How to vet a news source you have never heard of

Small is not the same as untrustworthy

A link arrives from a domain you do not recognize. The instinct to dismiss it is wrong as often as it is right. Some of the best reporting in the world comes out of two-person trade publications, regional papers, and single-beat newsletters. Some of the worst comes from sites with slick design, a plausible name, and no humans behind it.

Brand recognition is a weak proxy for reliability. What you want instead is a short list of structural signals — things a real newsroom has to have and a content farm rarely bothers to fake well. None is decisive alone. Together they resolve most cases in under ten minutes.

Look for named humans who accept consequences

Open the About or Staff page and look for people: names, roles, and ideally short bios with prior work you could verify. A real publication puts its people forward because bylines are how journalists build careers. An operation that does not want to be traced uses house bylines, invented staff names, or nothing at all.

Then check whether an individual byline appears on the article you were sent, and whether that byline has an archive of other work on the same site. A single name attached to forty articles across wildly unrelated beats in one week is a strong signal of automated production.

The deeper test is accountability. Is there a way to reach a specific editor? Journalism is a business of being wrong in public occasionally; publications that expect to be around next year make it easy to tell them so.

Find the corrections policy — and then find a correction

Almost every serious outlet publishes a corrections or ethics policy. Almost no content farm does, because corrections presuppose a standard you intend to be held to. So the presence of the policy is a signal.

The stronger signal is evidence the policy is used. Search the site for the word “correction” or “updated” and see whether real corrections exist, dated, describing what was wrong. A site with a beautiful ethics page and zero corrections across three years of daily publishing is either flawless or not actually checking.

Check who owns it and how it makes money

Ownership disclosure tells you where the pressure comes from. A trade publication owned by an industry association will cover that industry differently than an independent one, and that is not automatically disqualifying — it is information you need in order to read it correctly.

Revenue model matters similarly. Subscriptions align a publication with readers who will cancel if the work is bad. Advertising aligns it with volume. Sponsored content and affiliate revenue create direct incentives on specific stories, which is why disclosure norms exist around them. If you cannot tell how a site earns money at all, that is worth noticing.

Read the archive, not just the article

Click into the site's front page and read the last twenty headlines as a set. You are looking for coherence: does this look like a publication with a beat, or a pile of unrelated high-traffic topics? Content farms drift toward whatever is searchable — health scares, celebrity finance, product roundups — because the topic is chosen by keyword volume rather than editorial interest.

Also check the age and rhythm of the archive. A site claiming years of authority whose earliest post is four months old is misrepresenting itself. A site publishing sixty pieces a day with a staff page listing three people is not writing them.

Watch how it treats other people's work

This one is quietly the most diagnostic. Real publications link out. They name the outlet that broke the story, quote briefly, and send you there — partly from professional norm, partly because their own value is their reporting, not their traffic monopoly.

Operations that exist to capture attention do the opposite. They summarize another outlet's scoop at length without linking, attribute to “reports” instead of a name, or reproduce the substance of an article so completely that clicking through becomes pointless. If a site consistently gives you everything and credits no one, its business is arbitrage on someone else's work.

Cross-check one specific, checkable detail

Pick the most concrete fact in the article — a number, a date, a job title, a quoted sentence — and verify just that one thing elsewhere. Not the thesis, which is arguable, but a detail that is either right or wrong.

This is fast and surprisingly decisive. Fabricated and machine-generated articles tend to be directionally plausible and specifically wrong: the official has the wrong title, the figure is from a different year, the quote appears nowhere else. Publications with real editing get the checkable details right, because those are what editors check.

The ten-minute audit

In order, stopping as soon as you have your answer:

• Is there a named author with an archive of related work on this site?
• Does an About or Staff page list real people with verifiable prior work?
• Is there a corrections policy, and does the site actually publish corrections?
• Can I tell who owns the site and how it earns money?
• Do the last twenty headlines describe a coherent beat, or a keyword grab bag?
• When it discusses other outlets' reporting, does it link and name them?
• Does one concrete, checkable detail in the article hold up elsewhere?

Two or three failures is enough to set the piece aside. And when a source passes, you have learned something durable — you can trust it faster next time, which is the whole point of doing the audit once.

Keep a personal source list

The reason this audit is worth the effort is that it compounds. Every source you vet either joins a short list you can read quickly and confidently, or a shorter list you skip on sight. Within a few months of doing this deliberately, most links you encounter resolve instantly, and you spend your scrutiny only on genuinely new domains.

Write the list down, by beat rather than by brand — the outlet you trust on municipal government is rarely the one you trust on semiconductors. That specificity is what makes the list useful rather than a loyalty badge.