What Coverage Breadth Measures

This is a reference page for one of the six signals described on the methodology page. It goes into the detail that page summarizes in a paragraph.

The question this signal answers

Coverage breadth is worth 20 of the 100 points in a story score. It stands in for a judgement no ranking system can make on its own: how many independent newsrooms looked at an event and decided it was worth a reporter’s time. One outlet publishing first tells you something. Ten outlets publishing separately tells you more — and it tells you something freshness cannot.

What counts as one source

• The unit is a publisher, not an article. Sources are counted as distinct publisher domains, lower-cased with a leading www. removed. A publication that files three updates on one event counts once.
• Press releases are not counted. An issuer’s own announcement is not an independent report of it, so press-release members of a story are excluded from the count before the score is calculated.
• A story made only of press releases counts as one. However many sites carry the release, the scoring treats it as a single source — the issuer. The story page’s source-mix line reports it honestly as zero independent editorial domains.

Why syndication does not widen coverage

A corporate announcement can go out on a newswire, be picked up by a financial site and reprinted by a portal within minutes. Three domains, one document, no additional reporting. Counting that as three sources would let anyone with a distribution budget buy a high ranking, so CurrentWire treats a distribution chain as one source.

A release that real newsrooms then go and report is a different case: those newsrooms are counted normally, and the story becomes eligible for the ranked sections it was otherwise kept out of. The full rules for labeling and placing press releases are on the editorial standards page.

How the count becomes a score

Breadth is scored on a logarithmic curve that reaches its maximum at twelve independent sources. These are the exact values the ranking engine produces:

The curve is steep at the start and flat at the end on purpose. Going from one source to two is the single most informative step there is: it is the difference between an unconfirmed report and a corroborated one. Going from eleven sources to twelve tells you almost nothing you did not already know, and beyond twelve the score does not move at all — a story cannot climb by accumulating pile-on coverage.

Breadth against freshness: a worked example

Freshness is the largest single signal at 35 points, so it is fair to ask whether breadth ever actually wins. Consider two United States stories, one broadly covered but ten hours older, the other brand new and reported by nobody else. Geography is identical for both, and holding prominence and velocity equal, only three signals differ:

Story A is ten hours older and still ranks ahead. That is the intended behaviour: a list ordered purely by clock time is a feed, not a ranking. It is also why nothing stays on top forever — freshness decays continuously, and no story older than 72 hours is ranked at all, however broadly it was covered.

Where the count is visible

• Every story page names each publication covering the event, with the time of its report and a link to the original.
• The automated coverage analysis on a story page breaks the mix down: independent editorial domains, press releases, and opinion or analysis pieces.
• Corroborated details only list a phrase when it appears verbatim in reports from at least two distinct independent domains. Same-domain repetition never corroborates anything.
• The BREAKING label requires breadth, not just speed: four or more independent sources, or three including a Tier A publication, on top of a very high score and coverage under 90 minutes old. At most one story carries it at a time.

What coverage breadth is not

• Not a truth test. Twenty outlets can carry the same wire copy and all be wrong together. Breadth measures how many newsrooms judged an event significant, not whether their reporting is correct.
• Not popularity. It counts publication decisions, never traffic, clicks, shares or comments. CurrentWire does not rank on engagement and collects no engagement data to rank on.
• Not political balance. Which newsrooms covered an event is an observation. It is not a claim that the coverage was balanced, and it is never adjusted for the politics of the outlets involved.

Related references

• How CurrentWire ranks the news — all six signals and their weights.
• Publisher authority tiers — the other half of the source question: not how many, but who.
• How duplicate stories are merged — how reports of one event are grouped before they are counted.
• Sources — the publications currently represented in CurrentWire coverage.

Think a story’s source list is wrong? The corrections process explains how to report it.