Tracking Performance

A content pipeline without feedback is just a content factory. In this lesson, you'll set up performance tracking so you can see what's working and automatically improve your output over time.

Why Track?

After a few weeks of running the pipeline, you'll have 10-20 published pieces. Some will resonate, others won't. Without data, you're guessing. With tracking, you can:

See which headlines drive the most clicks
Identify your best-performing platform
Feed winning patterns back into the pipeline

Set up the performance tracker

Create a cron job that collects engagement metrics from your connected platforms:

openclaw cron add \
  --name "Content Analytics" \
  --cron "0 20 * * *" \
  --tz "America/New_York" \
  --session isolated \
  --message "Collect engagement metrics for all content published in the last 7 days. For each piece, report:
- Platform
- Title/hook
- Impressions or views
- Engagement (likes, comments, shares, clicks)
- Engagement rate (engagement / impressions)

Then provide:
1. Top 3 best-performing pieces this week
2. Bottom 3 underperformers
3. One pattern you notice (topic, format, or timing)
4. One specific suggestion for next week's content

Format as a clean report with emoji headers." \
  --source "twitter,linkedin,wordpress,convertkit" \
  --announce \
  --channel telegram \
  --to "CHAT_ID"

This runs every evening at 8 PM and delivers a performance report to your chat.

Create a weekly digest

For a higher-level view, add a weekly summary:

openclaw cron add \
  --name "Weekly Content Report" \
  --cron "0 9 * * 1" \
  --tz "America/New_York" \
  --session isolated \
  --message "Generate a weekly content performance report. Compare this week to last week. Include:

📊 Overview:
- Total pieces published
- Total impressions across all platforms
- Overall engagement rate
- Best day for engagement

🏆 Winners:
- Top post by engagement rate
- Top post by total reach
- Most commented piece

📈 Trends:
- Which topics performed best
- Which platform grew most
- Best time slot for posting

🎯 Recommendations:
- 3 topic ideas based on what worked
- Suggested schedule adjustments
- One experiment to try next week

Use data from the last 14 days for comparison." \
  --source "twitter,linkedin,wordpress,convertkit" \
  --announce \
  --channel telegram \
  --to "CHAT_ID"

Monday mornings

The weekly report arrives Monday at 9 AM — right before you start ideating for the week. Use the recommendations to guide your /idea prompts.

Feed insights back into the pipeline

This is the real power move: let performance data improve your content automatically.

Create a context file that the pipeline references:

openclaw analytics export \
  --format context \
  --days 30 \
  --output "~/.openclaw/content-insights.md"

This generates a file like:

## Content Performance Insights (Last 30 days)

### What Works
- How-to posts outperform opinion pieces by 2.3x
- Posts with specific numbers in headlines get 40% more clicks
- Tuesday LinkedIn posts get 60% more engagement than Friday

### Top-Performing Headlines
1. "5 AI Tools Under $50/Month..." (4.2% engagement)
2. "Why I Stopped Using ChatGPT for..." (3.8% engagement)
3. "The 15-Minute Morning Routine..." (3.5% engagement)

### Avoid
- Generic headlines ("The Future of AI")
- Posts longer than 1500 chars on LinkedIn
- Publishing on Saturday (lowest engagement day)

Now link this to your pipeline agents:

openclaw agent edit "Blog Writer" \
  --context-file "~/.openclaw/content-insights.md"

openclaw agent edit "Social Variants" \
  --context-file "~/.openclaw/content-insights.md"

Your agents now consider past performance when generating new content. The insights file auto-updates weekly.

Set up alerts for high performers

Get notified immediately when something takes off:

openclaw alert add \
  --name "Viral Alert" \
  --condition "engagement_rate > 5% OR impressions > 10000" \
  --check-interval "1h" \
  --source "twitter,linkedin" \
  --message "🚀 Your post is performing well! Consider:
1. Replying to comments to boost engagement
2. Creating a follow-up post on the same topic
3. Boosting the post if the platform supports it" \
  --channel telegram \
  --to "CHAT_ID"

Ride the wave

When a post goes viral (relative to your audience), the best move is to engage quickly. Reply to every comment within the first 2 hours — it signals to the algorithm that this is an active conversation.

The Complete Pipeline

Here's your full content pipeline from start to finish:

  💡 /idea in chat
       │
       ▼
  📋 Content Brief (auto-generated, queued)
       │
       ▼
  ✍️  Blog Post + Social Variants (pipeline generates all)
       │
       ▼
  👀 Review notification (30 min before publish)
       │
       ▼
  📅 Scheduled publishing (Mon-Fri, platform-optimized)
       │
       ▼
  📊 Daily + weekly performance reports
       │
       ▼
  🔄 Insights fed back into pipeline (auto-improving)

Time investment: ~15 minutes per content piece (review + approve), plus a few minutes reading your weekly report.

Output: 3-5 blog posts per month, 15-20 social posts, 4 newsletters — all consistent in voice and improving over time.

Going Further

Generate multiple headline variants and test them:

openclaw agent edit "Blog Writer" --append-message "
Also generate 3 alternative headlines. I'll pick the best one before publishing."

After a month of data, you'll know which headline patterns consistently win.

Extend the pipeline to repurpose older content:

openclaw cron add \
  --name "Content Repurposer" \
  --cron "0 10 * * 3" \
  --message "Review content from 30-60 days ago. Pick the best performer and create:
1. An updated version with new data
2. A different angle on the same topic
3. A 'Part 2' that goes deeper on the most-commented aspect"

Track what's working for others in your niche:

openclaw agent add \
  --name "Competitor Watch" \
  --message "Analyze the top 5 posts this week from these accounts: @competitor1, @competitor2, @competitor3. What topics and formats are getting engagement? Suggest 2 content ideas inspired by (but not copying) their best work."

Course Complete

You've built a full content pipeline that:

Captures ideas from chat
Generates blog posts and social variants
Publishes on schedule across platforms
Tracks performance automatically
Improves itself with feedback data

The pipeline gets better over time — every week of performance data makes the next week's content more targeted.

Knowledge Check

What's the most valuable part of feeding analytics back into the pipeline?