A negative article about your company feels like an emergency. It usually isn't. Most bad coverage has a half-life measured in days, and the single most common way companies turn a bad week into a bad quarter is by responding loudly to something almost nobody read.
So the first job isn't drafting a response. It's triage.
The triage decision
Run the article through four questions, in order.
1. Is it factually wrong, or is it just unflattering?
This is the distinction everything else depends on. "Our revenue was $4M, not $400K" is an error. "The reporter framed our pivot as a retreat" is an interpretation. Publications correct errors. They very rarely change framing, and asking them to makes you look like you don't understand the difference.
Be honest here. Founders routinely classify accurate-but-harsh reporting as "inaccurate." An editor will spot that in one read, and you'll have spent the credibility you needed for the next time something actually is wrong.
2. Who actually saw it?
Check the reality of the reach. Is this a top-tier outlet, or a low-traffic aggregator? Is it ranking for your brand name in search? Is it being shared by people your customers listen to? A critical post with 40 views on a site nobody reads does not need a response. It needs to be monitored.
The relevant test is not "does this exist" but "will a prospect or a partner encounter this." If it ranks on page one for your company name, it matters regardless of the outlet's size.
3. Is there a material consequence?
Is it affecting sales calls, a fundraise, a hiring pipeline, a regulator's interest? If your sales team is getting asked about it, that's a real signal. If nobody has mentioned it except you, that's a different signal.
4. Do you have something new to say?
A response with no new information is just volume. If your only contribution is "we disagree," you're adding a second day to a one-day story.
The four possible responses
Based on triage, you have four moves. Most situations resolve to the last two.
Request a correction when there's a specific, demonstrable factual error. This is the highest-value move available and the most underused.
Respond publicly when the story is significant, widely read, and you have substantive new information or a genuine correction of the record, and you can say it without relitigating every point. One clear statement, published on your own site, linked from your channels. Not a thread. Not a point-by-point rebuttal.
Respond privately when the relationship matters more than the record. An email to the reporter that says "here's context we should have given you earlier, no correction needed, happy to be a resource next time" costs nothing and frequently changes how the next article gets written.
Do nothing when the piece is accurate, small, or already fading. Doing nothing is an active decision, not a failure. Pair it with monitoring so you know if it grows.
How to request a correction
Correction requests succeed or fail on precision. Editors receive a steady stream of vague complaints from companies that didn't like an article. The ones that get acted on look completely different.
What works:
- Email the reporter first, copying the section editor only if the reporter doesn't respond within a business day. Going over someone's head as a first move poisons the relationship permanently.
- Quote the exact sentence from the article, with the paragraph location.
- State the correct fact in one line.
- Attach evidence. A filing, a signed document, a dated page, a regulator's database entry. Not "our records show," but an artifact they can verify independently.
- Say exactly what you're asking for. A correction, an update note, a removed sentence, an added line of context. Editors need a specific, actionable request.
- Keep it to one issue if possible. A list of eleven complaints reads as an attempt to discredit the article. Lead with the strongest one.
A workable version:
Hi [Name], one factual issue in today's piece on [topic]. Paragraph 6 states that [company] "[exact quote]." The correct figure is [X], per [specific source, linked]. Would you be able to correct that line? Happy to answer anything else on background. Thanks, [Name, title, phone]
What to expect: publications correct clear factual errors reasonably often, usually with an appended correction note rather than a silent edit. They will not remove an article, will not change a headline's tone, and will not let you review copy. Asking for any of those marks you as someone to avoid.
The Streisand risk
The named-after-a-lawsuit dynamic is real and it is the main reason to be conservative. Amplifying, demanding removal, or threatening legal action reliably converts a small story into a bigger one, because the response is itself newsworthy in a way the original wasn't. A legal threat against a journalist is a story. A furious founder thread is a story. A quiet correction request is not.
Three specific traps:
- Responding on a platform with more reach than the original. If the piece lives on a site with modest traffic and you post about it to a large audience, you have just published the criticism to your own customers.
- Creating new assets that rank. A long rebuttal page can end up ranking alongside the original for your brand name, so now there are two negative-adjacent results instead of one.
- Threatening before thinking. Legal action has a place, but it should be a considered decision with counsel, not a reflex. The reputational cost usually exceeds the legal benefit unless the falsity is severe and the harm is documented.
The durable fix is what you publish over time
Search results for your company name are not fixed. They're the outcome of what exists about you. If one critical article is the third result for your brand, the answer isn't removing it. The problem is that there are only four results.
The long game is having more legitimate material in circulation: your own press page with real announcements, third-party coverage of things you've actually done, executive commentary in trade outlets, and a steady record of company news. Each legitimate item is a result that ranks on your name. Over a year, a single bad article goes from defining your search page to being one line on it.
This also matters increasingly for how AI systems describe you. Assistants summarize what they find about a company, and a thin footprint means a single critical source carries disproportionate weight. It's worth monitoring how AI tools describe your brand for exactly that reason, though what makes any given source get picked up is still poorly understood, and nobody should sell you certainty about it.
A sober note on proportion
The hardest part of this is emotional, not tactical. The article that feels career-defining to you is, to almost everyone else, a thing they scrolled past. Give yourself a day before deciding. Have someone outside the company read it and tell you what they actually took away. Their answer is usually milder than yours.
Then do the smallest effective thing: fix errors, feed the record, and go back to work.
Building a steadier public record beats fighting one article. [See how it works](/how-it-works).