October 26, 2026
The alert arrives at 6:41, while you are still deciding whether to get out of bed. A building is on fire, a minister has resigned, a company has folded — the details are thin, but the tone is urgent. You read it, you form a quick impression, and by the time the coffee is ready the first version of the story is already settled somewhere behind your eyes.
The correction, if there is one, will come at two in the afternoon and reach a small fraction of the people who saw the alert. That gap — between how fast a claim travels and how slowly it gets fixed — is the whole argument in miniature. Nobody involved is careless. The system simply rewards the first version more than the accurate one.
The Price of a Fast First Draft
Speed costs in small coins, which is why the bill is easy to ignore. A name spelled wrong. A figure published without its denominator. A single source paraphrased into a firm fact by the second paragraph. None of these are dramatic failures, and each one is fixable. But they accumulate into something larger: a reader who has learned, from experience, to trust the headline less than the headline deserves.
There is a second cost, less visible. When the deadline for the first post is fifteen minutes away, the reporting shrinks to fit it. You call one person instead of three. You skip the one who would have said no. You take the press release at its word because checking it would push you past the hour. The story that runs is not wrong, exactly. It is thin, and thin stories bend easily when the next development arrives.
And the treadmill keeps moving. A newsroom that measures itself by publish rate will get more publishing. A newsroom that measures itself by what it got right, and what it had to fix, will get something else. The measurement is not a detail; it is the whole culture wearing a spreadsheet.
Slow Is a Sequence, Not a Delay
"Slower news" does not mean sitting on a story until it is cold or letting a competitor own the day. It means changing the order of operations: verify, then publish what is verified, then label clearly what is not. A live event can still be live. A court verdict can be reported within minutes of being read out — the verdict is a fact with a timestamp. What should not happen is characterising the judge's reasoning before anyone has read the ruling.
The practical distinction worth holding onto is between we don't know yet and we're not telling you. Readers forgive the first and resent the second. A line as plain as "Officials have not named the injured, and we are not repeating the name circulating on social media until it is confirmed" costs eleven seconds and buys hours of credibility.
A Ten-Minute Floor Before You Publish
Most avoidable errors happen in a window that is shorter than people think. Not an hour — about ten minutes, spent deliberately. Here is a sequence that fits almost any story:
- Write one sentence stating what you know and how you know it. "The council voted 7–4 to reject the plan; I read the minutes." If you cannot fill in the second half, you are not ready.
- Call the other side. Even a "no comment" received by telephone changes the shape of the piece, because it proves someone tried.
- Check that your two sources are actually two. Wire copy, a reposted social post and a rival's article are one source wearing three coats. Trace everything back to the origin before you count it.
- Find the denominator. "Infections doubled" says nothing without the starting number, the population and the period. "Record high" says nothing without a start date for the record.
- Read the document. Not the summary of the document, and not the release describing the document. The interesting part is usually on page nine.
- Decide now what you will say if you cannot confirm. Drafting that sentence before the pressure arrives is far easier than drafting it at speed with an editor waiting.
Ten minutes will not make you last. It will make you the outlet people quote when they want to be sure.
Context Is Not Padding
There is a lazy version of context — three paragraphs of history stapled to the top because someone said the piece needed background. Good context does work. It changes how a reader interprets the facts in front of them, and it usually takes fewer words than the padding does.
The comparison that reframes the number
A rise in reported incidents is not the same as a rise in incidents; reporting behaviour moves first and often moves most. A figure that looks alarming beside last month's may look ordinary beside the same month in each of the past five years. None of this is a reason to soften a story. It is a reason to give the reader the frame they need to judge it themselves.
The stranger test
Hand your first three paragraphs to a colleague from a different desk — someone who knows nothing about the story. Ask them what happened and what it means. If they come back with an impression your reporting does not support, the problem is in the opening, not in their reading. This takes four minutes and catches more than any style guide.
Corrections Belong in the Open
The instinct to bury a correction comes from treating it as reputation damage. It is the opposite. A visible, well-written correction is the strongest evidence a reader has that everything else on the site has been checked. A silent edit is a second error, committed against the people who already read the first one.
- Correct at the top when the error touches the main claim. A note at the foot of the page is a note for people who already finished reading.
- Say what was wrong, plainly. "An earlier version said the plant closed in March. It closed in May." No euphemisms, no "updated to reflect new information" when what happened was a mistake.
- Keep both timestamps — when the piece was published, when it was corrected. The gap is information, not embarrassment.
- Do not quietly fix a quote. If the wording was wrong, say so. A quote that changes without explanation teaches readers to doubt every other quote.
- Go back to the places people saw it. Correct the newsletter, the social post, the push alert thread. That is where most of the audience actually met the error.
- Thank the reader who caught it, publicly if they are willing. They did free quality control, and they will do it again.
If You Are Reading, Not Writing
Some of this is on the receiving end too, and it takes less effort than people expect.
- Check the timestamp before you share. A screenshot without a date is a rumour with good posture.
- Give a breaking story one news cycle before you settle your opinion on it. Early details — casualty counts, names, motives, charges — are the ones most likely to change.
- Notice whether an outlet publishes corrections at all. If you never see one, that is not a sign of accuracy. It is a sign of housekeeping.
- Click the follow-up. Reward the update as readily as you rewarded the alert.
This matters most with stories that touch health, money or the law. A first report on a drug, a scam or a court ruling is a starting point, not a basis for a decision. Speak to a qualified professional before acting on anything that carries real consequences.
Start With One Rule
None of this requires a rebuild. Pick a single rule and hold it for a month: nothing publishes with a number that has no source attached. Or: every contested claim gets one phone call before it goes out. Or: someone reads the first three paragraphs and says back what they think happened.
At the end of the month, count the corrections. Not to punish anyone, but because that number tells you what your process is actually producing — and a newsroom that knows its own error rate is already ahead of one that assumes it has none.
The headlines will keep breaking. The choice is whether they break on your terms or at your expense.
Photo: Sóc Năng Động / Pexels
