I Read the 60-Word Summary, Then the Full Story: What ShortNews Leaves Out

A summary makes you a deal. It says: trust me, I have read the long version, and this is the part that matters. You pay for that trust with the facts you never see. If the summary keeps the wrong fact and drops the right one, you walk away believing something technically true and practically wrong — and you never know, because the whole point of a summary is that you did not read the rest.

ShortNews is built entirely on that deal. Its homepage pitch is "News in 60 words. Zero noise." Three earlier posts in this series tested the format from the outside: the first counted the words and found the summaries run 10 to 24 words, far under the 60-word ceiling; the second counted the stories and found 27 homepage slots carrying 15 headlines; the third opened the story pages and found a real article underneath each summary. This time I tested the deal itself. I read four summaries first, wrote down what I believed had happened, and then read the full articles to score what each summary kept, cut, or softened — judged against the demo's own editorial policy, which promises every summary is distilled to "what happened, why it matters, and what happens next." One thing to be clear about first: nothing in the four summaries I read was false. The interesting failure in a summary is rarely a lie. It is a selection, and selection is where a short format earns or loses your trust.

How the test worked

For each story I did the same three steps: read only the headline and the labelled "In 60 words" summary and note what I would have repeated to someone else; read the full article underneath and list every fact it contains that the summary does not; then score the summary on whether it kept the story's load-bearing fact, whether it kept what happens next, and whether it softened anything the article says plainly. I picked the four stories the demo pushes hardest — the ones in the breaking-news ticker on every page: the earnings story, the blood test, the exoplanet image, and the championship final. Summary word counts are the demo's own labels; article word counts are mine, labelled approximate; read counts are the demo's own seeded numbers, never real traffic.

The earnings story: the summary that kept the numbers and lost the story

The earnings story is the best summary of the four, and it still loses the most interesting part of its own article. Headline: "Tech Giant Posts Record Quarterly Earnings, Shares Jump 12%". Summary, labelled 16 words: "Strong cloud growth drove the beat. Revenue hit $45B, up 22% year-over-year, as AI services demand surges." Read only that and you would believe the quarter was strong, revenue was $45 billion, up 22%, driven by cloud and AI. All true. The article, about 45 words, adds three facts that change the story: the company beat analyst expectations of $42 billion; cloud services grew 31%, fuelled by enterprise AI workloads; and the shares jumped 12% in after-hours trading, adding $180 billion in market value.

Each cut matters. The $42 billion figure is the difference between a good quarter and a beat — without it, "record earnings" is a mood, not a measurement. The 31% is the specific engine inside the vague "strong cloud growth" the summary kept. And the $180 billion is the consequence of the headline's own "Shares Jump 12%", the one number in the headline the summary never explains. The summary kept the headline numbers and dropped the context that gives them meaning — and restoring it would have cost about fifteen words it had to spare, since it used 16 of its 60.

The blood test: the summary that dropped its own headline's number

If the earnings summary is the best of the four, the blood test is the one that should worry you. Headline: "Breakthrough Blood Test Detects 50 Cancers Early". Summary, labelled 10 words: "A single blood draw could transform how we catch cancer." Read only that and you would believe there is a new blood test, it takes one draw, and it might change screening. The article, about 67 words, adds: the test detects more than 50 types of cancer at early stages; it has shown promising results in large clinical trials; it works by identifying tiny fragments of tumor DNA in the bloodstream, often before symptoms appear; regulators are reviewing the data; and if approved, it could join routine annual checkups within a few years.

The dropped fact that matters most is the "50". It is the single most specific, load-bearing number in the entire story — in the headline, the reason the story is news, the one thing a reader would repeat to someone else — and the summary drops it entirely. What remains could describe any cancer test announced in any year. The second cut is the status: the article says "promising results in large clinical trials" and regulators "reviewing the data", the difference between a proven thing and a promising thing. The summary's "could transform" softens all of that into one hopeful verb. Ten words, and the two most important qualifications — how many cancers, and how far from real — are both gone.

The exoplanet: wonder kept, future dropped

The exoplanet story shows the same cut from the other side: a summary that keeps the wonder and drops the future. Headline: "Astronomers Capture Clearest-Ever Image of Distant Exoplanet". Summary, labelled 11 words: "A direct image of an alien world 400 light-years away stuns scientists." The article, about 60 words, adds: a gas giant orbiting a star 400 light-years away; advanced coronagraph technology that blocks starlight made the image possible; atmospheric analysis suggests cloud bands similar to Jupiter's; and the team hopes the technique will eventually image rocky, Earth-like planets — and search their atmospheres for signs of life.

The summary keeps the distance and the awe, and drops three things a reader would actually want: what the planet is (a gas giant, not an Earth twin — the first question anyone asks about an "alien world"), how the image was made (the coronagraph, the breakthrough the headline is really about), and the sentence the article ends on, which is the best sentence in the whole story — the search for life on Earth-like worlds. That is the "what happens next" the demo's own editorial policy promises every summary will carry. It is in the article. It is not in the summary. The fix is cheap: the summary used 11 of its 60 words, and "a gas giant, photographed by blocking its star's light — a technique that may one day image Earth-like worlds and search for life" is 22 words.

The championship: one fact wearing a headline

The championship story is the thinnest summary of the four. Headline: "Championship Final Sets All-Time Viewership Record". Summary, labelled 10 words: "Over 120 million viewers tuned in for a historic championship showdown." The article, about 79 words, adds: more than 120 million viewers worldwide, the most-watched sporting event of the year; the game went down to the final seconds before a dramatic finish; the winning moment generated over 2 million social posts within an hour; and analysts credit the record audience to two storied franchises and a season-long rivalry.

The summary keeps the 120 million and drops everything that makes 120 million mean something: the record itself, the finish, the 2 million posts, and the rivalry that drew the audience. The article is a story about why people watched. The summary is a number with a game attached. Sports is the forgiving case — nobody makes a decision on a championship summary. But the shape of the cut is the same shape as the blood test: one number kept, the context that makes the number matter left behind. A format that does this to a game will do it to a clinical trial, and the reader has no way to tell which kind of story they are reading from the summary alone.

The pattern in the cuts

Four stories, four summaries, one pattern. Every summary was accurate, and every summary dropped the fact that made its story worth reading. The earnings summary dropped the expectations beat and the $180 billion. The blood test dropped the "50" and the unapproved status. The exoplanet dropped the gas giant, the coronagraph, and the search for life. The championship dropped the record, the finish, and the rivalry. In each case the summary kept one number — $45 billion, 120 million, 400 light-years — and dropped the second number, the mechanism, and the consequence.

The second half of the pattern is the unused budget. The summaries run 10, 10, 11 and 16 words against a 60-word ceiling — the demo's own labels, confirmed by reading them. The blood test summary could have added "detecting 50 cancers early, in trials, unapproved" in six words. The exoplanet summary could have added the gas giant and the search for life in eleven. The earnings summary could have added the $42 billion expectations and the $180 billion market value in ten. None of them needed to be longer than 30 words to keep everything they dropped. My honest opinion, and it cuts against the format's own marketing: the problem with these summaries is not that 60 words is too few. It is that the summaries do not use the words they already have. They read like they were written to a 10-word habit, not a 60-word limit — and the habit, not the limit, is what costs the reader the facts. To their credit, I found no falsehood in any of the four, no number contradicted by its own article. The summaries are honest. They are just thin — and thin is a different problem from wrong, with a different fix. Wrong needs a correction. Thin needs an editor who uses the budget.

The sourcing gap: the policy promises links the pages do not have

There is one more check a reader should run on any short-news site, and it is the one the demo fails outright. ShortNews's own editorial policy, last updated October 18, 2026, says: "Sources first: every summary links to the original reporting it is based on." That is a strong promise, and exactly the promise a summary format needs, because the summary asks you to trust a cut you cannot see. The link is the escape hatch — the way back to the full reporting when the summary drops something you needed.

I checked all four story pages for outbound links, and there are none. Each page ends with "Originally reported by ShortNews Demo" and nothing else — no source name, no source link, no "read the original" anywhere. The policy promises a link on every summary. The demo delivers a credit line with nowhere to go. That removes the reader's only defence against the cuts above: if the blood test summary drops the "50", the reader's options are the demo's own full article — a second summary of about 67 words — or the source link, which does not exist. To be fair, this is a demo, and its stories are seeded content dated October 13–19, 2026, so there may be no original reporting to link to. That is a reasonable explanation, and also the reason the policy should not make the promise in the demo at all. A policy is a contract with the reader, and a demo that prints the contract while shipping pages that break it teaches the operator the wrong lesson: that the policy is decoration. If you run a site on this script, the cheapest trust you can buy is a source link on every story, because it is the one feature that lets your reader check your work.

What happens when a summary is wrong

The one place the demo's paperwork is genuinely good is the corrections policy, also last updated October 18, 2026. It promises factual errors are corrected "as soon as they are confirmed — usually within hours", that significant corrections are noted on the article itself, and that there are "no silent edits". Errors are reported through the contact page with a link to the article and, if you have one, a reliable source, and "every report is read by a human editor". The story pages back that up with furniture: each carries a "Report this story" control with categories for Misinformation, Spam, Harassment, Hate speech, Violence and Other, plus A- and A+ text size controls. I did not submit a report — there is no live error to report on these four pages, and a test report to a demo would be noise — but the control exists, it is specific, and it sits on every page rather than buried in a footer. That is the right design.

Here is the honest limit of that safety net: a corrections policy fixes facts, and the problem with these summaries is not facts. Nothing I found was wrong enough to report. The blood test summary did not misstate the number of cancers — it omitted it. The exoplanet summary did not misdescribe the planet — it never said what kind it was. No correction fixes an omission, because there is nothing to correct, only something to add. The corrections policy is a good net for the wrong kind of failure. The failure these summaries actually have — thinness — needs an editor, not a corrections form.

The verdict: accurate, thin, and fixable for the price of a sentence

For a reader, my verdict is two-sided. The summaries are accurate — I read four against their full articles and found no falsehood, no contradicted number, no claim the article walks back. If you want the shape of the day in a minute, this format does what it says: the homepage pitch, "News in 60 words. Zero noise.", is honest about the length, and the summaries really do run 10 to 16 words. But a summary that drops the story's best fact is a summary you cannot quite trust, even when it is true, because the thing you needed was the thing it cut. I would read these summaries to find out what happened. I would not repeat them to anyone else without opening the article first — and that, for a format whose entire job is to save you the opening, is the verdict that matters.

For an operator, the fixes are cheap and all editorial, not technical. Use the budget: write summaries to 30 or 40 words, not 10, and spend the extra words on the second number, the mechanism, and what happens next. Keep the load-bearing number — the "50" in the blood test, the "most-watched of the year" in the championship. Add the source link the editorial policy already promises, on every story. And keep the corrections policy and the report control exactly as they are — the best paperwork on the demo. None of those fixes needs a new feature. They are one editorial decision each, and they would turn a format that is accurate but thin into one a reader could trust without opening every article — which is, after all, the only reason to write a summary in the first place.

Frequently asked questions

Are ShortNews summaries accurate?

On the four stories I tested — earnings, blood test, exoplanet, championship — yes. No falsehood, no number contradicted by its own full article. The problem is thinness, not accuracy: each summary drops the fact that makes its story worth reading.

What do the summaries leave out?

The pattern across all four: the summary keeps one number and drops the second number, the mechanism, and the consequence. The earnings summary drops the $42 billion expectations beat and the $180 billion market value. The blood test drops the "50 cancers". The exoplanet drops the gas giant, the coronagraph, and the search for life. The championship drops the record, the finish, and the 2 million social posts.

How long are the summaries, really?

The demo labels them 10, 10, 11 and 16 words — far under the 60-word ceiling. The full articles underneath run about 45 to 79 words by my count. The summaries use roughly a sixth of the budget the format allows, which is why the dropped facts are so cheap to restore.

Do the story pages link to the original reporting?

No. The editorial policy promises "every summary links to the original reporting it is based on", but none of the four story pages carries any outbound link. Each ends with "Originally reported by ShortNews Demo" and nothing else.

What happens if a summary gets something wrong?

The corrections policy promises fixes "usually within hours", with significant corrections noted on the article and no silent edits. Each story page carries a "Report this story" control with categories including Misinformation. I did not submit a test report, so the flow itself is untested. Note the limit: corrections fix wrong facts, not omissions, and thinness is an omission.

Is ShortNews a real product I can buy?

ShortNews is a short-news website script with a live demo at shortnews-demo.codester.shop. No price is stated here because the shop's front page carried no ShortNews listing when checked today — check the shop itself before assuming a price.

How is this different from the earlier ShortNews posts in this series?

The first post counted the words on the cards; the second counted the stories on the homepage; the third opened the story pages; the fourth tested the five-morning routine. This post reads each summary against its own full article and scores what was kept, cut, or softened — a fidelity test, not a count, a tour, or a routine.

Who a 60-word summary is for

If your day runs on headlines and you want the shape of the news in a minute, this format is built for you — the summaries are honest, the reading is fast, and the full article is one tap away when a story earns it. If you make decisions on what you read — a health story, a market story, anything with a number in it — read the article, not the card, because the card keeps one number and the article keeps the rest. And if you run a site like this, remember that your summary is a promise: you read the long version so your reader does not have to. Spend the words. Keep the load-bearing fact. Link the source. A summary that does those three things is worth trusting. One that does none of them is just short.

Four stories, read twice each. The summaries were true. The articles were better. That gap — between true and complete — is the whole test, and it is about a sentence wide.

Earlier in this series: Count the Words Before You Trust the News: A 60-Word Test of ShortNews — the word-count audit of the same demo. And Half the Homepage Is the Same Story Twice: A Repeat Audit of ShortNews — the duplication audit. And I Tapped the Card: What a 60-Word News Story Actually Gives You on Its Own Page — the story-page audit. And Five Mornings With ShortNews: Does 'News in 5 Minutes' Survive Day Two? — the routine test. The blog index carries the full series, and the tool this site is built around is free to use.