Free Growth Assessment
Marketing

Thought Leadership Examples From Tech Founders (and What Each One Gets Right)

Four founders, four pages read at the source. What each one gets right, in plain terms, and what an MSP owner can borrow without a research team.

By Holly Mack Last updated: November 22, 2026 12 min read

Four thought leadership examples from tech founders, read at the source, show four moves: a thesis, your own numbers, a named idea and one reader. Here’s what each gets right and what an MSP owner can borrow.

A man in a dark shirt writing by hand on a blank pad at a desk, with an open network rack in the alcove behind him
One page, one position, your name on it.

Search for thought leadership examples and you’ll notice a pattern.

One list I read for this piece names Microsoft’s Work Trend Index, Bessemer’s State of the Cloud and HubSpot’s State of Marketing among its eight, and each is a research program from a well-known company or firm. Impressive. Also not something a 30-person managed service provider can copy.

So owners read a list like that and decide thought leadership isn’t for them. On paper, fair. You don’t have a research team, and you shouldn’t build one.

But then you look at what a founder can do alone. One page. One position. Your name on it, on your own site, with a LinkedIn post pointing back.

For the service side, see our LinkedIn thought leadership services, and for the channel that carries the short version of your page to readers, our hub on LinkedIn marketing for MSPs.

This isn’t a reach problem. It’s a position problem. A founder with no stance has nothing to reach people with.

So I went looking for founders instead, and set one rule. I’d only use pieces I could open and read in full at the original source, because I won’t quote anyone from memory or from somebody else’s roundup.

Four made it, and I read them all on November 22, 2026. None is a LinkedIn post. They’re essays on a founder’s own site or a firm’s, which means the moves are about the thinking and not the channel, and they carry over to a LinkedIn post, as I’ll show.

Key Takeaways

The Short Version

Andreessen put a thesis in his headline. Hansson put his own bill on the page, then came back two years later with the new one. Graham named the idea. McKenzie wrote to one reader.

You don’t need all four moves in your first piece. Pick one, publish it on a page you own, and point your posts at it. For the concept, read what thought leadership marketing is and how it builds pipeline, and for the founder’s side of the channel, there’s our MSP founder LinkedIn playbook.

What Counts as a Thought Leadership Example

A thought leadership example is a piece of original thinking, published under a named person’s name, that takes a position the reader can test or use. It isn’t a product announcement, a recap of other people’s research, or a post any company’s logo could sit on.

I used four tests to pick these: a named founder, a piece I could read in full at the source, a position stated in plain words, and evidence the founder owns or clearly credits. That’s my screen, not an industry standard, and a different reader might pick four other people.

The Four at a Glance

FounderPiece and dateLength and homeThe move
Marc Andreessen“Why Software Is Eating the World,” August 20, 2011About 2,200 words, in The Wall Street Journal first, then on a16z.comA thesis in the headline
Paul Graham“Do Things that Don’t Scale,” July 2013About 3,800 words, on his own siteA named idea
David Heinemeier Hansson“Why we’re leaving the cloud,” October 19, 2022, and a follow-up on October 17, 2024About 1,000 and 900 words, on his blog on HEY WorldHis own numbers, then the update
Patrick McKenzie“Don’t Call Yourself A Programmer, And Other Career Advice,” October 28, 2011About 5,500 words, on his own siteOne reader, in their terms

Lengths are my own counts of each page’s body text on November 22, 2026, rounded.

Different years. Different lengths. One shared habit. Each one holds a view somebody could argue with.

Marc Andreessen, 2011: The Thesis Is the Headline

Start with the simplest version of the move. Andreessen’s Why Software Is Eating the World runs about 2,200 words, and the article’s first line repeats the headline: “Software is eating the world.”

He wrote it in 2011, when people were asking, in his words, “Isn’t this just a dangerous new bubble?” He doesn’t dodge it. He puts it in the opening paragraph, then states his answer a few paragraphs later: “My own theory is that we are in the middle of a dramatic and broad technological and economic shift in which software companies are poised to take over large swathes of the economy.”

Then there’s the stake. Andreessen tells you who he is to the topic before you’ve had time to wonder. He writes that he is “co-founder and general partner of venture capital firm Andreessen-Horowitz” and that he is also “personally an investor in LinkedIn.”

You can discount his view for that if you like. He’s already handed you the reason.

And he dates the claim. “Over the next 10 years, I expect many more industries to be disrupted by software.” A prediction with a number on it can be wrong, which is exactly why a reader believes the author means it.

I haven’t graded the prediction. This piece won’t. The lesson here is the form, not whether he was right.

Your version is smaller. Think of a headline in the same shape: “Co-managed IT only works when the client has someone to co-manage with.” That one is invented to show the form. It isn’t a C4 position or a finding.

But a reader can disagree with it, and that’s the point. A headline nobody could argue with isn’t a stance. It’s a topic.

Paul Graham, 2013: Name the Idea So People Can Carry It

Graham’s Do Things that Don’t Scale is a plain page on his own site, dated July 2013 and, by my count, about 3,800 words. He opens by describing a belief founders hold, that a startup either takes off or it doesn’t, and corrects it in one line: “Actually startups take off because the founders make them take off.”

Now the part that lasts. Stripe’s founders recruited early users by hand, and Y Combinator has a name for how. “At YC we use the term “Collison installation” for the technique they invented.” When anyone agreed to try Stripe, he writes, they’d say “Right then, give me your laptop” and set them up on the spot, which is the whole technique in one sentence.

A name is a handle. Without one, an idea is too slippery to pass along. “Collison installation” gives a founder something to say to another founder, and that’s how an idea travels.

One catch. Graham has a platform no MSP owner has, and the name carries partly because of who said it. Borrow the structure anyway.

You already run a method on every new client. You don’t call it anything. If your first week always starts with a walk through the server closet and a list of what’s unlabeled, give that walk a name a client could repeat to a friend. A named method is easy to refer, and an unnamed one is just something you do.

David Heinemeier Hansson, 2022 and 2024: Show Your Own Numbers, Then Come Back

Hansson is co-owner and CTO of 37signals, the company behind Basecamp and HEY. On October 19, 2022, he published Why we’re leaving the cloud, about 1,000 words on his own blog on HEY World. The stance is in the first paragraph. The proof sits a few paragraphs down: “We’re paying over half a million dollars per year for database (RDS) and search (ES) services from Amazon.”

Then he gives a reason a reader can picture. Staying in the cloud for the chance you’ll need it, he writes, is “like paying a quarter of your house’s value for earthquake insurance when you don’t live anywhere near a fault line.”

What I’d point at is the boundary. He tells you where the cloud wins: “The cloud excels at two ends of the spectrum, where only one end was ever relevant for us.” A claim with edges is easier to believe than a claim with none.

Then he came back. In Our cloud-exit savings will now top ten million over five years, dated October 17, 2024, he published the bill again: “For 2024, we’ve brought the cloud bill down from the original $3.2 million/year run rate to $1.3 million.” He also put the new Dell hardware at about $700,000, and he added the caveat a skeptic would raise: “Now, as with all things cloud vs on-prem, it’s never fully apples-to-apples.”

I’m not here to referee cloud versus on-premises. These are one company’s figures, stated by the person who runs its technology, and I can’t audit them. What I can say is that the central claim in each post comes with a number he put on the page (and yes, the bill is the interesting part).

Copy the follow-up first. A prediction is a post. The result two years later is a track record.

You own numbers nobody else has. Your PSA ticket history. How long a cutover really takes. How many first quotes turned into signed agreements.

Nobody has asked you to publish them, and that’s the opening. If you don’t have the number, leave it out. A claim without a number is an opinion. A claim with an invented number is worse.

Patrick McKenzie, 2011: Write for One Reader, in Their Terms

McKenzie writes as patio11, and his own site says he has run several software companies. On October 28, 2011, he published Don’t Call Yourself A Programmer, And Other Career Advice. By my count it’s about 5,500 words, the longest of the four, and it says who it’s for in the opening lines: “This post aspires to be README.txt for your career as a young engineer.”

Then it names its limits. “I wouldn’t trust this as the definitive guide, but hopefully it will provide value over what your college Career Center isn’t telling you.” Notice the edge.

Here’s the move. He doesn’t explain engineering to engineers. He explains how engineers look from the other side of the table, and one of his section headings says so: “Engineers are hired to create business value, not to program things.” An engineer reading that learns how the person signing the check sees them.

That’s translation. And it’s the same job an MSP owner has. Your buyer might not be the person who loves the tech, and could just as easily be an owner, a CFO or an office manager. The page they’d forward is the one that shows them how IT looks from their desk, in their words, down to what the invoice buys and what a bad quarter costs them.

A caution. It’s about careers, not marketing, so the move transfers and the advice doesn’t. And it runs long, which works only because every section is for the same reader. A page written for everyone gets forwarded by no one.

What These Thought Leadership Examples Share

Line the four up and the pattern is plain. One idea per page. A line worth repeating. Something to check.

Here’s each move as a question you can put to your own draft.

Four Moves, Four Questions

The moveSeen inA question for your draft
A thesis in the headlineAndreessen, 2011Could a sharp competitor disagree with the headline?
Your own numbersHansson, 2022 and 2024Does the page hold a figure only you could know?
A named ideaGraham, 2013Could a client repeat its name to someone else?
One readerMcKenzie, 2011Does the first paragraph say who this is for?

If every answer is no, you probably have a topic, not a position. A topic is fine for a post. Just don’t mistake it for a stance. More posts won’t fix a missing position.

How an MSP Owner Borrows These Moves

Here’s a working order. It’s mine, not a rule from any of these founders. The MSP in it is invented, a 30-person shop whose owner wants to write about co-managed IT, so nothing below is a C4 result or a client.

A woman and a man in business clothing laying three blank printed pages in a row on a long meeting-room table
Lay the moves out before you write a word.
  1. Write the thesis as one sentence a competitor could disagree with. Skip the version that says IT matters.
  2. Say your stake in the first paragraph. If you sell co-managed IT, say so, the way Andreessen did.
  3. Find one number only you have, from your PSA ticket history or your onboarding records, and write down where it came from. No number yet? Leave it out.
  4. Name the method you already run on every new client.
  5. Pick one reader and say who it is in the first two lines. An owner at a 60-person law firm, not “businesses.”
  6. Say what would change your mind, the way Hansson named the two cases where the cloud wins. Then put a date on your calendar a year out to publish the update.
  7. Publish on your own site first. Then write a short LinkedIn post that states the thesis and points to the page.

That last step is where the channel comes in. Whether the short version should be a post or an article is its own question, and we worked through it in LinkedIn articles vs posts, which covers what LinkedIn documents about each format and what three 2026 studies found.

One more thing, and it’s the point of the whole order. If it’s not repeatable, it’s not scalable. One strong page is a start. A founder who comes back to the idea every few months is building something readers can find.

What No Thought Leadership Example Can Tell You

Four essays prove nothing about leads. I picked them. So the sample is mine. Three of the four date from 2011 to 2013, and none of the four is about running an MSP.

Each of the four is well known now, and I don’t know how big any audience was on the day the piece went up, so I can’t tell you the structure caused any of it. I can tell you what the pages do, and that you can copy it for free.

Nothing here shows what founder content does for pipeline, either. This piece states no C4 result, because none is confirmed for it. If you want the measurement side, our post on thought leadership marketing covers how to measure it without fooling yourself.

So treat the four as a drawing board. They show the moves. Your own pages have to show whether the moves work for your buyers.

The better question isn’t which founder to copy. It’s which one position you’d defend to a skeptical client, and whether you have a page to put it on. If you’d like a second set of eyes on that, book a growth strategy call.

Common Questions Before Your First Piece

Which of these should a founder read first?

Hansson’s 2022 post. It’s short, about 1,000 words, and its central claim comes with a number.

Read Andreessen second for the thesis, Graham third for the naming, and save McKenzie for a day you want a long one.

Does thought leadership have to live on LinkedIn?

Not for these four. Each lives on a page its author or firm controls, and Andreessen’s ran in The Wall Street Journal first.

LinkedIn is a good place to point people at the page. Keeping the page itself on your own site means it doesn’t depend on a feed.

How long does a founder’s piece have to be?

900 to 5,500 words, going by these four. Hansson’s two posts run about 1,000 and 900. McKenzie’s runs about 5,500.

Length followed the job. Write until the reader has what they came for, then stop.

Can an owner with a small following pull this off?

The structure doesn’t need a following. The reach does, and I can’t tell you how big these authors’ audiences were when they published.

So start with one reader you can name, like a client, a referral partner or a prospect you’ve already met. A page that helps one of them is a page they can forward.

Do I write it myself or hand it to a ghostwriter?

The thesis, the numbers and the stake have to come from you. Those are the parts a reader checks.

A ghostwriter can shape the page and hold the schedule, and we have a page on LinkedIn ghostwriting if you want that help. Nobody else can supply your ticket history.

How would I know a piece like this is working?

Not by likes. Count the replies that quote your line back to you and the calls where the prospect says they read it.

Both are small numbers you can log by hand. I can’t promise either one shows up, and neither can anyone else.

READY TO ACT?

Find out which move your first piece should use

Talk to us about the position you’d defend, the number only you have, and where the page lives, so LinkedIn works as one part of a system.

Book a Growth Strategy Call