← 返回 Skills 市场
277
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install ltv-loyalty-winback
功能描述
Predict repeat-customer churn from purchase and tier data, then trigger branch-specific win-back workflows with VIP care or time-limited offers.
使用说明 (SKILL.md)
LTV loyalty churn alert & win-back
You connect historical purchase frequency (and recency) → silence / churn risk → automated activation workflows (journeys, messages, and ops hooks).
When NOT to use (should-not-trigger)
- Single purchase confirmation only ("thanks for your order") with no retention ask.
- Pure transactional templates with no lapse or segment context.
If the user later adds "many haven't bought again," switch to full workflow mode.
Gather context
- Recency: last purchase date distribution; spike in 90d+ silent (or merchant’s window).
- Frequency: orders per customer in last 12 months; baseline vs cohort.
- Monetary: AOV and cumulative spend — VIP threshold (define or ask).
- Tier / points: tier rules, points expiry date and notice rules.
- Channel: email, SMS, app push, CS outreach — constraints and consent.
For RFM-style cutoffs and workflow templates, read references/loyalty_playbook.md when needed.
Core workflow (always)
- Risk framing — Classify silence (e.g. at-risk: no order in 90d; high-risk: 120d+; formerly frequent now cold).
- Segment branch
- VIP / high-value → Dedicated CS care (named contact, proactive outreach, perk without trashing margin — early access, bundle concierge, not only generic % off).
- Standard buyer → Time-bound discount or incentive bait (deadline, code, stack rules clear).
- Activation workflow — Trigger, wait steps, exit rules, measure (reactivation rate, incremental revenue).
Mandatory outputs (full engagement)
Every full churn / loyalty answer must include:
- Risk summary — Who is silent, how many (or %), vs historical norm.
- Two branches — At least one block labeled VIP / high-value (CS care) and one Standard (discount bait), even if one segment is empty ("if no VIPs, still document threshold").
- Win-back plan table — concrete steps + channel + timing.
- Points expiry branch (if applicable) — Pre-expiry nudge sequence before burn.
Win-back plan table (required shape)
| Segment | Silence signal | Action | Channel | Timing |
|---|---|---|---|---|
| VIP | e.g. 90d + LTV top decile | Dedicated CS check-in + perk | Email + optional call | Day 0, Day 7 |
| Standard | e.g. 90d no order | Limited-time offer | Email/SMS | Day 0, Day 3, Day 10 |
| Points at risk | Expiry in 14d | Redeem nudge + basket builder | T-14, T-7, T-2 |
Adjust rows to data; keep VIP vs Standard distinction explicit.
VIP branch (content requirements)
- Named or role-based concierge line.
- Non-price-first value where possible (early access, restock hold, bundle suggestion).
- Escalation path if no response.
Standard branch (content requirements)
- Time-bound offer (end date).
- Clear code or auto-apply and exclusions.
- Single primary CTA.
Success metrics (suggest in every full run)
- Reactivation rate (ordered within 30d of campaign).
- Incremental revenue vs holdout.
- Unsubscribe / complaint rate (guardrail).
Split with other skills
- Affiliate reconciliation → not this skill.
- Abandoned checkout (session drop) → checkout skill; this skill is post-purchase lapse.
安全使用建议
This package is an advisory playbook (templates, segmentation rules, workflow tables) and contains no code or installers — it won't itself run or send campaigns. Before installing or using it, confirm: (1) whether the skill is expected to actually trigger messages/workflows or only to produce plans; if it will trigger anything, ask the author which platforms/APIs it uses and what credentials will be required; (2) only provide the minimal customer data needed (aggregate counts or hashed identifiers when possible) and ensure you have consent and proper PII handling; (3) if you plan to let an agent autonomously invoke this skill and connect it to your ESP/CRM, restrict the agent’s credentials and outbound permissions (or require manual approval) so campaigns are not sent without human review. If the author adds scripts or integration code later, re‑evaluate the skill for install risks and requested environment variables.
功能分析
Type: OpenClaw Skill
Name: ltv-loyalty-winback
Version: 0.1.0
The skill bundle is a legitimate marketing automation tool designed to identify customer churn risk and generate win-back strategies. The instructions in SKILL.md and the logic in references/loyalty_playbook.md are strictly focused on CRM workflows, segmenting users into VIP and standard categories, and suggesting retention campaigns based on purchase history. There are no signs of data exfiltration, malicious execution, or prompt injection attacks.
能力评估
Purpose & Capability
The name/description (predict churn, branch VIP vs Standard, and trigger win‑back workflows) matches the SKILL.md’s guidance, templates, and required outputs. However the skill also states it will “trigger” activation workflows and ops hooks but includes no install spec, credentials, or integrations. That suggests the skill is advisory (drafting plans) rather than actually executing campaigns; the claim to trigger workflows is ambiguous and should be clarified.
Instruction Scope
SKILL.md stays on‑topic: it instructs how to compute recency/frequency/monetary tiers, branch segments, design workflows, and produce required tables. It references only included local documentation (references/loyalty_playbook.md) and common marketing data (purchase history, tiers, consent). It does not instruct the agent to read unrelated system files, environment variables, or exfiltrate data.
Install Mechanism
Instruction-only skill with no install spec, binaries, or code to run. This is low risk from an install perspective; nothing will be written to disk as part of installing the skill package itself.
Credentials
The skill declares no required environment variables or credentials, which is proportionate for a planning/advisory skill. However, in practical use, executing the workflows it describes would require CRM/ESP/automation platform credentials and possibly access to customer PII or transaction data. The skill does not declare or request those, so responsibility to provide and securely manage integration credentials rests with the agent/user — confirm how/when those will be supplied.
Persistence & Privilege
No elevated privileges requested. always:false and default autonomous invocation are set (normal). The skill does not request to modify other skills or system configuration.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install ltv-loyalty-winback - 安装完成后，直接呼叫该 Skill 的名称或使用
/ltv-loyalty-winback触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v0.1.0
- Initial release of ltv-loyalty-winback for predicting repeat-customer churn risk and delivering targeted win-back workflows.
- Segments customers by value: VIPs receive white-glove CS care, standard buyers get incentive offers.
- Supports scenarios like growing lapsed cohorts, declining tier engagement, points expiry, and RFM/LTV segmentation needs.
- Excludes single-order confirmations and pure transactional messaging.
- Provides mandatory risk summary, explicit VIP vs Standard action branches, and a required win-back plan table.
- Outlines channel strategies, timing, and measurement metrics for reactivation success.
元数据
常见问题
Ltv Loyalty Winback 是什么？
Predict repeat-customer churn from purchase and tier data, then trigger branch-specific win-back workflows with VIP care or time-limited offers. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 277 次。
如何安装 Ltv Loyalty Winback？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install ltv-loyalty-winback」即可一键安装，无需额外配置。
Ltv Loyalty Winback 是免费的吗？
是的，Ltv Loyalty Winback 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Ltv Loyalty Winback 支持哪些平台？
Ltv Loyalty Winback 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Ltv Loyalty Winback？
由 RIJOY-AI（@rijoyai）开发并维护，当前版本 v0.1.0。
推荐 Skills