How I Used AI to Crack the Code of Viral Substack Notes (And You Can Too)
A step-by-step guide to analyzing your best-performing Substack notes and creating a repeatable system for engagement
We all know that notes are one of the biggest growth drivers on Substack. But which ones actually work and why?
I’ve tried all kinds of different approaches to posting notes. Some went unnoticed, others unexpectedly got hundreds of likes, and some didn’t get tons of likes but helped me connect with amazing writers who we all support each other now.
Every time I posted a note, I had no clue which one would hit. When I had accumulated several successful notes, I decided to get scientific about it. Instead of just guessing what worked, I fed my 12 best-performing notes to AI and asked it to find the patterns.
The results impressed me! AI is seriously good at analyzing content.
Here’s exactly how I did it, and how you can replicate this process with your own content.
My Experiment: 12 Notes, 1 Clear Question
I gathered my top-performing notes from the past 3 months:
- Note with 706 likes, 41 new subscribers
- Note with 143 likes, 43 new subscribers
- Note with 132 likes, 20 new subscribers
- And 9 others ranging from 40-76 likes, 2-14 new subscribers
So I just pasted all my note texts straight into Claude Sonnet 4 chat (ChatGPT or Perplexity work just as well) along with how many likes each one got and my subscribers count, then asked this simple question:
“Based on these Substack notes, identify three types of templates for the most successful notes.”
Just so you know, I used the free version of Claude. I had to manually copy all the note texts and stats into one document. I tried giving Claude a link to my Substack and asking it to analyze my best posts, but it told me “I’m not able to access external links or browse websites, including Substack profiles.”
During my academia career, I’ve done text analysis for years, so I had pretty high expectations for how this should be done. I’ll admit, Claude exceeded what I expected.
It didn’t just spit out some templates, it did this really solid analysis of my notes without me even asking for that.
It grouped them into three categories and broke down their structure and key elements. It even highlighted patterns in the opening lines I used in my notes. Plus, it identified what factors actually make notes successful.
I’m going to share the full analysis and the three successful note templates below so you can use them too.
You can also do this same thing with your own best-performing notes or posts to figure out what works for you.
Or flip it around - take your posts that barely got any likes and ask AI to analyze what went wrong and why people didn’t connect with them.
3 High-Performing Substack Note Templates: Claude Analysis
Template 1: The Milestone Celebration + Gratitude
Performance Range: 40-706 likes, 2-43 new subscribers
Structure:
1. Announce the milestone (subscriber count, time milestone, or achievement)
2. Share the numbers (posts written, notes published, current stats)
3. Express genuine gratitude to your community
4. Reflect on the journey (what it means beyond numbers)
5. Add celebratory emojis 🎉🥳😃
Example Opening Lines:
- "3 months on Substack! 🥳🎉😀"
- "44 Subscribers! 😃 I'm definitely not winning any speed races..."
- "A three weeks ago, I was daydreaming about hitting 50 subscribers..."
Key Elements:
- Vulnerability about your progress
- Emphasis on community over numbers
- Authentic celebration of small wins
- Personal reflection on growth
Template 2: The Strategy Share + Actionable Tips
Performance Range: 46-76 likes, 3-5 new subscribers
Structure:
1. Hook with a relatable question/observation
2. Share specific, actionable strategies (bulleted list works well)
3. Include timeframes and results for credibility
4. End with encouraging advice or key insight
Example Opening Lines:
- "A few people have asked me in the comments how I got my first 50 subscribers..."
- "What helped me jump from 44 to 96 subscribers:"
- "Wow, I thought it was a myth that after your first 100 subscribers, it's easier to grow..."
Key Elements:
- Specific daily/weekly actions
- Real numbers and timeframes
- Practical advice others can implement
- Personal experience as proof
Template 3: The Community Connection + Engagement Driver
Performance Range: 43-143 likes, 3-43 new subscribers
Structure:
1. Pose a thoughtful question or share a personal pivot/challenge
2. Share your own experience or perspective
3. Invite others to connect and participate
4. Create a clear call-to-action for engagement
Example Opening Lines:
- "Is anyone else here trying something completely new in mid-life?"
- "Dear Substackers, what do you think on using AI for writing?"
- "To help you get more visibility, please drop your link below..."
Key Elements:
- Direct questions to the community
- Personal vulnerability or major life changes
- Clear invitation for others to share
- Supportive, inclusive tone
Universal Success Factors Across All Templates:
Tone & Voice:
- Authentic and conversational (like talking to friends)
- Vulnerable about struggles and slow progress
- Generous with support for others
- Genuinely excited about small wins
Engagement Tactics:
- Use emojis strategically for emotion
- Include specific numbers for credibility
- Ask questions to drive comments
- Share both wins AND challenges
- Focus on community building over self-promotion
Timing & Consistency:
- Post 1-2 notes daily for visibility
- Celebrate milestones at any size
- Share the journey, not just destinations
- Be patient with growth (emphasize "slow and steady")
Community-First Approach:
- Support other creators actively
- Share others' content generously
- Focus on genuine connections over metrics
- Emphasize belonging and "finding your tribe"
How to Replicate This Analysis
Подведем итог, как провести подобный анализ самостоятельно.
Step 1: Gather Your Data
Export your 10-15 best-performing notes. Include:
- The full text
- Like counts
- New subscriber numbers
- Any other engagement metrics
Step 2: Ask AI the Right Questions
Use prompts like:
- “What are the common patterns in these successful posts?”
- “What templates can you identify?”
- “What tone and language patterns repeat?”
- “What structural elements do they share?”
Step 3: Create Your Templates
Ask your AI tool to turn the patterns into fill-in-the-blank templates you can reuse. For example, Milestone Template:
[NUMBER] [MILESTONE] on Substack! 🎉
[STATS LIST]
[HUMBLE ACKNOWLEDGMENT]
[GRATITUDE STATEMENT] ❤️
Step 4: Test and Iterate
Use your templates to create new notes. Track performance. Refine based on what works.
The Surprising Truth About Viral Content
But here’s the thing, even notes that follow these AI-generated templates aren’t guaranteed to go viral.
Sure, there are certain patterns that grab people’s attention. But people can always tell when you’re being authentic and genuine.
So while these frameworks are super helpful for organizing your content, the actual heart of your notes - your stories, feelings, experiences, your real connection with people. That’s what actually makes them work.
Or to put it in AI terms:
- **Authenticity over perfection**
- **Community over self-promotion**
- **Vulnerability over authority**
- **Specific help over generic advice**
I couldn’t write better.
Hope this was helpful! If it was, feel free to buy me a coffee. No pressure at all, but it always makes my day and keeps me motivated to share more good staff. 😃Really appreciate all the support!
Have you ever used AI to analyze your content and plan what to post? How’s that working out for you? I’d love to hear what you think about it.
This is the kind of thing I love AI for
great piece thank you!