How to summarise an article in your own words
A good summary is not a shorter version of an article. It is a statement of what the article argued, written for someone who will not read it.
That difference explains why most summaries are bad. People start at the first paragraph and compress as they go, which produces something with the original shape, the original order, and no sense of which parts mattered.
The method below works in the other direction: find the argument first, then write.
Step 1: Read it once without writing anything
You cannot judge what is central until you have seen the whole thing. Summarising as you read guarantees that the first section gets three sentences and the conclusion gets half of one.
Read it through. If it is long, note only the section headings — you are building a map, not taking notes yet.
Step 2: Find the thesis and write it in one sentence
Every article has one controlling claim. Find it and write it out in your own words before you do anything else, because everything in the summary either supports this sentence or gets cut.
It is usually in the final paragraph of the introduction or the first of the conclusion. If you cannot find it stated anywhere, it is implied — write the sentence that all the sections are collectively arguing for. The tests in how to find the main idea work at article scale too.
A test: if someone read only your thesis sentence, would they know what position the author took? “The article discusses remote work” fails — that is a topic. “Remote work raised productivity but weakened junior staff development” passes.
Step 3: Pull out the supporting claims, not the details
Go back through and extract the two to five claims the thesis rests on. Claims, not evidence — “productivity rose” is a claim, “a 2024 study of 61,000 workers found a 4% rise” is the evidence for it.
Most articles have fewer real claims than they appear to. A ten-page piece often has three, each supported by several examples that all make the same point. Spotting that the examples are interchangeable is most of the compression.
This is the step highlighting is genuinely useful for: mark only the sentences that state claims and the structure becomes visible.
Step 4: Write from your notes with the article closed
This single rule prevents most summarising problems. Close the article, look at your thesis sentence and your list of claims, and write.
With the text in front of you, you will unconsciously reach for the author’s phrasing, and the result is a patchwork of their sentences with a few words swapped — which reads awkwardly and is, in academic terms, plagiarism even with a citation. Writing from your own notes forces genuine restatement.
It also exposes gaps. If you cannot explain a claim without looking, you did not understand it, and that is worth knowing now rather than after you have submitted something.
Step 5: Check it against the original for distortion
Now reopen the article and check your summary against it. You are looking for three specific failures, in this order:
Dropped hedges
The most common and most serious. The author wrote “may contribute to” and your summary says “causes”. Compression naturally strips qualifiers, and stripping them changes the claim into one the author did not make.
Smuggled opinion
Words like “merely”, “admits”, “fails to”, “surprisingly”. If the author did not express that attitude, you have editorialised. A summary reports the argument; your assessment of it is a separate paragraph.
Inverted emphasis
Check the proportions. If the author spent two-thirds of the article on one mechanism and your summary gives it one clause, you have reported the wrong article.
How long should it be?
For most purposes, somewhere between five and ten per cent of the original. A 3,000-word article becomes 150–300 words: one sentence of thesis, one per supporting claim, one of conclusion.
If you cannot get it that short, you are still including evidence rather than claims. Go back to your list and check each line — anything containing a specific number, name or example is probably supporting detail that can go.
When not to summarise at all
Summarising destroys the original wording, which is fine for a briefing and wrong for revision, quotation and anything legal or technical where the exact phrasing carries the meaning. Highlighting vs summarising covers where the line sits.
The practical compromise is to do both: highlight the original so you keep a checkable, fully-worded version, then write the short summary from those marks. You end up with something you can hand to someone else and something you can go back and verify against.