What I shipped
About 100 posts over twelve months — mix of long-form deep dives (1500-3000 words), medium technical pieces (600-1000), and short commentary (300-600). Posted across LinkedIn, the personal site, and a couple of guest spots.
What compounded vs what didn’t:
What compounded
Deeply technical posts with code. The “how I built X in Go” posts brought the highest-quality readers — other senior engineers, hiring managers, conference organisers. The audience was small per post but high-signal.
Series with a coherent narrative. Posts that built on each other (“Series 1: the 5 patterns for Y”) brought repeat readers. Once someone read post 1, they’d often read 2 through 5. The series effect is real.
Specific numbers from production. “57% cost reduction at Tata Group” outperformed “FinOps best practices” by an order of magnitude on engagement and inbound conversation. Specifics build credibility; generalities don’t.
Counter-intuitive claims with evidence. “Why I picked the boring stack” did better than “How we built a modern stack” because the boring claim required justification. Surprise drives attention; the evidence holds it.
What didn’t compound
Generic listicles. “10 tips for Go developers” got the most views but the lowest conversation. The audience was wide and shallow; few of them came back. I stopped writing these after a few attempts.
Hot-take opinions without depth. A spicy take on AI hype got attention for a week and disappeared. The same effort spent on a deep technical piece had a half-life of months.
Conference recap posts. “Notes from KubeCon” got polite engagement and then nothing. The conference recording itself outranks any blog post about it on every search.
Brand-name praise pieces. “Why I love product X” — even when sincere — read as marketing. Audiences are sensitive to it.
What I’d tell someone starting
Pick a beat. “Senior Go engineer in regulated AI” is my beat. Posts that fit it land; posts that don’t drift. A clear beat helps the reader know what to expect; helps the writer know what to skip.
Ship before perfect. A post drafted today and shipped tomorrow beats a draft polished for two weeks. The discipline of weekly publishing is more valuable than the marginal quality improvement of polishing.
Reuse the same work three ways. A long-form post becomes a LinkedIn post becomes a tweet. Same idea; different forms; reach different audiences. The cost of writing the second and third forms is small after the first.
Anchor to code. Posts with grep-able code references hold up. Posts without them get dated. The reader can verify; the writer is forced to be precise.
Don’t read your own analytics too often. Once a quarter is fine. Weekly is destructive — you start optimising for vanity metrics instead of compounding ones.
What I’d cut from my own habit
In retrospect, I wrote too much in three categories:
- Conference recaps. Time would have been better spent on technical pieces.
- AI hype responses. The market self-corrects; my contribution didn’t add much.
- Polished interviews / hiring advice. Other writers cover this better; my comparative advantage isn’t there.
If I’d cut these and reallocated to technical depth, the compounding would have been ~30% greater.
What I learned about audience
The audience for technical writing isn’t homogeneous. There are three rough clusters:
- Other senior engineers. They read for the technical depth. They engage by sharing with peers; they don’t usually comment.
- Hiring managers. They read for evidence of capability. They engage by saving for later and reaching out months later.
- Junior engineers learning. They read for patterns to copy. They engage by asking questions.
Each cluster values different things. The senior engineers want depth; the hiring managers want specific accomplishments; the juniors want explanations. A post can serve more than one cluster; the best posts serve all three.
What twelve months changed
The compounding is real. A post from month 1 still gets occasional reads; a post from month 11 starts a slower compound. The body of work matters more than any single post.
The discipline is the asset. The audience is the byproduct.
For year two, the plan: same beat; deeper code anchoring; less hot-take; more series. The boring rule applies to writing too — depth and consistency compound; cleverness and breadth don’t.