A startup came to us with what its team called a strong case: eighteen articles about its funding round, published across eighteen outlets within forty-eight hours of the announcement. To the founders, that volume looked like exactly the kind of evidence that had cleared a Wikipedia biography draft for one of them, personally, a year earlier.
It wasn't the same kind of evidence, and reading the eighteen articles side by side showed why. Each opened with the same round size, the same investor list, the same line from the CEO about what the funding would be used for. None reported a detail the press release hadn't already supplied. When we lined up the language, whole sentences matched the release nearly word for word. Eighteen bylines, one source, retold eighteen times.
For organizations, coverage must include content that is independent of the subject — analysis, investigation, or verification the source produced itself, not material that merely reprints or restates a press release, a company website, or other information the subject supplied.
This is where the standard for a company diverges from the standard for a person. A magazine profile of a founder, built from an interview the founder gave, usually still counts: the outlet chose to run it, shaped it, decided what mattered. The same interview, repackaged as sourcing for the company itself, usually does not, because the test for an organization isn't only whether the source is independent of the subject; it's whether the content is. A press release quoted by eighteen outlets stays one source under that test, no matter how many bylines carry it.
So we read organizational coverage differently now, one question at a time: did anyone here place a call, ask a question the release didn't answer, check a number against a filing, offer a judgment the company didn't write for them. An article that fails all four is a repetition, however well it reads, however many of them exist.
It is not a comfortable message to deliver to a founder holding a folder of press clippings, and understating it does not make it kinder. What we tell them instead is that a stack of mentions proves reach, not coverage, and only one of those two counts for the case we are actually building.
