newsletter-publishing

新闻简报发布

面向新闻业和学术界的电子邮件简报创建与管理实用工作流。

适用场景

• 从头开始创建新闻简报
• 为新闻内容设计电子邮件模板
• 构建订阅者列表并进行细分
• 分析新闻简报的效果指标
• 制定新闻简报的编辑日历
• 在不同新闻简报平台之间迁移
• 提升送达率和打开率

新闻简报架构

内容策略框架

## Newsletter strategy document

### Core identity
- **Name**:
- **Tagline** (one line):
- **What readers get**: [specific value proposition]
- **Frequency**: [ ] Daily [ ] Weekly [ ] Bi-weekly [ ] Monthly

### Target audience
- Primary reader:
- What they care about:
- Why they'll subscribe:
- What they'll do with this info:

### Content pillars
1. [Core topic 1] - [how often]
2. [Core topic 2] - [how often]
3. [Recurring feature] - [how often]

### Voice and tone
- Formal ↔ Conversational: [1-5]
- Serious ↔ Light: [1-5]
- Reported ↔ Personal: [1-5]

### Success metrics (first 6 months)
- Subscriber goal:
- Target open rate:
- Target click rate:

单期简报结构模板

## [Newsletter Name] - Issue #[XX]
**Date**: [Date]
**Subject line**: [Subject]
**Preview text**: [First 50-90 characters readers see]

---

### Opening hook
[2-3 sentences that make readers want to keep reading]

### Main story
[Your primary content - 300-600 words for most newsletters]

### Secondary items (if applicable)
- **Quick hit 1**: [Brief item with link]
- **Quick hit 2**: [Brief item with link]

### Recurring section
[Weekly column, data point, recommendation, etc.]

### Sign-off
[Personal note, call to action, or preview of next issue]

---

**Unsubscribe** | **Preferences** | **Forward to a friend**

技术实现

HTML 电子邮件模板（响应式）

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8">
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  <title>{{newsletter_name}}</title>
  <style>
    /* Reset styles for email clients */
    body { margin: 0; padding: 0; width: 100%; }
    table { border-collapse: collapse; }
    img { border: 0; display: block; }

    /* Responsive container */
    .container {
      max-width: 600px;
      margin: 0 auto;
      font-family: Georgia, serif;
      font-size: 18px;
      line-height: 1.6;
      color: #333;
    }

    /* Dark mode support */
    @media (prefers-color-scheme: dark) {
      .container { background-color: #1a1a1a; color: #e0e0e0; }
      a { color: #6db3f2; }
    }

    /* Mobile styles */
    @media only screen and (max-width: 480px) {
      .container { padding: 15px !important; }
      h1 { font-size: 24px !important; }
    }
  </style>
</head>
<body>
  <table role="presentation" width="100%">
    <tr>
      <td align="center" style="padding: 20px;">
        <div class="container">
          
          <table width="100%">
            <tr>
              <td style="padding-bottom: 20px; border-bottom: 2px solid #333;">
                <h1 style="margin: 0;">{{newsletter_name}}</h1>
                <p style="margin: 5px 0 0; color: #666;">{{issue_date}}</p>
              </td>
            </tr>
          </table>

          
          <table width="100%">
            <tr>
              <td style="padding: 30px 0;">
                {{content}}
              </td>
            </tr>
          </table>

          
          <table width="100%">
            <tr>
              <td style="padding-top: 20px; border-top: 1px solid #ddd; font-size: 14px; color: #666;">
                <p>You're receiving this because you subscribed to {{newsletter_name}}.</p>
                <p>
                  <a href="{{unsubscribe_url}}">Unsubscribe</a> |
                  <a href="{{preferences_url}}">Update preferences</a>
                </p>
              </td>
            </tr>
          </table>
        </div>
      </td>
    </tr>
  </table>
</body>
</html>

Python 新闻简报发送器

from dataclasses import dataclass, field
from datetime import datetime
from typing import List, Dict, Optional
from enum import Enum
import hashlib

class SubscriberStatus(Enum):
    ACTIVE = "active"
    UNSUBSCRIBED = "unsubscribed"
    BOUNCED = "bounced"
    COMPLAINED = "complained"

@dataclass
class Subscriber:
    email: str
    name: Optional[str] = None
    subscribed_at: datetime = field(default_factory=datetime.now)
    status: SubscriberStatus = SubscriberStatus.ACTIVE
    tags: List[str] = field(default_factory=list)
    custom_fields: Dict = field(default_factory=dict)

    @property
    def hash_id(self) -> str:
        """Generate unique ID for unsubscribe links."""
        return hashlib.md5(self.email.encode()).hexdigest()[:12]

@dataclass
class NewsletterIssue:
    subject: str
    preview_text: str
    html_content: str
    plain_text: str
    scheduled_at: Optional[datetime] = None
    sent_at: Optional[datetime] = None
    issue_number: int = 0

    # Metrics
    sent_count: int = 0
    delivered_count: int = 0
    opened_count: int = 0
    clicked_count: int = 0
    bounced_count: int = 0
    unsubscribed_count: int = 0

    @property
    def open_rate(self) -> float:
        if self.delivered_count == 0:
            return 0.0
        return (self.opened_count / self.delivered_count) * 100

    @property
    def click_rate(self) -> float:
        if self.delivered_count == 0:
            return 0.0
        return (self.clicked_count / self.delivered_count) * 100

class NewsletterManager:
    """Core newsletter operations."""

    def __init__(self, name: str):
        self.name = name
        self.subscribers: List[Subscriber] = []
        self.issues: List[NewsletterIssue] = []

    def add_subscriber(self, email: str, name: str = None,
                       tags: List[str] = None) -> Subscriber:
        """Add new subscriber with double opt-in pending."""
        sub = Subscriber(
            email=email.lower().strip(),
            name=name,
            tags=tags or []
        )
        self.subscribers.append(sub)
        return sub

    def segment_subscribers(self, tags: List[str] = None,
                           min_engagement: float = None) -> List[Subscriber]:
        """Get subscribers matching criteria."""
        active = [s for s in self.subscribers
                  if s.status == SubscriberStatus.ACTIVE]

        if tags:
            active = [s for s in active
                     if any(t in s.tags for t in tags)]

        return active

    def calculate_engagement_score(self, subscriber: Subscriber) -> float:
        """Score subscriber engagement 0-100."""
        # Implementation would track opens/clicks per subscriber
        return 50.0  # Placeholder

订阅者管理

列表清理工作流

from datetime import datetime, timedelta

def clean_subscriber_list(manager: NewsletterManager,
                         inactive_threshold_days: int = 180) -> dict:
    """Identify and handle inactive subscribers."""
    cutoff = datetime.now() - timedelta(days=inactive_threshold_days)

    results = {
        'total': len(manager.subscribers),
        'active': 0,
        'inactive': [],
        'bounced': [],
        'unsubscribed': []
    }

    for sub in manager.subscribers:
        if sub.status == SubscriberStatus.BOUNCED:
            results['bounced'].append(sub.email)
        elif sub.status == SubscriberStatus.UNSUBSCRIBED:
            results['unsubscribed'].append(sub.email)
        elif sub.status == SubscriberStatus.ACTIVE:
            # Check last engagement
            engagement = manager.calculate_engagement_score(sub)
            if engagement < 10:  # Very low engagement
                results['inactive'].append(sub.email)
            else:
                results['active'] += 1

    return results

def run_reengagement_campaign(inactive_subscribers: List[str]) -> None:
    """Send win-back campaign to inactive subscribers."""
    # Send "We miss you" campaign
    # If no engagement after 2 attempts, mark for removal
    pass

订阅者细分

## Recommended segments

### By engagement
- **VIPs**: Open rate > 80%, always click
- **Engaged**: Open rate 40-80%
- **Casual**: Open rate 10-40%
- **At-risk**: Haven't opened in 90 days
- **Inactive**: Haven't opened in 180 days

### By interest (tag-based)
- Topic preferences from signup
- Content they've clicked
- Surveys/polls they've answered

### By source
- Organic (website signup)
- Referral (forwarded by friend)
- Social media
- Paywall/registration wall

主题行优化

高绩效模式

## Subject line formulas that work

### For news/journalism
- **Breaking format**: "Breaking: [Concise news]"
- **Numbers**: "[X] things we learned about [topic]"
- **Question**: "Why did [entity] do [thing]?"
- **Direct**: "[Topic]: What you need to know"

### For analysis/opinion
- **Take**: "The real story behind [event]"
- **Contrarian**: "Why everyone is wrong about [topic]"
- **Insider**: "What [industry] insiders know about [topic]"

### What to avoid
- ALL CAPS
- Excessive punctuation!!!
- Clickbait that doesn't deliver
- Spam trigger words (FREE, URGENT, ACT NOW)
- Misleading preview text

A/B 测试框架

import random
from typing import List, Tuple

def ab_test_subject_lines(subscribers: List[Subscriber],
                         subject_a: str,
                         subject_b: str,
                         test_percentage: float = 0.2) -> dict:
    """
    Test two subject lines on subset before full send.
    """
    test_size = int(len(subscribers) * test_percentage)
    test_group = random.sample(subscribers, test_size)

    # Split test group
    half = len(test_group) // 2
    group_a = test_group[:half]
    group_b = test_group[half:]

    remaining = [s for s in subscribers if s not in test_group]

    return {
        'group_a': {
            'subject': subject_a,
            'subscribers': group_a,
            'size': len(group_a)
        },
        'group_b': {
            'subject': subject_b,
            'subscribers': group_b,
            'size': len(group_b)
        },
        'remaining': {
            'subscribers': remaining,
            'size': len(remaining),
            'note': 'Send winner to this group after test period'
        },
        'test_duration_hours': 4
    }

提升送达率的最佳实践

电子邮件身份验证设置

## DNS records for deliverability

### SPF record

v=spf1 include:_spf.{{esp_sending_domain}} ~all

### DKIM
- Generate keys through your ESP
- Add TXT record with public key
- Verify signature is applied to outgoing mail

### DMARC

v=DMARC1; p=quarantine; rua=mailto:[email protected]

### Checklist before sending
- [ ] SPF, DKIM, DMARC configured
- [ ] Sending domain warmed up
- [ ] List is clean (no hard bounces)
- [ ] Unsubscribe link works
- [ ] Physical address in footer (CAN-SPAM)
- [ ] Test email received in inbox (not spam)

Gmail、Yahoo 和 Outlook 的批量发件人要求

批量发件人必须满足 Gmail 和 Yahoo 于 2024 年 2 月推出的身份验证和取消订阅要求。Microsoft Outlook 于 2025 年 5 月采用了类似的一套要求。自 2025 年 11 月起，Gmail 会以永久性 5xx 错误拒收不合规邮件，而不是暂缓处理；不合规的新闻简报现在会被硬退信。

适用对象。“批量发件人”是指每天向 Gmail 地址发送超过 5,000 封邮件的发件人。5,000 封的门槛按所有子域名共用的主域名计算，而不是按发送子域名分别计算。例如，某新闻编辑部从 transactional.example.com 每天发送 2,500 封，从 news.example.com 每天发送 2,500 封，就超过了门槛。

要求：

• 对发送域进行 SPF 和 DKIM 身份验证。 两者都必须通过验证。仅使用 SPF 已不再足够。DKIM 密钥长度必须至少为 1024 位。
• DMARC 策略至少设为 p=none。 生产环境中的发件人应在对齐后改用 p=quarantine 或 p=reject。
• 域名对齐。 SPF 或 DKIM 其中一项必须与可见 From: 标头中的组织域名对齐，不要求两者都对齐。允许使用宽松对齐模式。
• 一键取消订阅（RFC 8058）。 邮件必须包含带 HTTPS URL 的 List-Unsubscribe 标头，以及 List-Unsubscribe-Post: List-Unsubscribe=One-Click 标头。HTTPS 端点必须在两天内处理取消订阅请求，且不要求用户登录。邮件正文中也必须显示取消订阅链接。
• 垃圾邮件投诉率低于 0.3％，通过 Google Postmaster Tools 衡量。Google 建议将目标上限设为 0.1％；持续高于 0.3％ 会触发拒收。
• 发送 IP 必须有有效的 PTR 记录（反向 DNS）。 正向和反向 DNS 必须匹配。
• 入站连接必须使用 TLS（Google 自 2023 年 12 月起的要求）。

运营影响：

大多数可靠的 ESP 会在发送域通过验证后，处理身份验证、标头和 TLS。仍需由运营方负责的两项工作是控制投诉率和处理取消订阅；重新互动活动和及时清理列表对此至关重要。重新唤回长期不活跃的订阅者风险很高，正因为他们的投诉率是活跃订阅者的 5～10 倍；一次不当的重新互动活动就可能使投诉率超过 0.3％，导致整个发送域的邮件遭到拒收。

参考资料：

• Google，《Email sender guidelines》，support.google.com/mail/answer/81126
• Google，《Email sender guidelines FAQ》，support.google.com/a/answer/14229414
• Yahoo，《Sender Best Practices》，senders.yahooinc.com/best-practices/

垃圾邮件评分检查清单

## Before you send

### Content checks
- [ ] No spam trigger words
- [ ] Text-to-image ratio good (mostly text)
- [ ] All links are to reputable domains
- [ ] No URL shorteners (use full links)
- [ ] Plain text version included

### Technical checks
- [ ] From address matches sending domain
- [ ] Reply-to address is monitored
- [ ] Preheader text is set
- [ ] Images have alt text
- [ ] Links are not broken

分析与优化

关键指标仪表板

from dataclasses import dataclass

@dataclass
class NewsletterAnalytics:
    """Track newsletter performance over time."""

    issue: NewsletterIssue

    def summary(self) -> dict:
        return {
            'issue_number': self.issue.issue_number,
            'sent': self.issue.sent_count,
            'delivered': self.issue.delivered_count,
            'delivery_rate': self._pct(self.issue.delivered_count,
                                       self.issue.sent_count),
            'opens': self.issue.opened_count,
            'open_rate': self.issue.open_rate,
            'clicks': self.issue.clicked_count,
            'click_rate': self.issue.click_rate,
            'click_to_open': self._pct(self.issue.clicked_count,
                                       self.issue.opened_count),
            'unsubscribes': self.issue.unsubscribed_count,
            'unsubscribe_rate': self._pct(self.issue.unsubscribed_count,
                                          self.issue.delivered_count),
        }

    def _pct(self, numerator: int, denominator: int) -> float:
        if denominator == 0:
            return 0.0
        return round((numerator / denominator) * 100, 2)

# Benchmarks (journalism newsletters)
BENCHMARKS = {
    'open_rate': {'good': 40, 'excellent': 55},
    'click_rate': {'good': 4, 'excellent': 8},
    'unsubscribe_rate': {'acceptable': 0.5, 'concerning': 1.0},
}

平台比较

平台	最适合	定价模式	主要功能
Substack	以作者为中心、付费订阅	收入分成	内置支付功能
Buttondown	开发者、极简需求	按订阅者计费	原生支持 Markdown
Ghost	出版商、会员服务	固定费用	内置完整 CMS
beehiiv	注重增长	免费增值	推荐工具
Kit（原名 ConvertKit）	内容创作者	按订阅者计费	自动化
Mailchimp	小型组织	分级定价	简单易用的模板

法律合规

CAN-SPAM 要求（美国）

- [ ] Accurate "From" name and email
- [ ] Non-deceptive subject line
- [ ] Physical postal address included
- [ ] Working unsubscribe mechanism
- [ ] Unsubscribe honored within 10 days
- [ ] No purchased lists

GDPR 要求（欧盟订阅者）

- [ ] Explicit consent obtained (not pre-checked)
- [ ] Clear privacy policy linked
- [ ] Easy unsubscribe process
- [ ] Data export available on request
- [ ] Data deletion on request
- [ ] Record of consent stored

相关技能

• web-scraping，自动收集简报内容
• data-journalism，在邮件中加入数据可视化
• academic-writing，撰写清晰、有条理的内容
• newsroom-style，确保简报文案符合 AP Style
• fact-check-workflow，在内容发送到订阅者的收件箱之前核实说法
• ai-writing-detox，从草稿中去除 AI 写作模式

---

技能元数据

字段	值
版本	1.0.0
创建日期	2025-12-26
更新日期	2026-05-08
作者	Joe Amditis
领域	出版、营销
复杂度	中级