Long-form to short-form is usually described as shortening. That description is why most attempts produce a summary nobody remembers: shortening keeps the article's shape and removes its substance.
The alternative is to treat the long piece as a container and open it. A 2,000-word article is not one argument that needs reducing; it is four to six units that need separating.
Where to cut
Cut at seams, not at word counts. There are four kinds of seam:
Between claims. Each claim the article argues for is a candidate post. If the article has one claim and four examples, it has one post.
Between mechanism and conclusion. "X fails at scale" and "X fails because Y accumulates faster than Z" are two different posts, and the second is the better one.
Between assumption and reality. Anywhere the article says "most people think..., but...", there is a post that starts with the assumption and ends with the evidence.
Around artifacts. A checklist, a config, a script, a decision rule. These become posts with almost no rewriting, because they are already short by nature.
Three conversions that work
Article → thread. Take one claim, add its mechanism, end with an implication. Six to eight lines. The thread's job is attention, not completeness.
Article → carousel. Take one comparison or one cost breakdown and give each item a slide. Carousels work when the sequence is the argument, so only convert sections that have a natural order.
Article → single post. Take the sharpest claim plus one piece of evidence and stop. This is the highest-yield conversion and the one people skip, because it feels too small to count as distribution.
What does not survive the conversion
- The introduction. It exists to warm up a reader who already committed. Short-form has no room for warming up.
- The caveats. Keep one boundary line; delete the rest. Caveats protect the article from criticism and cost the post its readers.
- The recap. A conclusion that summarises tells the short-form reader the piece is over, which is the one thing you never want them to feel early.
- The transitions. "As we saw above" is meaningless when the reader arrived at this unit directly.
A worked sequence
Take a 2,000-word article about churn interviews.
- Pull the claim: "Almost nobody churns because of price."
- Pull the mechanism: "They churn because a second person on the account never got invited, so the product has no internal advocate."
- Pull the contrast: "Usage looks healthy right up until the cancellation."
- Pull the artifact: "If a customer cannot name a second user by day 14, treat the account as at risk."
Four posts, four different first lines, one source. None of them is a summary, and none requires writing from scratch.
Common mistakes
Cutting to a word count. Word counts produce fragments. Units produce posts.
Keeping the article's order. Order matters inside an article and is irrelevant across posts. Publish the sharpest unit first.
Publishing all units in one week. Four posts on Monday read as a dump. Four across two weeks read as a body of thought.
Rewriting instead of extracting. If the new post needs new evidence, it is not a conversion of the article — it is a new piece of work, and it should be scheduled as one.
Three rules for where to cut
- Cut at seams, not at word counts. Word counts produce fragments; seams produce posts. The seams are between claims, between mechanism and conclusion, between assumption and reality, and around artifacts — the checklist, the config, the decision rule.
- Link once, at the end. The short version carries the argument, the long version carries the depth. A post that opens with a link has spent its first line asking the reader to leave.
- A thin source produces one post — publish one. Padding is how a feed starts to feel like filler, and readers detect it faster than any metric will tell you.
- Transcripts are the exception worth making. Video and podcast material contains mechanisms that were never written down, and the spoken phrasing often survives as a stronger first line than anything you would type.