Overview
Korean D2C sellers in regular purchase categories such as cosmetics, food, pets, lenses, and razors without being buried in the average purchase cycle, by segment and product golden timeis a dedicated skill that calculates and induces repurchases.
One-line responsibility: Input category + average purchase cycle + customer segment + past repurchase data → automatically derive 3-stage golden time (remind, deadline, dormancy entry) + message tone per stage + lead scoring rules + retention chart guide.
vs commerce-ltv-cac-architect: The paired skill is a lifetime LTV/CAC model for one customer. This skill covers only repurchase timing triggersin depth.
3-Stage Golden Time Model
When average purchase cycle = T, split into 3 stages:
[구매 일] ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ [다음 구매 가능 시점]
D+0 T*0.8 T*1.1~1.2 T*1.5
│ │ │ │
│ ① 리마인드 ② 데드라인 ③ 휴면 진입
│ (80% 시점) (110~120%) (150%)
│ 감성·정체성 긴급성·할인 윈백·강한 인센티브
Stage 1: Remind (80% point)
Remind 20% earlier than the average cycle. Not "You haven't run out yet, right?" but identity recall:
| Category | Average cycle T | Remind point (0.8T) | Message tone |
|---|---|---|---|
| Cosmetics (serum) | 60일 | 48일 | "Are you using the serum from last time well?" — emotional |
| Razor | 30일 | 24일 | "Shave this weekend — start with a new blade" |
| Contact lenses | 30일 (1day) | 24일 | "One month done — prepare the next box" |
| Pet food | 30-45일 | 24-36일 | "🐶 Almost time to run out — this box" |
| Supplements | 30일 (1통) | 24일 | "This bottle is almost used up. If you want to continue, now" |
Stage 2: Deadline (110-120%)
When the average cycle has been slightly exceeded. urgency + discount:
| Category | Deadline (1.1T) | Message tone |
|---|---|---|
| Cosmetics | 66일 | "You've run out of the serum from last time, right? Order today for +10%" — clear CTA |
| Razor | 33일 | "New blade within 1 week — free shipping" |
| Contact lenses | 33일 | "This box closes — 30% off the next box" |
Stage 3: Dormancy Entry (150%)
If it exceeds 1.5 times the average cycle, dormancy. winback + strong incentive:
| Category | Dormancy entry (1.5T) | Message tone |
|---|---|---|
| Cosmetics | 90일 | "It's been a while. 30% off your first repurchase in 60 days + free shipping + new product sample" — strong reward |
| Razor | 45일 | "How was the razor last time? Leave a review and get a 50% coupon" — opinion + reward |
| Contact lenses | 45일 | "Time to meet again, at the first box price" (restore first purchase price) |
No response within 60 days of entering dormancy: Review for churn or attempt win-back only once per quarter (email-focused, zero cost).
Message tone matrix by stage
| Stage | 톤 | Channel priority | Incentive intensity |
|---|---|---|---|
| ① Remind (80%) | Evoke emotion and identity | App push > KakaoTalk Channel > Email | 0-5% (low — natural reminder) |
| ② Deadline (110-120%) | Urgency and clear CTA | KakaoTalk Channel > App push > SMS | 10-15% (medium) |
| ③ Dormancy entry (150%) | Win-back and strong rewards | Email > KakaoTalk Channel | 25-40% + gift (strong) |
Lead scoring rules
Assign scores by behavior, then identify purchase signals:
| Behavior | Score | Meaning |
|---|---|---|
| Revisit within 7 days of purchase | +10 | Satisfaction signal |
| 3 or more detail views after purchase | +15 | Strong interest |
| Add to cart (no purchase) | +20 | Immediate send trigger |
| Write a review | +25 | Loyalty signal — VIP candidate |
| Invite or share with friends | +30 | Referrer — Referral channel |
| Open email or push (within 30 days) | +5 | Active member |
| No app access for 60 days | -20 | Dormancy signal |
| Unsubscribe | -100 | Stop sending immediately |
Classify segments by cumulative score total:
- 100+: VIP candidate (5+ purchases + review + referral)
- 50-99: Loyal member (active repurchasing)
- 0-49: Regular member
- Negative: Dormancy and churn risk
Retention chart cohort analysis guide
Track repurchase rate changes by monthly cohort:
| 가입 월 | M+1 | M+2 | M+3 | M+6 | M+12 |
|--------|-----|-----|-----|-----|------|
| 2026-01 | 45% | 32% | 25% | 18% | 12% |
| 2026-02 | 48% | 35% | 27% | 20% | - |
| 2026-03 | 52% | 38% | 30% | - | - |
| 2026-04 | 55% | 40% | - | - | - |
| 2026-05 | 58% | - | - | - | - |
Interpretation criteria:
- Cohort M+1 at 40% or below → insufficient onboarding sequence (reinforce cs-channel-message Activation)
- M+3 at 25% or below → golden time trigger not working (this skill needs to be introduced)
- M+12 at 10% or below → win-back failed after dormancy entry (review loyalty program)
Improvement metrics (linked with vault-ecom + ltv-cac-architect):
- After introducing this skill, target M+3 cohort repurchase rate from 25% → 35%+
- When repurchase rate reaches 35%+, LTV/CAC ratio automatically rises 1.5x
Golden time standard values by category
| Category | Average cycle T | Remind 0.8T | Deadline 1.1T | Dormancy 1.5T |
|---|---|---|---|---|
| Cosmetics (serum, cream) | 60일 | 48일 | 66일 | 90일 |
| Cosmetics (sheet mask) | 45일 | 36일 | 49일 | 67일 |
| Razor | 30일 | 24일 | 33일 | 45일 |
| Contact lenses 1day | 30일 | 24일 | 33일 | 45일 |
| Contact lenses monthly | 30일 | 24일 | 33일 | 45일 |
| Pet food | 30-45일 | 24-36일 | 33-50일 | 45-67일 |
| Supplements (1 bottle) | 30일 | 24일 | 33일 | 45일 |
| Regular food (oatmeal, etc.) | 30-60일 | 24-48일 | 33-66일 | 45-90일 |
| Ink cartridge | 60-90일 | 48-72일 | 66-99일 | 90-135일 |
| Perfume (50ml) | 90-120일 | 72-96일 | 99-132일 | 135-180일 |
Individual seller measurement required: The table above is a general standard. Measure the actual cycle from your own data and apply it. A ±20% difference by category is typical.
Workflow
Input slots
| Item | Required | Example |
|---|---|---|
| Category | Required | Cosmetics serums, razors, contact lenses, etc. |
| Average purchase cycle (T) | Required | 60 days (or undecided → apply category standard) |
| Customer segment | Required | First purchase / 2nd purchase / regular (3+) / VIP (5+) |
| Past repurchase data | Recommended | Cohort repurchase rate or simple average |
| Channel availability | Recommended | App push, KakaoTalk FriendTalk, email, SMS opt-in status |
Calculation order
[Step 1] 평균 주기 T 확정 (입력 또는 카테고리 표준)
[Step 2] 3구간 골든타임 계산 (0.8T, 1.1T, 1.5T)
[Step 3] 세그먼트별 차등 적용
- 첫구매: 모든 3구간 활용
- 2회구매: 리마인드 + 데드라인만
- 단골: 리마인드만 (인센티브 약함)
- VIP: 휴면 진입 시점 트리거만 (휴면율 낮음)
[Step 4] 구간별 메시지 톤 + 채널 매핑
[Step 5] 리드 스코어링 룰 적용
[Step 6] 리텐션 차트 cohort 분석 가이드 출력
ai-slop-reviewer automatic chaining
After calculating the golden time message, moai-coworker:ai-slop-reviewer automatically called:
- AI cliché removal
- Natural tone by category (cosmetics = emotional / razors = practical / pets = emoji + friendly)
- Verification of the sincerity of incentive expressions
Immediately after slop review, moai-writer:korean-humanizeremoves the Korean AI feel (after slop review, required).
Usage example
"/commerce-repurchase-timer — 화장품 세럼, T=60일, 첫구매 세그먼트"
→ 리마인드 48일: "지난번 세럼 잘 쓰고 계신가요?" + 앱 푸시
데드라인 66일: "오늘 주문 시 +10%" + 카카오 친구톡
휴면 90일: "60일 만의 재구매에 30% + 무료 배송 + 신상 샘플" + 이메일
리드 스코어링 룰 + cohort M+3 25%→35% 목표
"/commerce-repurchase-timer — 면도기, T 미정, 카테고리 표준 적용"
→ T=30일 기본값. 리마인드 24일 / 데드라인 33일 / 휴면 45일
"/commerce-repurchase-timer — 반려동물 사료, T=35일, VIP 세그먼트만"
→ VIP 휴면 진입 시점(53일)만 트리거. 리마인드·데드라인 생략 (VIP 재구매율 80%+)
Quality checklist
- Confirm average cycle T (input or apply standard)
- Calculate all 3 golden time windows (0.8T, 1.1T, 1.5T)
- Specify message tone + channel priority per window
- Differentiate by segment (first purchase, 2nd, regular, VIP)
- Lead scoring rules (at least 8 behavioral scores)
- 1 retention chart cohort analysis guide
- ai-slop-reviewer review trace
Related skills
Chaining:
commerce-repurchase-timer (timing calculation) → moai-cs:cs-channel-message (app push, KakaoTalk, email copy) → commerce-message-compliance-kr (legal gate) → actual send
commerce-ltv-cac-architect— Full LTV/CAC model (pair, different owner)cs-channel-message— App push, KakaoTalk, email copy per window (including app push planning mode)commerce-message-compliance-kr— Network Act gate before sending
When not to use this skill
- Full LTV/CAC analysis:
commerce-ltv-cac-architectuse - New customer acquisition: This skill is for repurchase only. For new customers,
marketing-campaign-planneruse - Simple message copywriting:
cs-channel-messageuse (including app push planning mode) - First purchase onboarding:
cs-channel-messageUse the Activation stage
Reference materials
- Vault notes: "Data marketing that creates repurchases" + "ROAS→LTV CRM revenue design method"
- Standard cycles for Korean D2C subscription purchase categories (cosmetics, razors, lenses, pet food category statistics)