Most advice about content repurposing describes reformatting: take a blog post, cut it down, change the line breaks, add a carousel. That is not repurposing. That is distribution with extra steps, and readers can tell — they see the same argument in three shapes and stop clicking any of them.
Repurposing is a different operation. You take one piece of thinking and ask how many different arguments it contains. A well-argued 2,000-word post usually contains four to six. Each one can be published on its own, in a different week, to a different platform, without repeating the others.
What repurposing is not
Three operations get called repurposing and only one of them works.
Compression takes the article and makes it smaller. The result is a summary: accurate, skimmable, and impossible to quote. Compression is why most content calendars look busy and feel empty.
Reformatting keeps the argument and changes the container — the same paragraphs, now in slides. It works for a reader who already wants the argument and fails for everyone else, because the container was never the problem.
Extraction removes units from the article and lets each one become its own piece. This is the only version that produces more reach from the same work, because the output is genuinely new to anyone who did not read the original.
The four units
Almost every unit you can extract falls into one of four types. Naming them makes extraction mechanical instead of mystical.
Claims. The sentences the article argues for. "A 14-month sales cycle kills companies that look healthy" is a claim. "Sales cycles matter" is a topic. Only claims are publishable.
Mechanisms. The "because" behind a claim. These travel furthest, because most writing states conclusions and skips the reason. A mechanism is a post that teaches something the reader can apply elsewhere.
Contrasts. What people assume versus what actually happens. Contrasts are what make a first line stop a scroll, which is the only measurable job a first line has.
Artifacts. A checklist, a config, a script, a decision rule. Artifacts get saved, and saved content is what builds the reason to follow someone.
The workflow
- Inventory what you already published. Articles, transcripts, changelogs, newsletter issues. This is a queue of finished arguments, not a list of ideas.
- Extract in one pass, without writing. Highlight the four unit types. Output is a flat list of sentences. Do not draft anything in this step.
- Assign before you write. Rank units by surprise, match each to a platform, put a date on it. The calendar becomes a decision you already made.
- Produce against the assignment. Write the platform-specific first line, attach the unit and its evidence, stop. Minutes rather than hours.
- Gate and recycle. One question per asset: would you forward this to someone you respect? Anything that fails goes back to extraction, not to copy editing.
Where repurposing fails
Thin sources. If the article contains one claim and four examples, it yields one post. Trying for five produces five restatements of the introduction. The fix is upstream: write articles with more units, or publish fewer posts.
Uniform openings. Five posts that all start with "Most people think X" read as one post published five times. Every unit needs its own first line, written for that unit.
No mechanism. Posts without a mechanism get agreement and nothing else. Agreement is not distribution.
Volume without a queue. Publishing five units in one week empties the backlog and teaches your audience that the account is inconsistent. A queue of units is what makes a cadence survive a bad month.
How to tell it is working
Track units extracted per week, not posts published. Posts are shipments; units are inventory, and inventory is the only thing that makes next week possible.
Then track what readers repeat back to you. When someone quotes a sentence you wrote, that sentence is the seed of the next article — and the sign that the extraction found the sharpest claim in the source rather than the most convenient one.
Three rules that decide whether the work pays off
- Count units, not words. A 2,000-word piece with a claim, a mechanism and an artifact holds four to six posts. A piece with one claim and four examples holds one, because examples are evidence rather than separate arguments — and padding it produces five restatements of the introduction.
- Link once, at the end. Posts carry the argument; the article carries the depth. A post that opens with a link asks the reader to leave before you have earned the click, and a post that never links leaves the depth undiscovered.
- One unit, one first line — in every language. The rule does not change when you move between Chinese and English; only the failure mode does. There the problem is structural rather than lexical: the conclusion sits on line nine because that is where Chinese prose puts it, and the English reader stopped at line two.
- Keep the queue above five units. That number, not the publishing frequency, is what keeps a cadence alive through a launch week — and it is the difference between a system and a habit that ends at the first interruption.