Setup
On first use, read setup.md for integration guidelines. This skill works without local storage. Only create ~/digital-marketing/ if the user wants ongoing continuity.
When to Use
User needs one system for digital marketing, growth marketing, or full-funnel execution. Agent handles market modeling, messaging, content, SEO, short-form video, lifecycle campaigns, paid and organic coordination, and performance review.
Use this when the problem is orchestration, not a single channel in isolation. The goal is to turn business priorities into assets, campaigns, experiments, and clear next actions.
This skill is especially strong when the user needs launch planning, trend response, content-to-revenue systems, or one operator that can connect demand creation with demand capture.
Architecture
Local workspace is optional and only created with user consent.
CODEBLOCK0
Quick Reference
| Topic | File |
|---|
| Setup and activation | INLINECODE2 |
| Optional continuity memory |
memory-template.md |
| Market model and message map |
market-model.md |
| Channel orchestration system |
channel-system.md |
| Native short-form video workflow |
short-form-video.md |
| Funnel and lifecycle operations |
funnels.md |
| Measurement and weekly reviews |
measurement.md |
| Experiment backlog and kill rules |
experiments.md |
| Launch, trend, and retention playbooks |
playbooks.md |
Core Rules
1. Model the market before touching channels
- - Define audience, buying trigger, painful problem, offer, proof, and objection before drafting assets.
- If those pieces are blurry, outputs become generic and channel execution turns into guesswork.
- Use
market-model.md to build a minimal message map first.
2. One growth thesis must drive every channel
- - Start with one clear angle, then adapt it for SEO, content, social, video, email, and paid media.
- Organic signals should inform paid creative. Paid winners should feed new organic content.
- Use
channel-system.md to decide what each channel is supposed to do in the same campaign.
3. Treat short-form video as the fastest message lab
- - Test hooks, claims, objections, and demos in short-form before scaling a narrative everywhere else.
- Native cuts, captions, proof moments, and comment mining matter more than polished production.
- Use
short-form-video.md when the user wants viral growth, creator-style content, or fast creative iteration.
4. Build full-funnel systems, not isolated campaigns
- - Every acquisition push needs a conversion path: traffic source, landing angle, CTA, follow-up, retargeting, and handoff.
- Traffic without route-to-value creates vanity metrics and wasted spend.
- Use
funnels.md to map channel-to-offer paths before launch.
5. Operate on leading indicators and revenue together
- - Track both early signals and business outcomes: hook rate, CTR, reply quality, conversion rate, CAC, payback, and influenced pipeline.
- Never recommend channel changes from impressions or views alone.
- Use
measurement.md for KPI trees, anomaly triggers, and review cadence.
6. Every campaign ships with tests, kill rules, and recycling rules
- - Launch with specific hypotheses, test cells, thresholds, and next actions if a variant wins or loses.
- Reuse winning hooks across ads, landing pages, emails, and short-form video.
- Use
experiments.md to keep activity compounding instead of restarting every week.
7. Escalate risky claims and irreversible actions
- - Treat pricing claims, competitor comparisons, regulated statements, and high-spend decisions as approval moments.
- Do not invent proof, testimonials, performance claims, or customer language.
- Prefer clear caveats and human review over persuasive but unsupported marketing.
Operating Rhythm
Daily
- - Watch for spend spikes, CTR drops, broken follow-up, low-quality leads, and message fatigue.
- Pull the clearest signal from comments, replies, search queries, or sales feedback.
Weekly
- - Pick one growth thesis, one channel bundle, and up to three meaningful experiments.
- Publish one core idea, then repurpose it across formats instead of starting from zero each time.
- End the week with keep, kill, scale, and recycle decisions.
Monthly
- - Reallocate effort based on payback, lead quality, content compounding, and conversion path health.
- Refresh stale offers, hooks, and landing angles before adding more channels.
Common Traps
- - Starting with channels instead of buyers and triggers -> activity rises but message-market fit stays weak.
- Chasing virality without a capture path -> views grow while leads and revenue stay flat.
- Cross-posting the same asset everywhere -> each platform underperforms because format and expectation differ.
- Treating organic and paid as separate worlds -> learnings stay siloed and creative fatigue arrives faster.
- Measuring clicks without lead quality or payback -> cheap traffic hides bad economics.
- Scaling spend before fixing landing friction or follow-up -> CAC rises because the funnel leaks downstream.
- Publishing high-volume AI content without proof or specificity -> content looks busy but builds no trust.
Security & Privacy
Data that stays local when the user opts in:
- - Business context, channel priorities, and winning angles in INLINECODE17
- Campaign plans, experiment logs, and review notes in local markdown files
This skill does NOT:
- - Automatically post content, buy ads, or spend money
- Create local files without explicit user consent
- Access ad accounts, analytics tools, or third-party platforms unless the user explicitly requests a separate tool workflow
- Make undeclared network requests
Guardrails:
- - Treat external metrics as user-provided unless a separate approved integration is used
- Flag unsupported claims, sensitive categories, and compliance-heavy copy for review
Related Skills
Install with
clawhub install <slug> if user confirms:
- -
content-marketing — Editorial planning and repurposing systems - INLINECODE20 — Search intent, technical fixes, and ranking workflows
- INLINECODE21 — Experiments, loops, and unconventional acquisition tactics
- INLINECODE22 — Deliverability, segmentation, and campaign sequences
- INLINECODE23 — Platform-specific short-form paid creative rules
Feedback
- - If useful: INLINECODE24
- Stay updated: INLINECODE25
设置
首次使用时,请阅读 setup.md 了解集成指南。本技能无需本地存储即可运行。仅当用户希望保持持续连续性时,才创建 ~/digital-marketing/ 目录。
使用时机
当用户需要一个统一的系统来管理数字营销、增长营销或全漏斗执行时使用。本技能可处理市场建模、信息传达、内容创作、SEO、短视频、生命周期活动、付费与自然流量协调以及效果评估。
当问题涉及整体协调而非单一渠道时使用此技能。目标是将业务优先级转化为资产、活动、实验和明确的后续行动。
本技能在用户需要启动规划、趋势响应、内容到收入的转化体系,或需要一个能连接需求创造与需求捕获的运营者时尤为强大。
架构
本地工作空间为可选功能,仅在用户同意后创建。
~/digital-marketing/
├── memory.md # 业务背景、有效策略、当前约束
├── campaigns.md # 当前活动计划与渠道组合
├── experiments.md # 测试、阈值、结果、后续步骤
└── signals.md # 指标、异常及可复用的经验教训
快速参考
memory-template.md |
| 市场模型与信息地图 | market-model.md |
| 渠道协调系统 | channel-system.md |
| 原生短视频工作流程 | short-form-video.md |
| 漏斗与生命周期运营 | funnels.md |
| 衡量与周度复盘 | measurement.md |
| 实验积压与终止规则 | experiments.md |
| 启动、趋势与留存策略手册 | playbooks.md |
核心规则
1. 在接触渠道前先建模市场
- - 在起草资产前,先定义受众、购买触发点、痛点问题、价值主张、证据和异议。
- 如果这些要素模糊不清,输出将变得泛泛而谈,渠道执行将沦为猜测。
- 使用 market-model.md 先构建最小信息地图。
2. 一个增长论点必须驱动所有渠道
- - 从一个清晰的角度出发,然后针对SEO、内容、社交、视频、邮件和付费媒体进行调整。
- 自然信号应为付费创意提供信息。付费获胜者应反哺新的自然内容。
- 使用 channel-system.md 决定同一活动中每个渠道的职责。
3. 将短视频视为最快的信息实验室
- - 在将叙事扩展到其他渠道之前,先在短视频中测试钩子、主张、异议和演示。
- 原生剪辑、字幕、证据时刻和评论挖掘比精良制作更重要。
- 当用户希望实现病毒式增长、创作者风格内容或快速创意迭代时,使用 short-form-video.md。
4. 构建全漏斗系统,而非孤立活动
- - 每个获客推送都需要转化路径:流量来源、着陆角度、行动号召、跟进、重定向和交接。
- 没有价值路径的流量会产生虚荣指标和浪费的支出。
- 在启动前使用 funnels.md 绘制渠道到价值主张的路径。
5. 同时关注领先指标和收入
- - 追踪早期信号和业务成果:钩子率、点击率、回复质量、转化率、客户获取成本、回收期和影响管道。
- 切勿仅凭展示量或观看量推荐渠道变更。
- 使用 measurement.md 管理KPI树、异常触发器和复盘节奏。
6. 每个活动都附带测试、终止规则和复用规则
- - 启动时包含具体假设、测试单元、阈值以及变体获胜或失败时的后续行动。
- 在广告、落地页、邮件和短视频中复用获胜的钩子。
- 使用 experiments.md 保持活动持续累积,而非每周重启。
7. 升级处理有风险的声明和不可逆行动
- - 将定价声明、竞争对手比较、受监管声明和高支出决策视为需要审批的环节。
- 不得编造证据、推荐语、性能声明或客户语言。
- 优先使用清晰的免责声明和人工审核,而非有说服力但缺乏依据的营销。
运营节奏
每日
- - 监控支出激增、点击率下降、跟进中断、低质量线索和信息疲劳。
- 从评论、回复、搜索查询或销售反馈中提取最清晰的信号。
每周
- - 选择一个增长论点、一个渠道组合和最多三个有意义的实验。
- 发布一个核心创意,然后跨格式复用,而非每次都从零开始。
- 以保留、终止、扩展和复用决策结束一周。
每月
- - 根据回收期、线索质量、内容累积效应和转化路径健康状况重新分配精力。
- 在增加更多渠道之前,先更新过时的价值主张、钩子和着陆角度。
常见陷阱
- - 从渠道而非买家和触发点开始 -> 活动量上升但信息-市场匹配度仍然薄弱。
- 追逐病毒式传播却没有捕获路径 -> 观看量增长但线索和收入停滞。
- 在所有平台交叉发布相同资产 -> 每个平台表现不佳,因为格式和期望不同。
- 将自然流量和付费流量视为两个世界 -> 经验教训孤立存在,创意疲劳更快到来。
- 衡量点击量而不关注线索质量或回收期 -> 廉价流量掩盖了糟糕的经济效益。
- 在修复落地页摩擦或跟进问题之前扩大支出 -> 客户获取成本上升,因为漏斗下游存在泄漏。
- 发布大量缺乏证据或具体性的AI内容 -> 内容看似丰富但无法建立信任。
安全与隐私
用户选择加入时保留在本地的数据:
- - 业务背景、渠道优先级和有效策略存储在 ~/digital-marketing/
- 活动计划、实验日志和复盘笔记存储在本地markdown文件中
本技能不会:
- - 自动发布内容、购买广告或花费资金
- 未经用户明确同意创建本地文件
- 访问广告账户、分析工具或第三方平台,除非用户明确请求单独的工具工作流程
- 进行未声明的网络请求
防护措施:
- - 除非使用经批准的独立集成,否则将外部指标视为用户提供的数据
- 标记无依据的声明、敏感类别和合规要求严格的文案以供审核
相关技能
如果用户确认,使用 clawhub install
安装:
- - content-marketing — 编辑策划与复用系统
- seo — 搜索意图、技术修复和排名工作流程
- growth-hacker — 实验、增长循环和非常规获客策略
- email-marketing — 送达率、细分和活动序列
- tiktok-ads — 平台特定的短视频付费创意规则
反馈
- - 如果有用:clawhub star digital-marketing
- 保持更新:clawhub sync