sentiment-monitoring

センチメントモニタリング（舆情监控）

概要

センチメントモニタリングとは、小紅書でのブランドへの言及、コメント、会話を体系的に追跡・分析し、世間の認識を把握して評判リスクを早期に特定し、前向きなブランドイメージを維持するために先回りして対応することです。

使用する場面

次の場合に使用します：

• ブランドの健全性と世間の認識をモニタリングする
• マーケティングキャンペーン中のセンチメントを追跡する
• 否定的なフィードバックを早期に特定し、対処する
• PR 危機や評判の問題に対応する
• 競合他社に対するセンチメントやポジショニングを分析する
• ブランドセンチメントの経時的な変化を測定する
• 顧客からの苦情や懸念に対応する

次の場合には使用しないでください：

• 基本的なエンゲージメント指標をモニタリングする（data-analytics を使用）
• 日々のコンテンツパフォーマンスを追跡する（data-analytics を使用）
• 競合他社のコンテンツ戦略を分析する（competitive-analysis を使用）

基本パターン

変更前（世間の反応が見えない状態）：

❌ "No idea what people are saying about our brand"
❌ "Negative comment goes viral before we notice"
❌ "Can't measure if our campaigns improved perception"
❌ "Reputation crisis catches us completely off guard"

変更後（先回りしたセンチメント管理）：

✅ "Real-time alerts: sentiment dropped 15%, investigate immediately"
✅ "Identified negative trend emerging, addressed before crisis"
✅ "Campaign improved sentiment from 65% to 82% positive"
✅ "Comprehensive sentiment dashboard informs every decision"

3 段階のモニタリングフレームワーク：

1. モニタリング - 言及、コメント、会話をリアルタイムで追跡する
2. 分析 - センチメントを分類し、テーマを特定し、傾向を追跡する
3. 対応 - 先回りして関与し、問題に対処し、肯定的な意見を広める

クイックリファレンス

センチメントの種類	指標	ベンチマーク	対応の優先度
肯定的	賞賛、推薦、好意	目標 >75%	広めて、ユーザーに感謝する
中立的	質問、情報の問い合わせ	15-25%	情報を提供し、説明する
否定的	苦情、批判、問題	<15%	24 時間以内に対処する
危機的	拡散する否定的な意見、PR の悪夢	<5%	直ちに対応する

実装

ステップ 1：モニタリング基盤を整備する

モニタリング範囲を定義する：

What to Monitor:
1. Brand Mentions
   - Official brand name and variations
   - Product names and variations
   - Hashtags associated with brand
   - Account handles and nicknames

2. Executive and Spokespeople
   - Founder, CEO, key spokespeople
   - Influencers representing brand
   - Brand ambassadors and partners

3. Product Categories
   - Category terms where brand competes
   - Generic product terms + brand
   - Competitor mentions for comparison

4. Campaign Monitoring
   - Campaign-specific hashtags and keywords
   - Campaign launch to 30 days post-end
   - Influencer campaign mentions

Monitoring Frequency:
- Real-time: Crisis situations, major campaigns
- Hourly: During active campaigns or launches
- Daily: Standard brand health monitoring
- Weekly: Competitive analysis, trend reporting

モニタリングツールを設定する：

Platform Native Tools:
✅ Xiaohongshu Professional Account Analytics
   - Track comment sentiment on your posts
   - Monitor mention notifications
   - Review direct messages and inquiries

✅ Native Search and Alerts
   - Save searches for brand name
   - Monitor hashtag usage
   - Track location-based mentions

Third-Party Tools:
✅ Huitun Data (灰豚数据)
   - Brand mention tracking
   - Sentiment analysis over time
   - Competitor comparison

✅ Chanmama (蝉妈妈)
   - Real-time comment monitoring
   - Influencer sentiment tracking
   - Crisis alert system

✅ Social Listening Platforms
   - Mention.com, Brandwatch, etc.
   - Cross-platform sentiment tracking
   - Automated alerts and reports

Custom Monitoring:
✅ Manual review spreadsheet
   - Daily sentiment logging
   - Issue categorization
   - Response tracking

✅ Google Alerts for web mentions
   - News articles featuring brand
   - Blog posts and reviews
   - Forum discussions

ステップ 2：センチメントのベースラインを確立する

現在のブランドセンチメントを測定する：

Baseline Metrics to Track:
1. Overall Sentiment Score
   - Percentage of positive vs negative mentions
   - Track over time (weekly, monthly)
   - Benchmark: >75% positive is healthy

2. Sentiment by Content Type
   - Product posts sentiment
   - Educational content sentiment
   - Promotional content sentiment
   - Personal/lifestyle content sentiment

3. Sentiment by Audience Segment
   - New followers vs long-time followers
   - Different age demographics
   - Different geographic regions

4. Sentiment Drivers
   - What topics generate positive sentiment?
   - What issues trigger negative sentiment?
   - What questions or concerns come up repeatedly?

Baseline Data Collection:
- Analyze last 30 days of comments and mentions
- Classify minimum 500 comments as positive/neutral/negative
- Identify top themes and topics
- Calculate baseline sentiment percentages

センチメント分類ガイド：

Positive Indicators (友好的):
✅ Praise and compliments: "太棒了", "很喜欢", "好漂亮"
✅ Recommendations: "强烈推荐", "姐妹们快冲", "种草了"
✅ Purchase intent: "已经下单", "想要", "在哪里买"
✅ Gratitude: "谢谢分享", "学到了", "很有帮助"
✅ Emotional connection: "爱了爱了", "太心动了", "必须支持"
✅ Positive emojis: ❤️, 🔥, 👍, 😍, ✨

Neutral Indicators (中性的):
❓ Questions: "请问", "怎么", "多少钱", "哪里买"
❓ Information-seeking: "有没有", "是什么", "怎么样"
❓ Tagging friends: "@某某 看看", "@闺蜜"
❓ Factual comments: "收到", "知道了", "好的"

Negative Indicators (负面的):
⚠️ Complaints: "不好用", "踩雷", "失望"
⚠️ Criticism: "太贵了", "不推荐", "别买"
⚠️ Issues: "质量差", "客服差", "物流慢"
⚠️ Comparison: "还是XX品牌好", "不如XX"
⚠️ Skepticism: "广告吧", "像是假的", "不可信"
⚠️ Negative emojis: 👎, 😡, 😞, 💔

Crisis Indicators (危机级别):
🚨 Viral negative post with 1K+ engagements
🚨 Mass complaints about same issue
🚨 News media coverage of negative issue
🚨 Competitor or influencer attacking brand
🚨 Regulatory or legal investigation mentioned

ステップ 3：センチメントをリアルタイムでモニタリングする

毎日のモニタリングルーティン：

Morning Check (9:00 AM):
1. Review overnight mentions and comments
2. Check sentiment alerts and notifications
3. Identify any negative trends or spikes
4. Flag urgent issues for immediate response

Mid-Day Check (2:00 PM):
1. Monitor campaign post performance and sentiment
2. Respond to customer inquiries and complaints
3. Track influencer post sentiment
4. Engage with positive user-generated content

Evening Check (6:00 PM):
1. Final sentiment check of day's activity
2. Log sentiment metrics for daily report
3. Plan next day's responses and engagement
4. Escalate any emerging issues to team

Weekly Review (Friday PM):
1. Analyze sentiment trends for the week
2. Compare to baseline and previous weeks
3. Identify recurring themes or issues
4. Create weekly sentiment report
5. Recommend adjustments to strategy

リアルタイムアラートのトリガー：

Set Up Automated Alerts For:
- Sentiment drops >10% in 24 hours
- Negative comment spike (>50 negative comments in 1 hour)
- Viral negative post (>500 engagements in 1 hour)
- Mass complaints about same issue
- Competitor attack or negative comparison
- News media mention of brand
- Influencer with 100K+ followers mentions negatively
- Any comment with >100 likes that's negative

Alert Escalation:
Yellow Alert (Monitor closely):
- Sentiment down 10-15%
- Cluster of negative comments (not viral)
- Single issue raised by multiple users

Orange Alert (Prepare response):
- Sentiment down 15-25%
- Negative post gaining traction (100-500 engagements)
- Influencer or media mentions issue

Red Alert (Immediate action):
- Sentiment down >25%
- Viral negative post (>500 engagements)
- PR crisis emerging
- Mass complaints or coordinated attacks

ステップ 4：センチメントの傾向を分析する

傾向分析フレームワーク：

Weekly Sentiment Report:
Week of: [Date]

Overall Sentiment:
- Positive: 78% (↑ 3% from last week)
- Neutral: 16% (↓ 2% from last week)
- Negative: 6% (↓ 1% from last week)

Key Insights:
1. Sentiment Trend: [Improving/Stable/Declining]
2. Top Positive Themes: [What people love]
3. Top Negative Themes: [What people complain about]
4. Emerging Issues: [New concerns this week]
5. Resolved Issues: [Issues that improved]

Action Items:
1. [Specific actions to take next week]
2. [Content topics to address concerns]
3. [Customer service improvements needed]
4. [Product feedback to share with team]

センチメントを左右する要因を特定する：

Positive Driver Analysis:
What generates most positive sentiment?
- Product quality and results
- Educational content value
- Authenticity and transparency
- Customer service responsiveness
- Community and engagement

Negative Driver Analysis:
What causes most negative sentiment?
- Product issues (quality, not as expected)
- Price concerns
- Shipping or delivery problems
- Customer service experience
- Misleading marketing or expectations

Gap Analysis:
- Where does reality not match expectations?
- What promises are we not delivering on?
- What competitors do better?
- What improvements would have biggest impact?

競合他社のセンチメントをベンチマークする：

Monitor Competitor Sentiment:
- What's their overall sentiment score?
- What do customers love about them?
- What do customers complain about?
- What advantages do we have in sentiment?
- What can we learn from their sentiment issues?

Positioning Opportunities:
- Competitor weakness = Our opportunity
- Address pain points competitors ignore
- Emphasize strengths where competitors are weak
- Differentiate based on sentiment gaps

ステップ 5：センチメントに先回りして対応する

肯定的なセンチメントを広める：

Engage with Positive Mentions:
✅ Reply to positive comments: "Thank you so much! We're so happy you love it! 💕"
✅ Repost user-generated content to stories with credit
✅ Feature positive reviews in posts (with permission)
✅ Create "thank you" posts highlighting community love
✅ Surprise and delight: Send freebies to super fans
✅ Turn positive feedback into testimonials

Best Practices:
- Respond within 24-48 hours to positive comments
- Personalize responses (not generic copy-paste)
- Ask follow-up questions to deepen engagement
- Invite them to share more or tag friends
- Build relationships with brand advocates

Amplification Tactics:
- "Fan Friday" features of positive community members
- Create highlight of customer love and results
- Share testimonials in ads and marketing
- Identify and nurture potential brand ambassadors

中立的なセンチメントに情報を提供する：

Respond to Questions and Inquiries:
✅ Provide helpful, detailed information
✅ Link to relevant posts or resources
✅ Offer personalized advice when appropriate
✅ Invite them to DM for more complex questions
✅ Follow up to ensure their needs were met

Conversion Opportunities:
- Neutral questions often indicate purchase intent
- "Where can I buy?" = Ready to purchase, make it easy
- "How does this work?" = Needs education to convert
- "Is this right for me?" = Personalized consultation opportunity

Best Practices:
- Respond promptly (within 24 hours)
- Be thorough but concise
- Include clear call-to-action
- Track conversion from question to purchase

否定的なセンチメントに対応する：

Negative Comment Response Framework:
1. Acknowledge Quickly (within 24 hours, ideally 4-8 hours)
   "Hi [Name], we're so sorry to hear about your experience..."

2. Validate Their Experience
   "...we completely understand why you'd feel disappointed..."

3. Take Responsibility (even if not entirely your fault)
   "...we clearly didn't meet expectations here, and that's on us."

4. Offer Solution or Next Step
   "We'd love to make this right. Can you DM us your order details?"

5. Follow Up and Resolve
   - Address the issue privately
   - Provide refund, replacement, or solution
   - Follow up to ensure satisfaction

6. Turn Around if Possible
   - After resolving, ask if they'd reconsider
   - Many will update review or comment positively
   - Creates positive PR out of negative situation

Response Templates:
⚠️ Product Issue:
"Hi [Name], we're so sorry this didn't work for you! We stand behind our products 100%. Please DM us your order info and we'll send a replacement or full refund immediately - whichever you prefer. We want to make this right! 💕"

⚠️ Shipping/Delivery Issue:
"So sorry for the delivery delay, [Name]! That's frustrating. We're looking into your order status right now and will DM you with an update and expedited shipping on us. Thank you for your patience! 🙏"

⚠️ General Complaint:
"Thank you for sharing this feedback, [Name]. We take this seriously and want to address it. Could you DM us more details so we can understand what happened and make it right? We appreciate you bringing this to our attention. 🙏"

⚠️ Unreasonable or Troll:
(For clearly malicious or false complaints, respond once politely, then disengage)
"Hi [Name], we're sorry you feel this way. We'd love to understand more about your experience so we can improve. Please feel free to DM us with details. Wishing you the best! 😊"

Response Best Practices:
- Speed matters: Respond within 24 hours, ideally faster
- Empathy first: Validate their feelings before defending
- Take responsibility: Even if partly customer error
- Offer solutions: Don't just apologize, fix the problem
- Take it offline: DM for complex issues
- Don't argue: Public arguments never help
- Know when to disengage: Trolls won't be satisfied

危機対応プロトコル：

Crisis Identification:
- Viral negative post or mass complaints
- Sentiment drops >25% rapidly
- News media or major influencer covers issue
- Coordinated attack or misinformation

Immediate Response (First Hour):
1. Assemble crisis response team
2. Verify facts and assess situation
3. Determine root cause and responsibility
4. Draft initial response statement

Initial Response (First 4 Hours):
- Acknowledge issue publicly
- Express genuine concern and empathy
- Commit to investigation and resolution
- Provide timeline for next update
- Stop all scheduled promotional content

Example Initial Crisis Post:
"Hi everyone, we're aware of the concerns about [issue] and we're taking this very seriously. We're investigating right now to understand exactly what happened. We'll provide another update by [specific time, e.g., 6 PM today]. We're committed to making this right. Thank you for your patience. 🙏"

Ongoing Crisis Management:
- Regular updates (at least daily)
- Transparency about progress
- Show action being taken, not just words
- Address questions and concerns directly
- Don't delete legitimate criticism (builds trust)

Resolution:
- Clearly explain what happened
- Take responsibility if at fault
- Detail corrective actions implemented
- Offer restitution to affected customers
- Share lessons learned and improvements

Post-Crisis:
- Monitor sentiment recovery
- Rebuild trust through consistent quality
- Document learnings to prevent recurrence
- Update crisis response playbook

ステップ 6：センチメントから得た知見を活用する

製品開発へのフィードバック：

Identify Product Improvements:
- What features or benefits do users praise most?
- What complaints or issues come up repeatedly?
- What use cases do users discover that we didn't anticipate?
- What products do users wish we had?

Feed to Product Team:
- Weekly feedback summary from comments
- Feature request prioritization by sentiment volume
- User pain points and unmet needs
- Competitive advantages we should emphasize

Example Feedback:
"Sentiment analysis shows 87% positive feedback on our moisturizer's texture, but 23% of negative comments mention it's too heavy for oily skin. Opportunity: Create lightweight version for oily skin, or better communicate which skin types it's best for."

コンテンツ戦略の最適化：

Learn What Content Resonates:
- Which topics generate highest positive sentiment?
- Which formats get most engagement and praise?
- Which educational content is most helpful and shared?
- Which promotional content is accepted vs seen as "too salesy"?

Content Adjustments:
- Double down on high-sentiment content types
- Address negative sentiment topics with educational content
- Test new content formats based on user preferences
- Personalize content for different audience segments

Example:
"Tutorial content gets 92% positive sentiment vs 65% for promotional posts. Shift content mix from 60% promotional / 40% educational to 30% promotional / 70% educational."

カスタマーサービスの改善：

Identify Service Pain Points:
- What customer service issues generate complaints?
- Where is the customer journey breaking down?
- What questions or concerns come up repeatedly?
- Which touchpoints generate negative sentiment?

Service Improvements:
- Address top 3 complaint causes systematically
- Create FAQ content for common questions
- Improve response time metrics
- Train customer service team on sentiment trends

Example:
"Sentiment shows 18% of negative comments relate to shipping delays. Partner with faster shipping provider, provide tracking updates, set realistic delivery expectations upfront."

マーケティング戦略の調整：

Optimize Based on Sentiment:
- Pause campaigns generating negative sentiment
- Double down on campaigns with positive response
- Adjust messaging based on feedback
- Target segments where brand sentiment is strongest

Brand Positioning:
- Emphasize strengths that generate positive sentiment
- Address perceived weaknesses head-on
- Differentiate from competitors based on sentiment advantages
- Rebrand if sentiment issues are fundamental to positioning

Example:
"Our authenticity and transparency drive 89% positive sentiment, but price concerns drive 34% of negative comments. Emphasize value and quality in messaging, create price-accessible product line, communicate why premium pricing is justified."

ステップ 7：調査結果を報告し、共有する

社内報告：

Weekly Sentiment Report:
- Overall sentiment score and trend
- Positive/negative themes and examples
- Top issues and recommendations
- Campaign performance and sentiment impact
- Competitive sentiment comparison

Monthly Executive Summary:
- Sentiment trend over 3-6 months
- Impact of marketing initiatives on sentiment
- Brand health vs competitors
- Risk assessment and recommendations
- Resource allocation for improvements

Quarterly Strategic Review:
- Long-term sentiment trends
- Correlation between sentiment and business metrics
- Major reputation risks and opportunities
- Strategic recommendations for brand positioning
- Investment priorities for sentiment improvement

部門横断的な連携：

Share Insights With:
Product Team:
- User feedback and feature requests
- Product improvement priorities
- New product opportunities based on unmet needs

Customer Service:
- Common complaints and response templates
- Training on emerging issues
- Feedback loop for resolved issues

Marketing Team:
- What messaging resonates vs flops
- Campaign performance and sentiment impact
- Content topics and formats that work

Executive Team:
- Brand health dashboard
- Risk assessment and early warning
- Strategic recommendations

Legal/Compliance:
- Emerging regulatory or legal concerns
- Potential liability issues from sentiment
- Crisis preparedness

よくある間違い

間違い	発生する理由	対策
自社が所有するコンテンツだけをモニタリングする	追跡しやすいため	プラットフォーム全体でブランドへの言及をすべてモニタリングする
否定的なフィードバックを無視する	いずれ収まることを期待するため	事態が悪化する前に速やかに対処する
批判に対して防御的に反応する	個人的な攻撃として受け止めるため	共感を示し、責任を持って対応する
否定的なコメントを削除する	フィードをきれいに保ちたいため	問題に透明性を持って対処し、信頼を築く
センチメントのベースラインがない	開始時点の状況が分からないため	キャンペーンや変更の前にベースラインを確立する
センチメントではなく量だけを追跡する	測定しやすいため	量よりセンチメントの質が重要
問題への対応が遅い	対応に追われたり、アラートを見逃したりするため	自動アラートを設定し、迅速な対応を優先する
センチメントの傾向から学ばない	日々の業務に集中しているため	戦略的な知見を得るために傾向を分析する
競合他社のセンチメントを無視する	自社ブランドだけに集中しているため	機会と脅威を把握するために競合他社をモニタリングする
部門間で知見を共有しない	チームが縦割りになっているため	組織全体で広く知見を共有する

実社会への影響

ケーススタディ：感情分析による危機の回避

ある美容ブランドの感情モニタリングが、新製品のロットに関する問題の兆候を検知しました。

検知：

• 1 時間以内に、皮膚の炎症に関する否定的なコメントが 50 件以上投稿されたため、アラートが発動しました
• コメントはロット #XYZ123 に集中していました
• 感情スコアは 3 時間で 82% から 61% に低下しました

迅速な対応：

• 2 時間以内に問題を認める声明を発表しました
• 該当ロットの販売を直ちにすべて停止しました
• 影響を受けたお客様に全額返金を申し出ました
• 製造パートナーとともに調査を開始しました

解決：

• 製造工程における品質管理上の問題を特定しました
• 該当ロットをリコールしました（影響を 2,000 人以上に及ぶ可能性があったところ、200 人に抑えました）
• 新たな品質管理手順を導入しました
• 全プロセスを通じて透明性のあるコミュニケーションを行いました

結果：

• 感情スコアは 2 週間以内に 79% まで回復しました
• お客様から透明性のある対応が高く評価されました
• 影響を受けたお客様の多くが、ブランドの熱心な支持者になりました
• ブランドに恒久的な損害を与えかねない危機の拡散を防ぎました

データに裏付けられた知見：

• 否定的な感情に 24 時間以内に対応するブランドは、顧客維持率が 40% 高くなります
• 問題に透明性をもって対処することは、問題を無視するよりも感情を改善します
• 苦情が満足のいく形で解決されたお客様の 70% は、リピーターになります
• 感情モニタリングは早期検知により、危機の影響を 60～80% 軽減します
• 積極的な感情管理により、6 か月でブランド健全性スコアが 25～35% 向上します
• 感情モニタリングを効果的に行っているブランドは、行っていないブランドより 2～3 倍速く成長します

関連スキル

必須：data-analytics を使用して、感情データを定量的に分析します
必須：community-management を使用して、感情分析の知見に基づきエンゲージメントを図ります
必須：customer-service を使用して、モニタリングで特定された問題を解決します

包括的なモニタリングにおすすめ：

• social-listening - 高度なソーシャルリスニング手法
• crisis-management - PR 危機や評判に関する問題への対応
• customer-feedback - フィードバックを体系的に収集・分析
• competitive-intelligence - 競合他社の感情や戦略をモニタリング
• brand-reputation-management - 評判の積極的な保護と構築