Subscribers, Not Followers: Auditing a Month of Notes Performance
Objective: Given a month of synthetic Notes performance data (likes, restacks, and subscriber conversions per Note), apply the lesson's core distinction between vanity engagement and subscriber-converting engagement to decide which Notes to repeat and which to stop making.
You run content for Stitch Fix's editorial newsletter arm on Substack, the San Francisco personalized-styling company that went public via NASDAQ IPO in 2017. Marketing wants to know which of last month's 12 Notes actually grew the subscriber base versus just farming likes.
Rank 12 Notes by subscribers-per-Note rather than likes, flag the one high-like, low-subscriber Note as a false positive, and recommend which 3 formats to repeat next month.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, and a calculated column is all this analysis needs
The process
1 step
Step 01 of 01
The lesson's algorithm-signals section states Notes that drive subscribe clicks get more distribution than Notes that just collect likes, and the Key Takeaways section warns a 500-like Note that drives 5 subscribers loses to a 50-like Note that drives 30.
Note #7 has 940 likes but only 6 new subscribers. Note #3 has 180 likes and 61 new subscribers. Which one gets repeated next month?
Procedure
- Import the 12 rows with columns Note, Format, Likes, Restacks, New Subscribers
- Add a calculated column: (New Subscribers / Likes) * 100
- Sort descending by that calculated column, not by raw Likes
- Flag any Note in the top 3 by Likes that falls outside the top 3 by the calculated column as a false positive
- Cross-reference the top 3 by conversion rate against their Format column to find the repeatable pattern
Note #7 (Poll, meme format): 940 likes, 55 restacks, 6 subscribers -> 0.6 per 100 likes. FALSE POSITIVE, high likes, low conversion. Note #3 (Generosity Note): 180 likes, 41 restacks, 61 subscribers -> 33.9 per 100 likes. TOP PERFORMER. Note #9 (Vulnerable confession): 210 likes, 18 restacks, 44 subscribers -> 20.9 per 100 likes. Note #1 (Contrarian take): 305 likes, 29 restacks, 38 subscribers -> 12.5 per 100 likes. ... 8 more rows, sorted by conversion rate
Healthy
The repeat list is built from the conversion-rate ranking, generosity Notes and vulnerable confessions dominate the top 3.
Unhealthy
The repeat list is built from raw like counts, which would put the poll-meme Note back in rotation next month.
What this means
Likes measure attention, subscribers-per-100-likes measures whether that attention converts to the platform's actual currency, an owned email address.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Team keeps making poll and meme Notes because they get the most likes | Switch the monthly reporting metric from total likes to subscribers-per-100-likes | 5 min |
| No one can tell which format to double down on | Sort the export by conversion rate before every monthly content review | 30 min |
Final deliverable
A ranked table of 12 Notes by subscribers-per-100-likes, with the false-positive Note flagged and a 3-format repeat recommendation.
See a reference example
Peloton Interactive, Notes performance review, October REPEAT NEXT MONTH 1. Generosity Note (33.9 subs/100 likes) 2. Vulnerable confession (20.9 subs/100 likes) 3. Contrarian take (12.5 subs/100 likes) FLAGGED FALSE POSITIVE Poll/meme Note: 940 likes but only 0.6 subs/100 likes, high engagement, near-zero funnel value DO NOT REPEAT Behind-the-scenes screenshot with no CTA framing: 4.1 subs/100 likes
Success criteria
You're done when you can:
- Correctly identifies the high-like, low-conversion Note as a false positive
- Ranks all 12 Notes by the calculated subscribers-per-100-likes metric
- Recommends exactly 3 formats to repeat, matching the top 3 by conversion rate