Overview
When an e-commerce seller is considering entering the subscription business, rather than blindly starting with a "monthly flat-rate model," 5-question self-diagnosis + 4 model classification + Korea market fit this is a dedicated skill for designing in stages.
One-line responsibility: Input category + average purchase cycle + average order value + current revenue → automatically generate answers to 5 questions + recommended fit model + lock-in and churn prevention strategy.
5-Question Self-Diagnosis
| # | Question | Review Point |
|---|---|---|
| Q1 | Is consumption frequency sufficient? | Only suitable for categories consumed at least once within 30 days |
| Q2 | Is the price predictable? | Stable price with no monthly volatility (KRW 5,000–50,000 range) |
| Q3 | Is there lock-in value? | Signup cost ≤ first-month savings effect (clear payback point) |
| Q4 | Is there a churn prevention mechanism? | Auto-renewal, win-back during dormancy, personalized recommendation infrastructure |
| Q5 | Has the market been validated? | 3+ subscription cases in the same category in the Korea market |
All 5 YES: Enter subscription model design immediately / 4 or fewer: Strengthen single-item sales, then re-evaluate
4 Subscription Model Classification
Model 1: Consumer Goods (consumption timing motivation)
- Monthly automatic delivery — reliable supply is the value
- Korea cases: Oi6, SSG Dawn Delivery, Market Kurly Dawn Delivery, Coupang Rocket Fresh
- Suitable categories: food, household goods, pet food, diapers
Model 2: Experience (experience diversity)
- New curation every month — discovery and surprise are the value
- Korea cases: VIPS membership, Collins curation, Aritaum member
- Suitable categories: cosmetics, perfume, snacks, books
Model 3: Relationship (membership + benefits)
- VIP membership — rights and access beyond price discounts are the value
- Korea cases: Coupang Rocket Wow, SSG Universe, Naver Membership
- Suitable categories: marketplace, e-commerce platform
Model 4: Personalized (personalization data)
- 1:1 personalized recommendations — accumulated data is the value (leaving means data loss)
- Korea cases: Pillly, Hwahae personalized supplements, personalized cosmetics
- Suitable categories: supplements, cosmetics, diet meal plans
Lock-in vs Churn Prevention Message Matrix
| Stage | Message Strategy | Korea Case |
|---|---|---|
| Signup | First month free or trial + strong value promise | Coupang Rocket Wow 1 month free |
| Onboarding (D+7) | First delivery review + next delivery preview | Friendly message after Oi6's first box |
| Retention (M+1-3) | Value reminder · add personalized recommendations | Collins monthly curation story |
| Dormancy signal | Value re-reminder + dormancy option (no cancellation) | Market Kurly pause option |
| Just before churn | 1:1 message + strong recovery | "Are you really leaving? 1:1 consulting" |
| After churn | Winback only once after 30 days | Restore first signup price |
Korea market suitability diagnosis
[Step 1] 카테고리에 한국 구독 사례 3개+ 있는가?
└─ NO → 사례 0~2개라면 시장 미성숙. 단품 판매 우선.
[Step 2] 한국 소비자 정기 구매 패턴 있는가?
└─ NO → "필요할 때마다 사는" 행동 패턴이면 구독 부적합
[Step 3] 락인 비용 > 첫 달 가치인가?
└─ NO → 락인 비용이 크면 가입 자체 발생 안 함
[Step 4] 이탈 시 데이터 손실 있는가?
└─ NO → 데이터 손실 없으면 이탈 장벽 약함 (맞춤 모델 제외)
Usage example
"/commerce-subscription-strategist — 비건 영양제, 30일 주기, 월 5만원, 신생 D2C"
→ 5가지 질문 답변 + 모델 4 맞춤 추천 + 락인 시퀀스
Follow-up chain (text output review)
This skill's 6-stage copy for the signup·onboarding·retention·dormancy·churn message matrix is moai-coworker:ai-slop-revieweris automatically post-processed and chained with. Subscription messages directly influence user decisions such as new signups, churn prevention, and winback, so AI slop removal is essential.
commerce-subscription-strategist → ai-slop-reviewer → (선택) cs-channel-message
Related skills
commerce-repurchase-timer— recurring purchase category (stage just before subscription)commerce-ltv-cac-architect— subscription LTV calculationcommerce-integrated-strategy— macro channel·business strategy (pair)ai-slop-reviewer— message matrix AI slop review (auto chain, HARD)moai-writer:korean-humanize— Korean AI tell removal (after slop review, required)