The Honest Arithmetic of Doing LinkedIn Properly

Seven posts of this series have described what doing LinkedIn properly involves. None of them exaggerated, and none of them included a single tactic from the pile we threw out in the second post. This one adds it up.

The reason it's worth adding up is that every individual item on the list is reasonable, and the total is not. That gap — between the reasonableness of each part and the unreasonableness of the whole — is the real reason founder marketing fails, and almost nobody ever sees it, because nobody writes the list down.

Every quarter

1. Decide the theme — one arguable position a competitor couldn't publish, rather than a topic list.
2. Define the audience precisely: who signs, what they currently believe, and what they'd have to stop believing.
3. Expand the theme into three to five distinct angles, written down.
4. Extract your voice from a real sample of your own writing, and write down your refusals.
5. Read the entire previous quarter of your own posts in one sitting and count how many genuinely distinct claims survived.

Every week

1. Choose this week's angles against a record of what you've already used.
2. Find real source material for each — something that happened, rather than a subject to cover.
3. Draft.
4. Rewrite for the tells.
5. Check against the corpus for sameness.
6. Post.
7. Read and reply to comments in the hours afterwards.
8. Notice what happened, without over-reading it.

Every post

1. Choose the vantage deliberately.
2. Choose the lead framing before writing rather than during.
3. Ground it in something real rather than in your value proposition.
4. Run the eleven tells.
5. Never invent a specific. Get the fact or cut the sentence.

What that list is made of

Look at what those columns require and they separate cleanly into different kinds of work, done at different rhythms, by people who in a larger company would not be the same person. It was never one job with a lot of steps.

There is a strategist — theme, audience, the belief gap, the angles. Quarterly, unpublishable, and the hardest thinking in the set.

There is a planner — tracking which angles and vantages have been used, keeping the quarter's output diverging, maintaining a rhythm that survives a bad week.

There is someone who reads the world for material — because a post grounded in something real requires a supply of real things, and that supply doesn't accumulate by itself.

There is an editor — voice, tells, refusals, and the willingness to cut a paragraph you like because it isn't yours.

And, last and smallest, there is a writer. The part everyone assumes is the job is roughly a fifth of it.

When it was my job I was the orchestrator and the producer both. I set the messaging framework and the multichannel programme, and then I either bought the content, produced it myself, or handed it to someone on the team.

Read that sentence again for the options in it. Bought, produced, delegated. A founder has none of the three, which is the entire difference between a hard job and an impossible one.

Five kinds of work, on three different clocks, for a company where marketing sits somewhere around the eleventh priority. That's the honest shape of it.

The design flaw that makes it unwinnable

Here is the property that ties the whole list together, and it's worth stating exactly, because it explains something that otherwise looks like a personal failing.

Every item on that list is free to skip on the day and expensive only in aggregate.

Skip the theme: nothing happens Tuesday. Skip the corpus check: nothing happens Tuesday. Skip the vantage decision, the tells pass, the framing choice: nothing happens Tuesday. There is no error message, no missed deadline, no unhappy person in your inbox.

The bill arrives in month three, as forty posts that say the same thing in the same voice with the same tells — and by then it doesn't present as “we skipped the quarterly framework.” It presents as “LinkedIn didn't really work for us.”

You can see the aggregate result of that design in the research. A peer-reviewed study of 1,070 startup firms (Lilien & Mintz, Industrial Marketing Management, 2024) found that 45% of early-stage B2B startups do no systematic marketing at all — not bad marketing, none. That number is usually read as a discipline failure or a budget failure. I'd read it as the predictable output of a task nothing ever forces.

It is close to the worst possible design for a task assigned to a busy person. Everything urgent in a founder's day punishes you immediately for skipping it. This punishes you invisibly, later, in a form you won't correctly diagnose. Against that competition it doesn't lose because you lack discipline. It loses because it is structurally unable to compete for attention.

Two honest readings

You've now read the list. There are two reactions to it, and both are correct.

“Yes — that's my job.” If that's your reading, you're a marketer, and you should go and do it. Nothing on that list is busywork or invented difficulty; it's a fair description of the craft, and it's the work that separates a feed that compounds from one that fills space. You didn't need this series to tell you that. You might find the fourth post useful, because corpus convergence is under-discussed even among people who do this professionally.

“I want to do this, and I'm never going to get to it.” If that's your reading — and if you're running a company, it probably is — that is not a confession and it isn't a discipline problem. It's an accurate assessment of your situation.

I want to be precise here, because the lazy version of this argument says founders don't care about marketing. That's wrong, and it's insulting to people who plainly do. What I've found since becoming one is narrower and more awkward: I want to do marketing and I don't have the time. And despite twenty years of doing it professionally, holding continuity of message while staying genuinely engaging is hard — it was hard when it was my whole job and it's harder now that it's competing with a fundraise.

Every founder who has bought a course, hired an agency, or opened an AI writing tool and quietly abandoned it three weeks later arrived at the same assessment by a slower route.

One thing the honest answer is not: a lighter version. A lighter version of that list is what everybody already does, and it's what produces the quarter of posts that all say the same thing. The list isn't padded. It's what good requires.

I've had both reactions, years apart. For most of my career that list was simply the job, and I'd have read it and thought it slightly under-specified. I run a company now, I look at the same list, and I know I won't get to it — because the company survives on sales closing and money being raised, and marketing loses that argument every single week.

Nothing about the list changed. What changed is that it stopped being anybody's job.

What we did about it

This is the point where it would be normal to offer you a template.

We didn't build one, because the list above is not a knowledge problem. You could read this series twice, agree with all of it, and still fail to do it — and you would be behaving rationally.

Groundworks AI is built around that list rather than around the writing. Writing was never the hard part; a tool that only writes hands the empty box back to you along with all five jobs. The machinery is the point — the quarterly theme and its angles, the vantages, the divergence across a body of work, the voice held steady across fifty posts, the tells, and a cadence that survives the weeks the company is on fire.

We're building that in the open, and some of it is shipped while some of it is still ahead of us. Claiming otherwise at the end of a piece that spends a thousand words on never inventing a specific would be a strange way to finish. What I can say plainly is that the list is real, because it's the one we run against ourselves.

Three things stay with you, and two of them are deliberate omissions from the arithmetic above.

Substance. The number you actually measured, the thing a customer actually said, the position you'd defend in an argument. A system can hold your theme, your angles and your voice; it cannot know what happened to you last week. So when it needs a fact it doesn't have, it asks — and it never invents one. That's deliberate. An invented specific is the last item on the tells list, and it would go out under your name.

Your profile, and the time you spend commenting on other people's posts. Both were on the short list of things that genuinely work, back in the second post, and neither appears in the arithmetic. That's not an oversight and it isn't because they're unimportant — it's because nobody can do either one on your behalf, and a system that tried would be doing you harm. The list above is the production work. It was never the whole of being on LinkedIn.

Judgement. Read the draft, and say yes, or change this, or not this week.

That's the ask. Everything else above is machinery, and machinery was never a founder's job.

See how it works →