buyWords

Writing Insights

What actually works on LinkedIn, read off real posts.

Operator notes from running a roster of voices. No engagement-bait listicles, no growth-hacking gospel (we do neither). Every note here was read off real published posts, and the receipts are on this page: the numbers, and the links to check them.

Real voices
5
Impressions, counted
76,505
Posts dissected below
9
Em dashes in them
0
  • Mechanics1
  • Openers1
  • Employee content1
  • Analytics1
  • Specificity1
  • Distribution1
  • Cadence1
MechanicsField note 014 min read

Why your best post ended four lines too late.

The point landed. Then you kept typing. Where attention actually drops, and why the edit nobody makes is the last paragraph.

The full index

7 field notes
  1. Openers4 min readThe cold open beats the hook.Hooks announce themselves. Cold opens just start. One of them gets read past line one.
  2. Employee content5 min readMany voices, one team: writing for a whole team at scale without the blend.A whole roster of employees, one team, zero blended-into-mush posts. The system that keeps each voice its own.
  3. Analytics5 min readThe metric that actually explains what lands.Likes are a lagging vanity number. The leading signal is velocity: how fast a post catches in its first hour.
  4. Specificity4 min readSpecific beats clever, every time.The abstract insight scrolls past. The oddly specific detail stops the thumb. Why the number you almost cut is the post.
  5. Distribution4 min readThe first comment is yours to lose.You hit post and walked away. The hour you skipped is the hour that decides reach. What the accounts that travel do in the first sixty minutes.
  6. Cadence5 min readConsistency is a format problem, not a willpower one.You don't fall off because you lack discipline. You fall off because every post starts from a blank page. Fix the page, not the person.

Read from real posts

These notes are not theory. They came off real posts.

Nine posts Michael Maximoff actually published, Chief Growth Officer, Co-Founder @ Belkins. Every figure is real and every date links to the live LinkedIn post, so you can check any row. This is the raw material the notes above were read from.

  • 220words, median post
  • 6 of 9/9open in first person
  • 2–15blank-line beats
  • 0em dashes, counted
ShippedFirst line, his wordsWordsBeatsER*Views
Apr 4, 2026My co-founder is obsessed with AI. Can anyone relate?220120.5%12,187
Apr 13, 2026I was driving, and after the first 20 minutes, I said to myself, “What an incredibly insightful conversation, was I really part of it?”18890.5%8,374
May 4, 2026We grew the BLKNS community to 3,600 members. Then we stopped investing. Why?338150.9%7,454
Mar 9, 2026I'm watching Apollo, Outreach, SalesLoft, Nooks, Instantly and many others all converge into the same tool.249131.5%3,861
Apr 2, 2026Just wrapped up an interview with marketing GOAT Udi Ledergor, ex-CMO and Chief Evengelist at Gong. Excited to share this one soon 💣4724.3%2,914
Feb 24, 2026You shouldn't have to pay to learn.14494.4%2,523
May 6, 2026Upd from Belkins: we’re steadily trying to figure out Lead Gen 3.0.296154.2%2,115
May 3, 2026Abbas Somji, thank you for inviting me onto your podcast. It’s refreshing to be on the other side of the interview for a change.6333.6%2,095
Apr 9, 2026I bet 90% of tech leaders spend at least 1–2 hours a day working on AI. People reading this fall into this group.239103.9%2,012
9 posts · totals1.6%43,535

Words and beats read on wider screens*Engagement rate, derived: (reactions + comments + shares) / views. These 9 are the publicly quoted subset of Michael’s 31 tracked posts.

The rules we read by

The same rules are wired into the engine.

These notes are not content marketing. They are how the writing engine actually thinks. Below are its real system-prompt rules, quoted word for word, each one behind a note above.

  • “Compress quickly. Get to the point with very little runway.”backs · The cold open beats the hook
  • “Posts should end 20-35% earlier than feels comfortable.”backs · Ended four lines too late
  • “Use contrast well. The market does something. The client does the opposite.”backs · Specific beats clever
  • “Never use: The em dash symbol (AI tell).”backs · 0 em dashes, counted below
  • “Fabricated specifics are the single biggest failure mode of AI-written posts.”backs · Never invent a number

The Writing Assistant applies every one of them, in each profile’s own voice. The advice and the product are the same thing.

Reading about it is step one. Writing it is the engine.

Put a profile in and the assistant drafts in their voice, with every rule above already running. The trial is the whole platform, not a demo.

Start your 7-day trialSee the writing assistant

7-day full-access trial. $0.95 to start. Cancel anytime.