Back to Details

marketing-email-automation

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

EMAIL AUTOMATION — WORKFLOW OS (OPERATIONAL)

邮件营销自动化 — 工作流操作系统（运营级）

Built as a no-fluff execution skill for designing and operating revenue-safe, deliverability-safe email automation across B2B and B2C.
Structure: Core workflows and segmentation in SKILL.md. Platform setup in 
references/
. Revenue economics in 
references/email-economics.md
. Templates in 
assets/
.
Scope note: For outbound/cold email, see marketing-leads-generation. For email HTML/CSS implementation, use a coding-focused workflow (this skill is strategy + operations).
这是一项无冗余的落地技能，专为B2B和B2C场景设计，可搭建兼顾营收安全与交付性安全的邮件自动化体系。
结构说明：核心工作流与用户细分规则见SKILL.md。平台配置指引在
references/
目录。营收经济效益相关内容在
references/email-economics.md
。模板文件存于
assets/
目录。
范围说明：关于 outbound/冷邮件营销，请查看marketing-leads-generation。若需邮件HTML/CSS开发，请使用代码聚焦型工作流（本技能侧重策略与运营）。

Modern Best Practices (January 2026)

2026年最新最佳实践

Operational reality in 2026:
  • Authentication and alignment are enforced (SPF/DKIM/DMARC).
  • One-click unsubscribe is required for bulk sending.
  • Inbox placement is driven by multi-signal engagement and complaint rates.
  • Inactive recipients degrade reputation; enforce suppression/sunset policies.
  • Open rates are unreliable due to privacy features (e.g., Apple MPP); measure downstream actions.
  • AI-powered personalization is common; ship with guardrails and measure incrementality.
  • Interactive email is possible (e.g., AMP for Email in Gmail), but client support is limited; treat it as progressive enhancement.
  • EU Accessibility Act (EAA) effective June 2025; WCAG 2.2 AA is the standard.
References:
  • Sources: 
    data/sources.json
  • Deliverability checklist: 
    references/deliverability.md
  • Accessibility requirements: 
    references/accessibility.md
2026年的运营现状：
  • 邮件认证与对齐机制已强制执行（SPF/DKIM/DMARC）。
  • 批量发送邮件必须支持一键退订。
  • 收件箱投递率由多维度互动信号与投诉率共同决定。
  • 不活跃收件人会损害发件人声誉；需严格执行抑制/淘汰策略。
  • 由于隐私功能（如Apple MPP），打开率数据已不可靠；需衡量下游行为指标。
  • AI驱动的个性化已普及；需设置防护规则并衡量增量效果。
  • 交互式邮件已实现（如Gmail支持的AMP for Email），但客户端支持有限；应将其作为渐进式增强功能。
  • EU无障碍法案（EAA）于2025年6月生效；WCAG 2.2 AA为合规标准。
参考资料：
  • 数据源：
    data/sources.json
  • 交付性检查清单：
    references/deliverability.md
  • 无障碍要求：
    references/accessibility.md

Expert: Automation vs Campaigns (Quick Calibration)

专家校准：自动化 vs 营销活动（快速区分）

Use this to avoid "scheduled campaigns in disguise."
  • Email campaigns: One-to-many broadcasts optimized for short-lived goals (announcement, promotion).
  • Email automation: Trigger-based workflows optimized for reliability, restraint, and state transitions.
  • Lifecycle messaging: The set of automations that move users through product and revenue states (activation → value → expansion → renewal) with consistent suppression and stop rules.
  • Revenue-driven communication system: Cross-functional control system where email is one actuator among others (product UX, sales, support, paid) measured for incrementality, not attribution comfort.
Organizational failure mode when automation is treated as scheduled campaigns:
  • Teams optimize local email metrics while silently degrading global outcomes (deliverability, qualified pipeline, retention, margin) because messaging is decoupled from real user state.
用此标准避免将自动化变成"伪装的定时营销活动"。
  • 邮件营销活动：一对多的广播式发送，针对短期目标（如公告、促销）优化。
  • 邮件自动化：基于触发机制的工作流，针对可靠性、克制性与用户状态转换优化。
  • 生命周期触达：一套自动化体系，通过统一的抑制与停止规则，引导用户完成产品使用与营收状态的转换（激活→价值实现→拓展→续费）。
  • 营收驱动的沟通系统：跨职能控制系统，邮件是其中一个执行环节（其他环节包括产品UX、销售、客服、付费推广），需以增量效果而非归因便利性为衡量标准。
将自动化视为定时营销活动的组织失败模式：
  • 团队仅优化邮件局部指标，却悄然损害全局结果（交付性、合格线索池、留存率、利润率），因为触达信息与真实用户状态脱节。

Core: Lifecycle Architecture (Behavior, Not CRM Defaults)

核心：生命周期架构（基于行为，而非CRM默认标签）

Define stages by user-state constraints

以用户状态约束定义阶段

Lifecycle stages are not CRM labels. A lifecycle stage is a set of constraints that implies:
  • What message can help right now
  • What message will be noise or cause harm
Define stages by:
  • Intent (is the user trying to solve the problem now?)
  • Ability (do they have access, setup, permissions, budget?)
  • Value realization (have they experienced the core value loop?)
  • Risk (churn risk, buyer risk, compliance risk)
  • Route (self-serve vs sales-assisted vs partner)
生命周期阶段并非CRM标签。一个生命周期阶段是一组约束条件，意味着：
  • 当前哪些信息对用户有帮助
  • 哪些信息是噪音或会造成干扰
通过以下维度定义阶段：
  • 意图（用户是否正试图解决当前问题？）
  • 能力（用户是否具备访问权限、已完成设置、拥有预算？）
  • 价值实现（用户是否体验过核心价值循环？）
  • 风险（流失风险、购买风险、合规风险）
  • 路径（自助服务 vs 销售辅助 vs 合作伙伴渠道）

What moves users between stages

驱动用户阶段转换的因素

Stage transitions should be driven by observable behavior and system events (product, billing, support, sales), not time.
Examples (event-level, platform-agnostic):
  • Activation: first meaningful action completed (not "account created")
  • Value: repeated value loop within a window (not "opened email 3 times")
  • Expansion readiness: feature ceiling hit, invites attempted, usage saturation, admin intent signals
  • Churn risk: value loop stops, negative support signals, failed payments, downgrade intent
Strategically wrong signal that looks correct:
  • Using email engagement (opens/clicks) as a proxy for lifecycle movement; it confuses attention with progress.
阶段转换应由可观测的行为与系统事件（产品、计费、客服、销售）驱动，而非时间。
示例（事件级，平台无关）：
  • 激活：完成首次有意义的操作（而非"账户创建"）
  • 价值实现：在周期内重复完成核心价值循环（而非"打开邮件3次"）
  • 拓展就绪：触及功能上限、尝试邀请他人、使用饱和、管理员意图信号
  • 流失风险：核心价值循环中断、负面客服信号、支付失败、降级意图
看似正确但策略错误的信号：
  • 将邮件互动（打开/点击）作为生命周期转换的代理指标；这混淆了用户注意力与实际进度。

Core: Workflow Design (Expert Boundaries)

核心：工作流设计（专家边界）

Branch only when the best next action materially changes due to:
  • Different constraints (setup/permissions/budget)
  • Different route (sales-assisted vs self-serve)
  • Different risk profile (compliance, churn, payment failure)
  • Different intent (decision support vs execution support)
Collapse logic when differences are cosmetic and can be handled by:
  • Segmented entry rules (separate enrollment) instead of deep branching
  • Shared core path + modular blocks (same workflow, different payload)
  • Stop rules (exit on success) rather than "if/else forever"
Automation complexity is harming performance when:
  • You cannot explain "why did this user get this message" without manual platform tracing
  • Users qualify for multiple flows and experience frequency spikes
  • Branches rely on stale or inconsistently populated fields
  • Too many simultaneous changes prevent causal learning
Early warning sign of unmaintainable automation:
  • The team cannot answer "who is in this flow and why" without opening the ESP and manually stepping through conditions.
仅在以下情况分支逻辑：不同约束（设置/权限/预算）、不同路径（销售辅助 vs 自助）、不同风险等级（合规、流失、支付失败）、不同意图（决策支持 vs 执行支持）导致最佳后续操作发生实质性变化。
当差异仅为表面性时，应简化逻辑：
  • 用细分准入规则（单独入流）替代深层分支
  • 共享核心路径+模块化内容块（同一工作流，不同内容 payload）
  • 用停止规则（成功即退出）替代"无限if/else"
自动化复杂度损害性能的信号：
  • 若不手动追踪平台日志，无法解释"该用户为何收到此邮件"
  • 用户符合多个工作流条件，导致发送频率激增
  • 分支依赖过时或填充不一致的字段
  • 过多同步变更导致无法进行因果分析
自动化难以维护的早期预警信号：
  • 团队无需打开ESP平台并手动遍历条件，就无法回答"哪些用户在该工作流中，原因是什么"

Core: Platform Literacy (Non-Tool-Centric)

核心：平台认知（非工具导向）

Platform-agnostic decisions (must be stable even if you switch tools):
  • Lifecycle stage model and transitions
  • Event taxonomy and identity rules (user vs account vs contact; merge/split)
  • Consent, suppression, and frequency policies (including global caps)
  • Measurement strategy (incrementality, holdouts, decision criteria)
  • Routing rules between product, sales, support, and email
Legitimately platform-dependent decisions:
  • Data model constraints (profile vs event tables, custom objects)
  • Segmentation expressiveness (real-time vs batch; lookback windows)
  • Identity resolution behavior (dedupe rules, stitching)
  • Deliverability tooling availability (sending pools, domain controls)
  • Workflow evaluation semantics (how often conditions re-evaluate)
Common mistake from over-trusting platform "best practices":
  • Treating default attribution windows and engagement filters as if they represent causality.
平台无关的决策（切换工具后仍需保持稳定）：
  • 生命周期阶段模型与转换规则
  • 事件分类与身份规则（用户 vs 账户 vs 联系人；合并/拆分规则）
  • 同意、抑制与频率策略（包括全局上限）
  • 衡量策略（增量效果、对照组测试、决策标准）
  • 产品、销售、客服与邮件之间的路由规则
确实依赖平台的决策：
  • 数据模型约束（用户档案 vs 事件表、自定义对象）
  • 细分功能的灵活性（实时 vs 批量；回溯窗口）
  • 身份解析行为（去重规则、身份拼接）
  • 交付性工具可用性（发送池、域名控制）
  • 工作流评估语义（条件重新评估的频率）
过度信任平台"最佳实践"的常见错误：
  • 将默认归因窗口与互动过滤器视为因果关系的代表。

Core: Automation Archetypes (What They Solve)

核心：自动化原型（解决的问题）

Workflow TypeTriggerPurposeTimeline
New subscriber / opt-inConsent capturedSet expectations, route to intent pathDay 0-7
New customer / startFirst purchase / startDrive first value loopDay 0-30
Decision supportIntent signalReduce decision risk, progress to commitmentWeeks 1-8
Abandoned CartCart created, no purchaseRecover revenueHours 1-72
Re-engagementInactive subscriberWin back or clean listDay 30-90
Upsell/Cross-sellPurchase completeIncrease LTVDay 7-30 post-purchase
Renewal/RetentionSubscription nearing endPrevent churnDay -30 to -1
工作流类型触发条件目的时间范围
新订阅用户/选择加入捕获用户同意建立预期，引导至对应意图路径第0-7天
新客户/服务启动首次购买/服务启动推动首次核心价值循环第0-30天
决策支持意图信号触发降低决策风险，推进承诺达成第1-8周
购物车遗弃召回创建购物车但未完成购买挽回营收1-72小时
重新激活订阅用户不活跃赢回用户或清理列表第30-90天
追加销售/交叉销售购买完成提升客户终身价值（LTV）购买后第7-30天
续费/留存订阅即将到期防止流失到期前30至1天

Core: Workflow Design Framework

核心：工作流设计框架

text
WORKFLOW: [Name]
TRIGGER: [Event that starts flow]
GOAL: [Primary conversion action]
DURATION: [Total time span]

ENTRY → STEP 1 → WAIT → BRANCH (state A/state B) → STEP 2 → ... → EXIT

Exit when: Goal met | Unsubscribe | Time limit reached
See assets/workflow-blueprint.md for full template.
text
工作流：[名称]
触发条件：[启动工作流的事件]
目标：[主要转化动作]
时长：[总时间跨度]

准入 → 步骤1 → 等待 → 分支（状态A/状态B） → 步骤2 → ... → 退出

退出条件：达成目标 | 退订 | 达到时间限制
完整模板请查看assets/workflow-blueprint.md。

Core: Nurture Sequences (Reality Check)

核心：培育序列（现状审视）

What nurture is actually optimizing for:
  • Reduce time-to-value and time-to-commitment by removing uncertainty
  • Increase decision confidence (not "engagement") through clarity and risk reduction
  • Detect route: self-serve users progress; sales-assisted users surface intent and constraints
Why most nurture fails even with good content:
  • Wrong stage model: messages do not match the user's constraint (they cannot act yet, or already acted)
  • No stop rules: users keep receiving messages after conversion, churn, or route change
  • No cross-channel routing: email tries to replace product UX, sales, or support, so it becomes noise
  • Optimization for clicks: high click-through can be "confusion," not progress
How nurture changes once product or sales signals exist:
  • Shift from generic persuasion to contextual assistance (unblock setup, remove friction, coordinate handoffs)
  • Use product and sales events as primary triggers; email becomes a reinforcement and routing layer
  • Segment by next constraint (permissions, onboarding step, stakeholder, pricing fit), not by persona labels
Optional scaffolds (if you need examples, not copy guidance):
  • assets/welcome-sequence.md
  • assets/nurture-sequence.md
  • assets/cart-abandonment.md
培育序列的实际优化目标：
  • 通过消除不确定性，缩短价值实现时间与承诺达成时间
  • 通过清晰沟通与风险降低，提升决策信心（而非"互动率"）
  • 识别用户路径：自助用户推进转化；销售辅助用户暴露意图与约束
多数培育序列内容优质但效果不佳的原因：
  • 阶段模型错误：信息与用户当前约束不匹配（用户无法行动，或已完成行动）
  • 无停止规则：用户完成转化、流失或路径变更后仍持续收到信息
  • 无跨渠道路由：邮件试图替代产品UX、销售或客服，导致信息沦为噪音
  • 以点击为优化目标：高点击率可能源于"困惑"而非"进度"
当产品或销售信号出现后，培育序列的调整方向：
  • 从通用说服转向情境化协助（解决设置障碍、减少摩擦、协调交接）
  • 以产品与销售事件为主要触发条件；邮件成为强化与路由层
  • 按下一约束条件（权限、入门步骤、利益相关者、定价适配）细分，而非按用户角色标签
可选参考框架（仅作示例，非抄袭指南）：
  • assets/welcome-sequence.md
  • assets/nurture-sequence.md
  • assets/cart-abandonment.md

Core: Segmentation Framework

核心：细分框架

Behavioral Segments

行为细分

SegmentDefinitionUse For
EngagedClicked (or took downstream action) in last 30 daysPrimary campaigns
ActiveLogged in / used product recentlyOnboarding, upsell
At-riskNo engagement 31-60 daysRe-engagement
DormantNo engagement 61-90 daysWin-back
ChurnedNo engagement 90+ daysSunset or remove
细分群体定义用途
活跃互动过去30天内有点击（或下游行为）核心营销活动
产品活跃近期登录/使用产品入门引导、追加销售
风险预警31-60天无互动重新激活
休眠61-90天无互动用户赢回
流失90天以上无互动淘汰或清理列表

RFM Segmentation (E-commerce)

RFM细分（电商场景）

SegmentRecencyFrequencyMonetaryStrategy
ChampionsRecentOftenHighReward, refer
LoyalRecentOftenMediumUpsell
PotentialRecentRareLowNurture
At RiskOldOftenHighWin back
HibernatingOldRareLowRe-engage or remove
细分群体最近一次互动互动频率消费金额策略
核心客户近期频繁高奖励、推荐计划
忠诚客户近期频繁中追加销售
潜力客户近期偶尔低培育转化
风险客户远期频繁高赢回
沉睡客户远期偶尔低重新激活或清理

Zero-Party Data Collection

零方数据收集

Zero-party data = information users explicitly and voluntarily provide (preferences, intentions, feedback). Unlike behavioral data, this comes directly from the user's stated intent.
Why it matters (2026):
  • Third-party cookies deprecated; first-party and zero-party data are now primary
  • 48% of consumers prefer brands that collect data transparently
  • 93% of marketers view first-party data as critical for future-proofing
Collection methods:
MethodExampleBest For
Preference centersContent frequency, topic interests, channel preferencesList health, relevance
Progressive profilingAsk 1-2 questions per touchpoint over timeLow friction, high completion
Surveys/quizzesGamified "find your style" or product matchEngagement + data capture
Onboarding questionsRole, company size, goals during signupB2B segmentation
Post-purchase feedback"How did you hear about us?"Attribution clarity
Implementation principles:
  • Clear value exchange: explain why you're asking and what they get
  • Progressive, not invasive: 1-2 questions per interaction, not 10-field forms
  • Respect stated preferences: if they say "weekly," don't send daily
  • Store with consent timestamp: GDPR/CCPA requires audit trail
Dynamic content from zero-party data:
  • Product recommendations based on stated preferences (not just browsing)
  • Content blocks that change based on declared interests
  • Send frequency that respects explicit preferences
  • Lifecycle stage based on stated goals, not inferred behavior
零方数据 = 用户主动且自愿提供的信息（偏好、意图、反馈）。与行为数据不同，它直接来自用户声明的意图。
2026年的重要性：
  • 第三方Cookie已废弃；第一方与零方数据成为核心数据源
  • 48%的消费者更青睐透明收集数据的品牌
  • 93%的营销人员认为第一方数据对未来发展至关重要
收集方法：
方法示例最佳场景
偏好中心内容频率、主题兴趣、渠道偏好列表健康度、相关性优化
渐进式画像每次互动仅询问1-2个问题低摩擦、高完成率
调查/测验游戏化的"找到你的风格"或产品匹配测验互动+数据收集
入门问题注册时询问角色、公司规模、目标B2B用户细分
购后反馈"你是如何了解到我们的？"归因清晰度
实施原则：
  • 清晰的价值交换：解释询问原因与用户收益
  • 渐进式而非侵入式：每次互动仅问1-2个问题，而非10字段表单
  • 尊重声明偏好：若用户选择"每周"，则不发送每日邮件
  • 存储同意时间戳：GDPR/CCPA要求审计追踪
基于零方数据的动态内容：
  • 基于声明偏好的产品推荐（而非仅浏览记录）
  • 根据声明兴趣调整内容块
  • 遵循明确偏好的发送频率
  • 基于声明目标而非推断行为的生命周期阶段

Core: Message Design Principles (Non-Copy)

核心：消息设计原则（非文案层面）

  • Minimize mismatch: map each message to one user constraint and one next action
  • Prefer stop rules over reminders: exit on success, suppress on route change
  • Use frequency caps and suppression as first-class design surfaces, not afterthoughts
  • Treat incentives as a last resort (they can train discounting and attract low-LTV behavior)
  • Optimize for downstream outcomes (activation, retention, qualified pipeline), not surface engagement
  • 最小化不匹配：每条消息对应一个用户约束与一个后续动作
  • 优先使用停止规则而非提醒：成功即退出，路径变更即抑制
  • 将频率上限与抑制策略作为核心设计环节，而非事后补救
  • 将激励措施作为最后手段（会培养用户的折扣预期，吸引低LTV行为）
  • 优化下游结果（激活、留存、合格线索池），而非表面互动率

Deliverability as a System (Not a Checklist)

交付性作为系统（而非检查清单）

Deliverability is three coupled systems:
  • Technical system: authentication (SPF/DKIM/DMARC), alignment, headers, infrastructure, bounces
  • Behavioral system: whether recipients consistently signal "this is wanted" (read, reply, save, move) and avoid "this is spam"
  • Reputation system: ISPs infer your sender quality from aggregate behavior, consistency, and list hygiene over time
One action that improves short-term performance but harms long-term deliverability:
  • Increasing send volume by re-mailing unengaged users to "recover revenue" often creates a delayed reputation crash (complaints up, inbox placement down, future revenue down).
交付性由三个相互关联 的系统组成：
  • 技术系统：认证（SPF/DKIM/DMARC）、对齐机制、邮件头、基础设施、退信处理
  • 行为系统：收件人是否持续发出"此邮件为所需"的信号（阅读、回复、保存、移动），并避免标记为垃圾邮件
  • 声誉系统：ISP通过长期的聚合行为、一致性与列表卫生，推断发件人质量
短期提升性能但损害长期交付性的行为：
  • 通过重新发送给不活跃用户"挽回营收"来增加发送量，通常会导致延迟的声誉崩溃（投诉率上升、收件箱投递率下降、未来营收减少）。

AI-Powered Email Automation (2026)

AI驱动的邮件自动化（2026年）

AI capabilities are mainstream. Treat them as accelerators, not substitutes for measurement, consent, and deliverability hygiene.
Production-ready AI capabilities:
CapabilityWhat It DoesWhen to Use
Send time optimizationPredicts best send time per recipientHigh-volume campaigns, engagement optimization
Subject line generationAI-generated variants for testingA/B testing, creative scaling
Content personalizationDynamic blocks based on behavior + preferencesNurture sequences, product recommendations
Predictive segmentationIdentifies churn risk, purchase likelihoodRetention, upsell targeting
Copy generationFull email drafts from promptsCampaign scaling, first drafts
Guardrails (non-negotiable):
  • AI suggestions must pass deliverability review (no spam triggers, no misleading claims)
  • Human approval for anything customer-facing until confidence is established
  • A/B test AI vs human copy; measure downstream outcomes, not just opens
  • Stop rules: if AI-generated content underperforms, revert to human baseline
Autonomous campaigns (emerging):
  • Real-time systems can adapt timing, offer, and content based on behavior and context.
  • Use only when measurement infrastructure is mature (incrementality, holdouts) and governance is explicit.
Warning: AI can optimize for surface metrics (opens, clicks) while degrading system outcomes (deliverability, retention, margin). Always validate with downstream outcomes.
AVOID Over-automation without controls: Complex flows without monitoring and suppression eventually damage reputation. Review automation performance monthly and retire stale logic.
AI能力已成为主流。将其视为加速器，而非衡量、同意与交付性卫生的替代品。
可落地的AI能力：
能力功能适用场景
发送时间优化预测每个收件人的最佳发送时间高容量营销活动、互动率优化
主题行生成AI生成多个变体用于测试A/B测试、创意规模化
内容个性化基于行为+偏好的动态内容块培育序列、产品推荐
预测性细分识别流失风险、购买可能性用户留存、追加销售定向
文案生成根据提示生成完整邮件草稿营销活动规模化、初稿撰写
必须遵循的防护规则：
  • AI建议必须通过交付性审核（无垃圾邮件触发词、无误导性声明）
  • 面向客户的内容需经人工批准，直至建立足够信心
  • A/B测试AI与人工文案；衡量下游结果，而非仅打开率
  • 停止规则：若AI生成内容表现不佳，恢复为人工基线
自主营销活动（新兴）：
  • 实时系统可根据行为与上下文调整时间、优惠与内容。
  • 仅在衡量基础设施成熟（增量效果、对照组测试）且治理明确时使用。
警告：AI可能优化表面指标（打开率、点击率）但损害系统结果（交付性、留存率、利润率）。始终用下游结果验证。
避免：无控制的过度自动化：无监控与抑制的复杂工作流最终会损害声誉。每月审核自动化性能，淘汰过时逻辑。

Core: Deliverability Controls (Operational)

核心：交付性控制（运营层面）

Use 
references/deliverability.md
for the full compliance checklist (including BIMI and enforcement status). Use 
assets/email-audit.md
for audits.
完整合规检查清单（包括BIMI与执行状态）请查看
references/deliverability.md
。审计模板请使用
assets/email-audit.md
。

Revenue Attribution (Hard Part)

营收归因（难点）

Why last-touch attribution fails for automation:
  • Automation participates across a long decision journey; last-touch over-credits the final nudge and under-credits earlier state changes.
  • It rewards late-stage "harvesting intent" patterns that may not be incremental.
  • It breaks when routes mix (product-led + sales-led) and when offline touches exist.
How experts estimate email contribution without overclaiming:
  • Treat attribution dashboards as accounting, not truth.
  • Prefer incrementality: holdouts, randomized suppression, geo splits, or step-wedge rollouts.
  • Evaluate system outcomes: activation rate, time-to-value, retention cohorts, pipeline velocity, margin protection.
Attribution signal to trust more than dashboards:
  • Lift measured via a properly designed holdout/suppression test on the automation (conversion/retention difference between eligible users who were and were not emailed).
最后点击归因在自动化场景失效的原因：
  • 自动化参与整个漫长的决策旅程；最后点击过度归功于最终推动，而低估了早期状态变化的贡献。
  • 它奖励后期"收割意图"模式，但这些模式可能不具备增量效果。
  • 当路径混合（产品驱动+销售驱动）或存在线下触点时，归因会失效。
专家如何在不过度主张的情况下估算邮件贡献：
  • 将归因仪表盘视为记账工具，而非事实。
  • 优先使用增量衡量方法：对照组测试、随机抑制、地域拆分或逐步推广测试。
  • 评估系统结果：激活率、价值实现时间、留存 cohort、线索流转速度、利润率保护。
比仪表盘更可信的归因信号：
  • 通过设计合理的自动化对照组/抑制测试衡量的提升效果（符合条件的用户中，收到邮件与未收到邮件的用户在转化/留存上的差异）。

Economics & Retention (Use Subfiles)

经济效益与留存（使用子文件）

Use 
references/email-economics.md
for attribution models, RPE framework, segment economics, churn reduction ROI, and channel mix decisions.
How automation contributes:
  • Retention: reinforces the value loop, removes friction, and routes users to help before they stall
  • Expansion: drives feature adoption and account-level enablement when usage ceilings or admin intent appear
  • Churn reduction: detects risk states (usage drop, failed payments, negative support signals) and triggers corrective paths with tight suppression
归因模型、RPE框架、细分群体经济效益、流失降低ROI与渠道组合决策请查看
references/email-economics.md
。
自动化的贡献方式：
  • 留存：强化核心价值循环、减少摩擦、在用户停滞前引导至帮助渠道
  • 拓展：当触及使用上限或出现管理员意图信号时，推动功能采用与账户级赋能
  • 流失降低：检测风险状态（使用量下降、支付失败、负面客服信号），并触发带有严格抑制规则的纠正路径

Metrics & KPIs

指标与KPI

Use metrics to detect system health, not to "win the dashboard":
  • Open rate: trend only (MPP).
  • Click/action rate: engagement signal; prefer downstream action.
  • Revenue per email: accounting signal; validate with incrementality.
  • Conversion and retention cohorts: primary outcomes.
Retention-related email metric that can look positive while the business degrades:
  • Rising "re-engagement" click rate driven by incentives while retention cohorts and gross margin deteriorate (you are buying activity, not improving value).
See 
references/email-economics.md
for definitions and calculations.
用指标检测系统健康度，而非"美化仪表盘"：
  • 打开率：仅看趋势（受MPP影响）。
  • 点击/行动率：互动信号；优先看下游行为。
  • 每邮件营收：记账信号；需用增量效果验证。
  • 转化与留存 cohort：核心结果。
看似正面但实则损害业务的留存相关邮件指标：
  • 由激励措施驱动的"重新激活"点击率上升，但留存 cohort与毛利率恶化（你在购买用户活动，而非提升价值）。
定义与计算方法请查看
references/email-economics.md
。

Platform Setup Guides

平台配置指南

  • references/hubspot-setup.md
  • references/klaviyo-setup.md
  • references/mailchimp-setup.md
  • references/accessibility.md — WCAG 2.2, EAA compliance, dark mode
  • references/segmentation-strategies.md — Behavioral, RFM, lifecycle, zero-party data segmentation
  • references/automation-workflows.md — Lifecycle flows, trigger design, branching, timing, benchmarks
  • references/email-copywriting.md — Subject lines, preheaders, CTAs, mobile-first copy
  • references/ab-testing-email.md — Test design, sample sizing, winner selection, platform features
  • references/re-engagement-winback.md — Sunset policies, win-back sequences, list hygiene
  • references/transactional-email.md — Transactional vs marketing email, delivery infra, compliance boundaries
  • references/email-design-systems.md — Modular templates, responsive patterns, dark mode, client rendering
  • references/hubspot-setup.md
  • references/klaviyo-setup.md
  • references/mailchimp-setup.md
  • references/accessibility.md — WCAG 2.2、EAA合规、深色模式
  • references/segmentation-strategies.md — 行为、RFM、生命周期、零方数据细分
  • references/automation-workflows.md — 生命周期工作流、触发设计、分支、时间安排、基准数据
  • references/email-copywriting.md — 主题行、预标题、CTA、移动端优先文案
  • references/ab-testing-email.md — 测试设计、样本量、胜者选择、平台功能
  • references/re-engagement-winback.md — 淘汰策略、赢回序列、列表卫生
  • references/transactional-email.md — 事务性 vs 营销邮件、交付基础设施、合规边界
  • references/email-design-systems.md — 模块化模板、响应式模式、深色模式、客户端渲染

Templates

模板

TemplatePurpose
workflow-blueprint.mdDesign any workflow
welcome-sequence.mdNew subscriber welcome
nurture-sequence.mdLead nurturing
cart-abandonment.mdE-commerce recovery
email-audit.mdHealth check template
email-roi-calculator.mdROI calculation
模板用途
workflow-blueprint.md设计任意工作流
welcome-sequence.md新订阅用户欢迎序列
nurture-sequence.md线索培育序列
cart-abandonment.md电商召回序列
email-audit.md邮件健康检查模板
email-roi-calculator.mdROI计算模板

Interactive Email (AMP for Email)

交互式邮件（AMP for Email）

AMP for Email enables in-email actions without leaving the inbox. Users can complete forms, browse carousels, submit reviews, and update preferences directly in the message.
Adoption status (2026):
  • Support is limited and uneven across clients; treat AMP as progressive enhancement, not a default UX.
  • You must ship a full HTML fallback that preserves intent (not just a broken placeholder).
Interactive elements and use cases:
ElementUse Case
FormsSurveys, reviews, preference updates
CarouselsProduct browsing, feature highlights
AccordionsFAQ, detailed specs, terms
Live dataInventory counts, pricing, appointment slots
CheckoutOne-click purchase, cart updates
Implementation requirements:
  • Sender must be registered with Google (AMP for Email sender registration)
  • DKIM/SPF/DMARC must pass
  • AMP MIME part required alongside HTML fallback
  • Content must be dynamic (server-rendered, not static)
When to use AMP:
  • High-intent moments: abandoned cart, booking confirmation, review request
  • Data collection: NPS surveys, preference updates, feedback forms
  • Real-time content: inventory, pricing, appointment availability
When NOT to use AMP:
  • Simple announcements (overhead not justified)
  • Audiences primarily on Apple Mail (no AMP support)
  • When HTML fallback would be significantly different experience
Fallback requirement: Always provide HTML fallback. AMP is progressive enhancement, not replacement.
Testing: Use Litmus, Email on Acid, or Gmail's AMP Playground to validate rendering across clients.
AMP for Email支持在邮件内直接完成操作，无需离开收件箱。用户可直接在邮件中填写表单、浏览轮播、提交评价、更新偏好。
2026年采用现状：
  • 客户端支持有限且不均衡；将AMP视为渐进式增强功能，而非默认UX。
  • 必须提供完整的HTML降级方案，确保核心意图不受影响（而非仅显示失效占位符）。
交互式元素与适用场景：
元素适用场景
表单调查、评价、偏好更新
轮播产品浏览、功能亮点展示
折叠面板FAQ、详细规格、条款
实时数据库存数量、定价、预约时段
结账一键购买、购物车更新
实施要求：
  • 发件人必须在Google完成AMP for Email发件人注册
  • 必须通过SPF/DKIM/DMARC认证
  • 需同时提供AMP MIME部分与HTML降级方案
  • 内容必须为动态（服务器渲染，而非静态）
何时使用AMP：
  • 高意图场景：购物车遗弃、预约确认、评价请求
  • 数据收集：NPS调查、偏好更新、反馈表单
  • 实时内容：库存、定价、预约可用性
何时不使用AMP：
  • 简单公告（投入产出比不足）
  • 受众主要使用Apple Mail（无AMP支持）
  • HTML降级方案与AMP体验差异过大
降级要求：始终提供HTML降级方案。AMP是渐进式增强，而非替代方案。
测试：使用Litmus、Email on Acid或Gmail的AMP Playground验证跨客户端渲染效果。

CRO vs Email Automation Boundary

CRO与邮件自动化的边界

Email automation supports CRO when it:
  • Routes the right users back into the funnel at the right stage (intent match)
  • Reduces friction and uncertainty around the next commitment (activation, checkout, demo)
Email automation conflicts with CRO when it:
  • Trains the market to wait for discounts or bypasses qualification gates
  • Drives low-intent traffic back to conversion pages (inflates visits, lowers close quality)
Optimization mistake that improves email metrics but reduces funnel efficiency:
  • Maximizing click-through by broadening eligibility and adding incentives, which increases page traffic while lowering qualified conversion rate and harming deliverability.
邮件自动化支持CRO的场景：
  • 在正确的阶段将合适的用户重新引导至漏斗（意图匹配）
  • 减少下一个承诺环节（激活、结账、演示）的摩擦与不确定性
邮件自动化与CRO冲突的场景：
  • 训练市场等待折扣，或绕过资格审核门槛
  • 将低意图流量重新引导至转化页面（增加访问量，但降低合格转化率）
提升邮件指标但降低漏斗效率的优化错误：
  • 通过放宽准入条件与添加激励措施最大化点击率，这会增加页面流量，但降低合格转化率并损害交付性。

Red Flags (Subtle, Practitioner-Level)

危险信号（细微，从业者层面）

Three statements that signal non-expert operation:
  1. "We just need more emails in the flow to increase conversions."
  2. "If the email gets clicks, the automation is working."
  3. "Our platform attribution shows email drove X%, so we should scale volume."
三个表明非专业操作的说法：
  1. "我们只需在工作流中添加更多邮件即可提升转化率。"
  2. "如果邮件有点击，自动化就是有效的。"
  3. "我们的平台归因显示邮件贡献了X%，所以应该扩大发送量。"

Decision Tree (Email Triage)

决策树（邮件问题排查）

text
Complaints rising (or inbox placement dropping)?
├─ Suppress unengaged; tighten eligibility
├─ Reduce frequency; add global caps
└─ Audit list sources + consent + auth alignment

Bounces rising?
├─ Remove hard bounces immediately
├─ Fix list acquisition hygiene (double opt-in where appropriate)
└─ Verify infrastructure and authentication

Clicks/actions down but deliverability stable?
├─ Stage mismatch: rebuild lifecycle entry/exit rules
├─ Route mismatch: coordinate with product/sales/support signals
└─ Validate impact with holdouts; do not "add more emails" first

Revenue up but retention/margin down?
└─ Incentive training or low-LTV acquisition: tighten segmentation and protect pricing integrity
text
投诉率上升（或收件箱投递率下降）？
├─ 抑制不活跃用户；收紧准入条件
├─ 降低发送频率；添加全局上限
└─ 审核列表来源+同意机制+认证对齐

退信率上升？
├─ 立即移除硬退信用户
├─ 优化列表获取卫生（适当场景启用双重确认）
└─ 验证基础设施与认证

点击/行动率下降但交付性稳定？
├─ 阶段不匹配：重建生命周期准入/退出规则
├─ 路径不匹配：协调产品/销售/客服信号
└─ 用对照组测试验证效果；不要先"添加更多邮件"

营收上升但留存/利润率下降？
└─ 激励训练或低LTV获取：收紧细分规则，保护定价完整性

Anti-Patterns

反模式

Anti-PatternInstead
Batch-and-blastSegment and trigger
No segmentationBehavioral segments
Multiple competing goalsOne primary outcome per message
Ignoring mobile constraintsMobile constraints first (tap targets, load, readability)
Buying listsBuild organically
Set-and-forget automationsMonthly performance reviews
Relying on open ratesUse click rate, conversions, revenue
反模式替代方案
批量群发细分+触发
无用户细分行为细分
多个竞争目标单条信息对应一个核心结果
忽略移动端约束优先考虑移动端约束（点击目标、加载速度、可读性）
购买邮件列表有机搭建列表
一劳永逸的自动化每月性能审核
依赖打开率使用点击率、转化率、营收

International Markets

国际市场

This skill uses US/UK market defaults. For international email marketing:
NeedSee Skill
Regional compliance (CASL, LGPD, PIPL)marketing-geo-localization
Regional send time optimizationmarketing-geo-localization
Cultural messaging adaptationmarketing-geo-localization
Alternative channels (WhatsApp, LINE, WeChat)marketing-geo-localization
Auto-triggers: When your query mentions a specific country, region, or compliance framework, both skills load automatically.
本技能默认采用美英市场规则。针对国际邮件营销：
需求对应技能
区域合规（CASL、LGPD、PIPL）marketing-geo-localization
区域发送时间优化marketing-geo-localization
文化适配信息marketing-geo-localization
替代渠道（WhatsApp、LINE、微信）marketing-geo-localization
自动触发：当查询提及特定国家、地区或合规框架时，两个技能将自动加载。

Related Skills

相关技能

  • marketing-geo-localization — International marketing, regional compliance
  • marketing-leads-generation — Lead capture
  • marketing-content-strategy — Content for emails
  • marketing-cro — Landing page optimization
  • startup-go-to-market — Channel strategy
  • marketing-geo-localization — 国际营销、区域合规
  • marketing-leads-generation — 线索捕获
  • marketing-content-strategy — 邮件内容策略
  • marketing-cro — 落地页优化
  • startup-go-to-market — 渠道策略

Usage Notes (Claude)

使用说明（Claude）

  • Stay operational: return lifecycle stage model, trigger rules, suppression/frequency policy, and workflow diagrams
  • For revenue/ROI questions, reference 
    references/email-economics.md
  • Always include deliverability requirements (authentication, hygiene)
  • Do not invent benchmark data; use ranges or state "varies by industry"
  • 聚焦落地：返回生命周期阶段模型、触发规则、抑制/频率策略与工作流示意图
  • 针对营收/ROI问题，参考
    references/email-economics.md
  • 始终包含交付性要求（认证、列表卫生）
  • 不要编造基准数据；使用范围或说明"因行业而异"