Every small B2B company arrives at the same conclusion within about four months of deciding to do content marketing: this is more work than we thought, and we cannot keep it up.
The maths is not subtle. A serious article takes six to ten hours. LinkedIn wants posts several times a week. The newsletter goes out monthly. Someone said you should be doing video. That is a full-time job, and the person doing it also has a real job, which is why the blog has three posts and the last one is from March.
Repurposing is the standard answer, and the standard version of it is bad. “Turn your blog post into ten LinkedIn posts” produces ten posts that read like fragments of a blog post, because they are. Nobody engages with them, the author concludes content does not work, and the blog goes quiet again.
The version that works is not chopping. It is a system where one piece of real thinking gets expressed in the native form of each channel it appears on. This article covers how to build that, what it costs weekly, and where it stops working.
The Source-First Model — One Real Thing, Then Everything Else
The unit is not a blog post. It is an argument.
Repurposing fails when the source is thin. If the article is a competent summary of things generally known, there is nothing to repurpose — you can reformat it, but every derived piece will be as forgettable as the original. Reformatting weak content produces more weak content, faster. That is the whole failure mode.
So the model starts one level up. Once a month — not once a week, once a month — you produce one substantial piece built on something only you can say:
- A number from your own operations
- A decision you made and what happened
- A pattern across your customers that outsiders cannot see
- A strongly held position on a contested question in your niche
- A process you actually run, documented in full
That piece is the source. It should take you a real day. Everything else that month derives from it, and derives from it easily, because there is something in it worth deriving.
The test for whether you have a source or a filler. Could a competitor have written the same article by reading the same three public pages? If yes, you do not have a source. You have a summary, and the market has plenty.
One source per month is enough. A small B2B company publishing twelve genuinely original pieces a year, each expressed across three or four channels, is out-publishing almost everyone in its category on the only dimension that matters. Twelve is a target you can hit. Fifty-two is a target that will be abandoned in April.
What Each Channel Actually Needs
Repurposing is translation, not extraction.
The mistake is treating a channel as a container of a different size. It is not — it is a different language with its own grammar, and a piece written in one and pasted into another reads exactly like what it is.
The article (source). Long, structured, built to be found by search and read by someone with a problem. Owns the full argument, the caveats, the evidence. This is the thing that still earns traffic in three years and the only asset in the list you actually own.
LinkedIn. Wants one idea, stated with a position, in a form that makes sense with no context. A LinkedIn post is not a summary of the article — it is one of the article’s claims, argued on its own terms, with the specific detail that makes it credible. The article might contain six such claims. That is six posts, each of which stands alone, spread across three weeks.
What it is not: “I wrote a new blog post about inventory metrics. Link in comments.” That is an advertisement for content, not content, and the platform’s distribution treats it accordingly. The outbound and posting mechanics on LinkedIn get into the platform-specific side.
The newsletter. Wants the thinking behind the article, not the article. The person on your list already gets your public writing if they want it. What they cannot get elsewhere is why you looked into this, what surprised you, what you got wrong first. The newsletter is where you are a person rather than a publication, and that is what makes it get opened. Link to the article at the end for whoever wants the full version.
Short video. Wants one claim and a face saying it. Two minutes, no production. The claim from the article that would make someone in your industry stop scrolling because they disagree with it.
A slide or a diagram. The article’s core structure as one image. The framework, the sequence, the comparison table. This is the highest-reuse asset in the set — it goes in the LinkedIn post, the newsletter, a sales deck, and a proposal. Make it once, use it fifteen times.
A customer email. The most under-used derivation and often the most valuable commercially. Take the argument, cut it to four sentences, send it to the eleven customers for whom it is directly relevant, with a line saying why you thought of them. That is not marketing. It is a reason to be in touch that is not “checking in.”
The Weekly Routine
A month, in the order it actually happens:
Week 1 — Produce the source. One day, blocked, not fragmented. Write the article. This is the only week that requires real creative energy, and protecting it is the whole discipline. If it gets eaten by client work every month, the system does not exist.
Week 2 — Derive the set. Two hours. From the finished article, pull:
- Four to six standalone claims, each written as its own LinkedIn post
- One newsletter angle — the behind-the-thinking version
- One diagram or framework image
- One short video script, if you do video
- The customer email, with a named list of who gets it
Write them all in one sitting. This matters: deriving in a batch while the argument is fresh takes a fraction of the time it takes to come back on a Tuesday in three weeks and reconstruct what you meant.
Weeks 2–4 — Publish on a schedule. Everything is already written. Publishing is a ten-minute task, not a creative one. Schedule it and stop thinking about it.
End of month — Look at what worked, briefly. Which claim got engagement? Which one produced a reply from someone who could buy? That is not vanity metrics — it is the input to next month’s source. The claim that generated argument is usually the next article.
Total: roughly one day plus three hours a month, producing an article, five or six LinkedIn posts, a newsletter, a video, a diagram, and a set of customer touches. That is a sustainable load for a founder who has other work, which is the only relevant test.
Where AI Helps, and Where It Ruins It
The honest split.
AI is genuinely useful in this system for:
- Deriving structure. “Here is my article. List the six distinct claims it makes, each as a standalone statement.” That is a mechanical task and models do it well.
- Format translation. Turning your written claim into a video script, or an outline into a diagram brief.
- First-pass editing. Cutting a 300-word draft to 120.
- The boring variants. Subject lines, alt text, the third rewrite of a sentence that is nearly right.
AI is corrosive when used for:
- Producing the source. The source’s entire value is that it contains something only you know. A model does not know your numbers, your customers, or what you got wrong in 2024. Ask it to write the article and you get a competent summary of the public internet — precisely the thing we established has nothing worth repurposing.
- Writing the LinkedIn posts unsupervised. Everyone in your industry can now recognise this instantly, and recognising it costs you more credibility than posting nothing.
The reliable rule: AI on the derivation, never on the source. The thinking is the asset. The reformatting is the chore. Automate the chore. Our note on AI writing assistants in B2B operations covers where that line sits across other tasks.
Where Repurposing Stops Working
Three failure modes, all common.
The channel with no audience. Repurposing into a channel where you have nobody is not repurposing, it is shouting. If your buyers are not on a platform, a post there costs you time and returns nothing regardless of how well it is adapted. Small B2B companies routinely maintain four channels, of which one produces every lead. Cut the other three and do the one properly.
Repetition without new value. The same argument, sixth time, to the same people. There is a real limit to how many times one idea can appear before your audience stops seeing you. The limit is higher than most people fear — you are far less repetitive to your readers than to yourself — but it is not infinite, and the tell is engagement flattening on derived pieces while the source still performs.
The system running with nothing in it. The month you have no source and derive anyway. You will produce five posts from an article that was itself thin, and the whole month is noise. When there is no source, publish nothing that month. Silence costs less than filler, and it is much easier to recover from.
The measurement that keeps you honest: track replies and enquiries, not impressions. A LinkedIn post with 400 views that produced one message from a director at a target account outperformed a post with 12,000 views that produced applause. In small B2B, the audience that matters is small enough to count, and vanity metrics are the fastest route to concluding that content does not work while doing it wrong.
The reason repurposing has a bad name is that it is usually taught as a productivity trick — same content, more places, less effort. Applied that way it is a machine for spreading thin material across more surfaces, and it does not work, and it makes the people doing it feel like frauds.
The version that works inverts the effort. You do more work on one thing per month, not less, and the repurposing is what makes that affordable — because a day spent on something genuinely worth saying is only justifiable if it reaches everyone it should reach, in the form each of them will actually read. Get that right and the system runs for years on one focused day a month. Get it wrong and you will produce a great deal of content and nobody, including you, will remember any of it.
Sources: Google Search Central — Creating helpful content · European Commission — Data protection