Ten essays in: what compounded
A short retrospective at the 10-post mark. Which essays drove signups, which got picked up, which were boring and necessary anyway.
The three that compounded
Out of ten essays, three carried disproportionate weight:
The OG card factory 17 dynamic share-cards out of one Vercel function. The technical post that producers actually shared because the share-cards are visible artifacts. Why we rank by saves, not followers The philosophical post tying the leaderboard to a metric thesis. Producers who ranked top-20 cited it when sharing their profile. Three callers, one RPC The narrowest, nerdiest post — and the one that most clearly stated a generalizable rule. The kind of post a backend engineer bookmarks.What didn't work
The retrospective posts (100 days, 30 features) underperformed expectation. They read well to people already in the orbit but didn't pull anyone new in. Lesson: retrospectives are for the existing audience, not for acquisition.
The "how the algorithm works" post landed somewhere in the middle. Useful for the curious, not actionable enough to share.
What worked anyway
The boring infrastructure posts (12 functions, building the scraper) didn't drive signups, but they drove the right kind of inbound — engineers asking how we structured something. Those conversations turned into two product partnerships and a couple of producer claims.
"Inbound from indie hackers" isn't a metric in the dashboard. But the people who showed up that way were the most engaged 1% of all signups.
Retrospectives are for the existing audience. Pattern essays are for the future audience.
The pattern that's emerging
Three traits separate the essays that worked from the ones that didn't:
- Generalizable rule baked in. "Two callers is a coincidence. Three is a contract you haven't written yet" travels further than "we made an RPC."
- Tied to a visible artifact. The OG card post worked because the cards themselves are evidence — every share is implicit advertising for the post.
- A specific aesthetic preference. "Save count beats follower count" is a taste claim, not a feature description. Taste claims travel.
The next ten essays should lean into all three.
What's next
A bunch more SEO pages — we just shipped 25 across producer lanes (Metro Boomin, Kanye, J. Cole, Young Thug, Juice WRLD, Trippie, 21 Savage, Roddy Ricch, Ice Spice, Sexyy Red, Lil Uzi, and a dozen genre lanes). Two or three new dynamic OG cards on the docket. And probably another nerdy infrastructure essay when something narrow enough to push down is sitting at three callers.