimate pending quotes.

金主的資金分批Investor Tranches

每一筆都綁定一道關卡，並加約一成預備金。Each tranche is tied to a gate, with roughly 10% contingency.

第四筆的計算假設嘖嘖募得 300 萬元，扣掉 8% 平台費後淨入約 276 萬元。這個目標超過前面找到的所有台灣寵物科技案例，偏積極。如果只募到 100 萬元，第四筆要增加約 180 萬元，或者首批生產減半。The fourth tranche assumes the zeczec campaign raises TWD 3 million, about TWD 2.76 million net of the 8% fee. That beats every Taiwan pet-tech campaign found above, so it is ambitious. If only TWD 1 million is raised, the fourth tranche grows by about TWD 1.8 million, or the first run is halved.

第 13 個月起的每月營運成本Monthly Running Cost from Month 13

全職後的固定成本約每月 47 萬元：四位全職人員約 43 萬元，加上雲端約 4 萬元。這也是「金主承諾至少 12 至 18 個月薪資」這個離職條件背後的數字：一年約 560 萬元。After going full-time, fixed costs are about TWD 470,000 a month: about TWD 430,000 for four full-time staff and about TWD 40,000 for cloud. That is the number behind the resignation condition that the investor commits at least 12 to 18 months of salaries: about TWD 5.6 million a year.

群眾募資路徑Crowdfunding Path

建議先在嘖嘖募資，預計 2027 年 11 月上線；目標設低保證達標，實際以 100 萬元為基本情境、300 萬元為努力目標。海外平台留到第二年。Start on zeczec, going live in November 2027. Set a low headline goal to be sure of success, plan around TWD 1 million as the base case and aim for TWD 3 million. Overseas platforms wait until year two.

平台比較Platform Comparison

募資數字怎麼來Where the Numbers Come From

• 假設平均贊助金額約 5,000 元（一組項圈加鏡頭，含數個月訂閱）。Assume an average pledge of about TWD 5,000 (one collar-plus-camera set with a few months of subscription).
• 100 萬元需要約 200 位贊助人，300 萬元需要約 600 位。TWD 1 million needs about 200 backers; TWD 3 million about 600.
• 對照：台灣寵物科技案例中，贊助人數最多的是 KittySpring 的 915 人，但它的單價只有 1,000 多元。For comparison, the most-backed Taiwan pet-tech campaign was KittySpring with 915 backers, but at a price of only about TWD 1,000.

方案設計Reward Design

• 多貓家庭組合 Multi-cat bundles： 一台鏡頭加兩到三個項圈，打中核心客群，也拉高客單價。One camera with two or three collars targets the core customer and lifts order value.
• 綁定訂閱 Bundled subscription： 參考 PETUS 的做法，用 3、6、12 個月訂閱區分方案，提前驗證訂閱意願。Like PETUS, tier rewards by 3, 6 and 12 months of subscription to test willingness to subscribe early.
• 開箱分身 Twin before unboxing： 贊助後就能先上傳貓咪照片、拿到風格化分身，讓贊助人在等出貨時就開始分享，這也是募資期間的擴散工具。Backers upload cat photos right after pledging and receive a stylised twin, so they share while waiting for delivery; this doubles as a campaign growth tool.

上線前的準備Before Launch

1. 等待名單 Waiting list： 從第 2 階段開始由創辦團隊累積，上線前至少要有數百人。The founding team builds it from Phase 2, aiming for at least several hundred people by launch.
2. 真實試用影片 Real beta footage： 30 至 50 戶試用家庭的片段，就是最有說服力的募資內容。Clips from the 30 to 50 beta households are the most persuasive campaign content.
3. KOL 預熱 KOL warm-up： 上線前一個月開始發布試用內容。Start publishing trial content a month before launch.
4. 認證先行 Certification first： NCC 證書最好在上線前拿到，避免出貨延誤。嘖嘖的四階段審核也會看執行能力。Get the NCC certificate before launch to avoid shipping delays; zeczec's four-stage review also checks execution capability.

台灣募資的真正價值是三件事：第一批付費用戶、可以拿去海外證明需求的數據，以及通路總部要求的「試賣成績」。The real value of a Taiwan campaign is threefold: the first paying users, data that proves demand to overseas partners, and the trial-sale record retail chains ask for.

海外擴張Overseas Expansion

建議順序是：2028 年下半年先做香港與新加坡的小規模測試，同時準備日本；2029 年再進美國。越後面的市場越大，但競爭、認證與行銷成本也越高。這是金主人脈最能發揮的一段。The suggested order: a small test in Hong Kong and Singapore in the second half of 2028 while preparing Japan, then the US in 2029. Each later market is bigger, but so are competition, certification and marketing costs. This is where the investor's network matters most.

市場規模Market Size

美國多貓家庭的成長，和我們在台灣看到的趨勢一致，讓「多貓辨識」這個差異點可以直接帶過去。The growth of multi-cat homes in the US matches what we see in Taiwan, so multi-cat identification carries straight over as a point of difference.

各市場的做法Approach by Market

香港與新加坡（2028 年下半年） 中文內容可以直接沿用，可以透過嘖嘖的新加坡平台做跨境募資，是成本最低的第一個海外測試。市場規模與認證規定還沒查。Hong Kong and Singapore (second half of 2028). Chinese content carries over as is, and zeczec's Singapore platform allows cross-border crowdfunding, making this the cheapest first overseas test. Market size and certification rules are not yet researched.

日本（2028 年下半年準備，2029 年初上線） 貓文化強、距離近，可以用 Makuake 進場。要注意日本是 Catlog 的主場，我們的差異要放在「影像精華加數位分身」，而不是健康數據。需要日文化的 App 與內容、符合日本無線電法的認證，以及日本當地的經銷夥伴。Japan (prepare in the second half of 2028, launch early 2029). Strong cat culture, close by, with Makuake as an entry route. Japan is Catlog's home market, so our difference must be video highlights plus the digital twin, not health data. Needs a Japanese app and content, Japanese radio certification and a local distribution partner.

美國（2029 年） 市場最大，但競爭最密，Furbo、Petcube、Whisker、Petlibro 都在。建議用 Kickstarter 加亞馬遜進場，需要比亞洲市場高很多的行銷預算。United States (2029). The largest market but the most crowded, with Furbo, Petcube, Whisker and Petlibro. Enter through Kickstarter plus Amazon, with a far larger marketing budget than in Asia.

認證Certification

硬體設計時就要想到海外： 選用已取得各國認證的藍牙模組，可以大幅降低認證成本與時間，這是第 2 階段就要和代工廠談清楚的事。Design the hardware with overseas markets in mind: a Bluetooth module already certified in each market sharply cuts certification cost and time, and needs settling with the manufacturer in Phase 2.

金主在海外擴張的角色The Investor's Role Overseas

1. 經銷與通路人脈 Distribution and channels： 日本、美國的寵物用品經銷商與零售通路。Pet-product distributors and retailers in Japan and the US.
2. 第二輪募資 Second funding round： 主導或引薦投資人，金額依市場進場計畫而定。Lead or introduce investors, sized to the market-entry plan.
3. 跨境物流與當地設立 Cross-border logistics and local set-up： 倉儲、退換貨、是否設立當地公司。Warehousing, returns, and whether to set up local entities.
4. 公關與媒體 PR and media： 當地媒體與寵物圈的引薦。Introductions to local media and the pet community.

進入海外前要先有的數字Numbers Needed Before Going Overseas

• 台灣用戶的留存率與訂閱轉換率Retention and subscription conversion among Taiwan users
• 每位用戶的雲端成本與毛利Cloud cost and gross margin per user
• 多語言 App 與客服流程A multilingual app and support process

各市場的進場預算還沒有可靠依據，建議在第 4 階段，用台灣的實際獲客成本與留存數字去推算。There is no reliable basis yet for each market's entry budget; estimate it in Phase 4 from Taiwan's actual acquisition cost and retention.

政府補助與其他資金來源Government Grants and Other Funding

結論先講：政府資源可以補一部分研發費用，但不能取代金主。第一年真正對得上時程的只有 SBIR 先期研究，實際可期待約 50 萬到 110 萬元；國發基金天使投資方案適合搭配第二輪募資；其他是金額較小的補貼。接受補助換來的是查核、計畫內容被鎖定與資訊公開，這些會和本劇本的「關卡隨時可停」、「兼職期不具名」、「深圳代工」與「香港先行」產生摩擦，需要事先安排。Bottom line first: government money can cover part of R&D, but it cannot replace the investor. In year one only SBIR Phase 1 fits the schedule, realistically worth about TWD 0.5 to 1.1 million; the National Development Fund angel scheme fits the second funding round; the rest are small subsidies. In exchange, a grant brings audits, a locked project scope and public disclosure, which rub against this playbook's stop-at-any-gate design, the unnamed part-time phase, Shenzhen manufacturing and the Hong Kong-first overseas plan. These need arranging in advance.

可用資源一覽Options at a Glance

SBIR 實際能拿到多少How Much SBIR Could Realistically Provide

第 2 階段預算 251 萬元中，和研發相關的約 226 萬元：技術負責人 56 萬、韌體外包 40 萬、App 工程師 33 萬、打樣 50 萬、鋁模 20 萬、動畫 21 萬、雲端 6 萬；品牌識別 25 萬不算研發。Of Phase 2's TWD 2.51 million budget, about 2.26 million is R&D-related: technical lead 560,000, firmware contract 400,000, app engineer 330,000, prototyping 500,000, mould 200,000, animation 210,000 and cloud 60,000. Brand identity (250,000) does not count as R&D.

• 最好的情況 Best case： 全部可列支，補助上限約 113 萬元。Everything is eligible, capping the grant at about TWD 1.13 million.
• 深圳外包不能列支時 If Shenzhen contracting is ineligible： 扣掉韌體外包與打樣 90 萬元，可列支約 136 萬元，補助上限約 68 萬元。Removing the 900,000 of firmware and prototyping leaves about 1.36 million eligible and a cap of about 680,000.
• 核定刪減 Cuts at approval： 一篇 2026 年申請教學指出核定金額可能被刪減 10% 至 50%，查核點卻不會跟著減少。A 2026 application guide notes awards may be cut by 10% to 50% while milestones stay the same.

所以合理的規劃金額是 50 萬到 110 萬元，約是第二筆資金 280 萬元的二到四成。而且它是先墊後補：已查到的地方型與潛力新創規定，都是簽約時撥一半、結案審查通過後再撥另一半，中央型的細節以契約為準。金主仍然要先把第二筆全額準備好，補助款是事後回補。A sensible planning figure is therefore TWD 0.5 to 1.1 million, about 20% to 40% of the 2.8 million second tranche. It also arrives in arrears: the local and early-stage SBIR rules we found pay half at signing and half after the final review, and the central programme's terms are set in the contract. The investor still has to have the full second tranche ready; the grant reimburses part of it later.

要讓它可行的前期部署Groundwork Needed to Make It Work

建議時程The Suggested Sequence

1. 2026 年 11 月 November 2026： 設立公司、資本到位、確認非陸資。Incorporate, pay in capital, confirm there is no mainland Chinese capital.
2. 2026 年 12 月至 2027 年 1 月 December 2026 to January 2027： 準備簡報並送件，研發主題是項圈感測、影像精華辨識與數位分身同步。Prepare the pitch and apply; the R&D themes are collar sensing, video-highlight recognition and digital-twin sync.
3. 2027 年 2 至 3 月 February to March 2027： 書面與簡報審查。一篇 2026 年申請教學估計，從投件到第一筆款項約 4 至 6 個月，所以也可能延到 5 至 7 月。Written and pitch review. One 2026 guide estimates 4 to 6 months from submission to first payment, so it may slip to May to July.
4. 2027 年 3 月底關卡一 Gate 1, end of March 2027： 通過才簽約；未通過就不簽。核定後不簽約怎麼處理，送件前先向計畫辦公室確認。Sign only if the gate is passed. Check with the programme office beforehand what happens if an approved award is not signed.
5. 2027 年 4 月起 From April 2027： 執行，對齊第 2 階段；結案後請領尾款，之後配合 3 年成效追蹤。Carry out the project alongside Phase 2; claim the balance after the final review, then report results for three years.

接受補助要承擔的事What Accepting a Grant Commits Us To

1. 自籌一半、先墊後補 Matching funds and payment in arrears： 補助款最多是總經費的一半，超支不補，沒花完的要繳回。The grant covers at most half the cost; overspending is not covered and underspending is returned.
2. 計畫內容被鎖定 A locked scope： 與核定內容不符就要辦變更，原則是「目標不減、經費不增」。這和本劇本「關卡未過就調整或停止」的設計有衝突：中途停止可能被認定違約，面臨解約、追繳補助款，且 5 年內不得再申請。Any departure from the approved plan needs a formal change, on the principle that targets cannot shrink and funding cannot grow. This conflicts with the playbook's adjust-or-stop gates: stopping midway may count as breach, with termination, repayment and a five-year ban.
3. 行政負擔 Administrative load： 期中報告、期中實地審查、會計師查核、結案報告與簡報，結案後 3 年每年回報成效。對兼職中的創辦團隊是實際的時間成本。Interim report, on-site interim review, accounting review, final report and presentation, then three years of annual results reporting. For a part-time founding team this is real time.
4. 資訊公開 Disclosure： 受補助名單與計畫名稱可能公開，研發人員要有投保紀錄。Grant recipients and project titles may be public, and R&D staff need insurance records.
5. 政府的實施權 Government licence： 研發成果歸公司所有，但政府可為研究目的無償、非專屬實施，不影響我們自己商業化。Results belong to the company, though the government may use them non-exclusively and free of charge for research; our own commercial use is unaffected.
6. 不能重複補助 No double funding： 同一或類似計畫不能再向其他政府單位申請。The same or a similar project cannot be funded by another government body.

風險Risks

對金主的影響Impact on the Investor

• 好處 Benefits： 補助不稀釋股權，等於事後退回第二筆的二到四成；通過政府審查也是對外的背書。Grants do not dilute anyone and effectively refund 20% to 40% of the second tranche later; passing a government review is also a credential.
• 資金形式 Form of funding： 因為補助款不超過實收資本額，金主的錢要以股款進公司，用借款或可轉債的形式會讓資本額太小、補助上限跟著變小。這會影響之前討論的投資工具選擇。Because the grant cannot exceed paid-in capital, the investor's money should go in as share capital; loans or convertible bonds keep capital small and shrink the grant ceiling. This bears on the funding instrument discussed earlier.
• 背景查核 Background check： 金主本人與其公司如果有陸資，整個公司就不能申請。Any mainland Chinese capital in the investor or his companies disqualifies the whole company.
• 國發基金進場時 If the National Development Fund comes in： 會稀釋金主持股；金主與現有股東不能當搭配天使，必須找外部天使；官股進來後要定期提供財務與營運資訊，董事會與股東會要事前通知、會後提供議事錄。It dilutes the investor; neither he nor existing shareholders can be the matching angel, so an outside angel is needed; afterwards the company must report financial and operating information regularly and notify the Fund before board and shareholder meetings, sending minutes afterwards.
• 個人稅務 Personal tax： 金主若以個人名義投資，可能適用天使投資租稅優惠，由他的會計師判斷。Investing personally may qualify him for angel tax relief, for his accountant to assess.
• Sam 的角色 Sam's role： 如果 Sam 拿顧問股份或顧問報酬，他也不能當國發基金的搭配天使。If Sam takes advisor shares or fees, he cannot be the Fund's matching angel either.

對海外發展的影響Impact on Overseas Expansion

助力 Help： SBIR 自 2026 年起增設專利申請費補助，支持同步佈局國內外專利；展覽補助可以補貼日本與美國的參展；政府計畫紀錄對海外通路與投資人是加分。SBIR added patent-filing support from 2026, covering domestic and foreign patents; trade-show subsidies can offset shows in Japan and the US; a government-programme record helps with overseas channels and investors.

限制 Constraints：

1. 香港與大陸 Hong Kong and mainland China： 經濟部的研發成果運用辦法規定，運用對象或區域在大陸、香港、澳門時需報經濟部核准。這是否適用於 SBIR 要在簽約前確認。單純出口賣產品一般不等於移轉研發成果，但授權當地夥伴製造、或把設計與韌體交給深圳廠商，就可能觸及。The Ministry's rules on using R&D results require approval when the recipient or territory is mainland China, Hong Kong or Macau. Whether this applies to SBIR must be confirmed before signing. Simply exporting products is generally not a transfer of results, but licensing a local partner to manufacture, or handing designs and firmware to a Shenzhen manufacturer, might be.
2. 海外架構調整 Offshore restructuring： 美國或日本投資人可能要求把公司放到境外控股公司之下。結案後 3 年成效追蹤期間把研發成果移出台灣公司，可能受到檢視；國發基金若已是股東，架構調整也要取得其同意。這兩點是推論，需律師確認。US or Japanese investors may want the company under an offshore holding company. Moving R&D results out of the Taiwan company during the three-year follow-up may draw scrutiny, and if the Fund is a shareholder its consent would be needed. Both points are inferences for a lawyer to confirm.
3. 投資人背景 Investor backgrounds： 第二輪若有中資背景的投資人（包括經由香港的資金），可能讓公司被認定為陸資而失去補助資格；國發基金也表示政府投資與中資不宜同時成為股東。Second-round investors with Chinese backing, including money routed through Hong Kong, could make the company count as mainland-funded and lose grant eligibility; the Fund also says government money and Chinese capital should not sit on the same register.

建議Recommendation

• 第一年 Year one： 只申請 SBIR 先期研究，以 50 萬到 110 萬元規劃，關卡一之後才簽約，預算不依賴它。Apply only for SBIR Phase 1, plan on TWD 0.5 to 1.1 million, sign only after Gate 1, and never make the budget depend on it.
• 第二年 Year two： 全職後再評估 SBIR 研究開發（做海外版產品），以及搭配第二輪募資申請國發基金天使方案。After going full-time, consider SBIR Phase 2 for the overseas product and the Fund's angel scheme alongside the second round.
• 不建議 Not recommended： 第一年就申請大型多年期補助，或為了補助改變代工與海外市場的順序。Large multi-year grants in year one, or changing the manufacturing route or market order just to qualify.

來源 Sources：旺得富：SBIR 2026 年起提高補助額度、2026 年 SBIR 新制與舊制比較、恒遠數位：SBIR 申請流程 2026、中小及新創企業署：SBIR 潛力新創研發補助申請須知、臺中市 SBIR 管考說明會簡報、經濟部研發成果運用問題彙總、國發基金創業天使投資方案常見問題、國發基金天使投資加碼至百億（經濟日報）、臺北市貿易推廣補助 2026、臺北市產業發展研發補助申請須知、KPMG：產創條例修正稅務快訊、PRO360：SBIR 顧問費用行情、yours.net：SBIR 資格與加值應用規定

風險與關卡判準Risks and Gate Criteria

關卡判準（數字待第 0 階段與金主訂定）Gate Criteria (thresholds to be set with the investor in Phase 0)

主要風險Main Risks

以上是規劃框架，不是法律或財務建議。股東協議、勞務出資與稅務請律師與會計師確認。This is a planning framework, not legal or financial advice. Have a lawyer and an accountant confirm the shareholder agreement, service-based equity and tax treatment.

資料來源Sources

市場Market

• 農傳媒：農業部 114 年家犬家貓飼養數量調查
• 經濟日報：寵物市場 3 年內拚破千億（台經院）
• 中央社：2025 年新生兒 10 萬 7812 人
• 日本寵物食品協會 2025 年犬貓飼養調查（いきもののわ整理）
• Pet Food Industry：美國養貓家庭增至 5,300 萬
• APPA：美國寵物產業 2025 年 1,580 億美元

群眾募資Crowdfunding

• adHub：嘖嘖抽成與流程
• 數位時代：嘖嘖累計破 200 億、進軍新加坡
• 嘖嘖專案頁：KittySpring、PETUS、Pawbby、amicoipet
• Makuake：收費說明
• Fundpop：Kickstarter 費用（2026）

認證與製造Certification and Manufacturing

• 安盛國際驗証：NCC 認證費用
• MarkReady：FCC 認證費用（2026）
• 無線コネクト：2026 年 3 月起藍牙認證費調漲
• Xinyang：新創小量射出模具成本（2026）
• Tapo C210 評測（定價）

人力與補助Staffing and Grants

• GetCareers：台灣職稱薪資行情（依主計總處 2025 年調查）
• faqs.tw：動畫師月薪 2026
• 聯合新聞網：SBIR 補助上限調高