- One methodology, every channel
- Source credibility and reach weighted, never counted flat
- Syndication collapsed to the story, not the pickups
- The same definition of negative, everywhere
One engine, or the comparison is fiction.
Most monitoring platforms are an acquisition trail wearing a single logo. A press feed bought here, a social feed licensed there, a video score from a third vendor. Each scores sentiment by its own method, on its own scale, against its own idea of neutral. Put those three numbers side by side on a slide and you have produced a chart, not a comparison. Brillaince runs press, social, YouTube and AI answers through one analysis engine, so when social sentiment falls four points while press holds steady, that gap is real and it is safe to act on.
One engine also means one source model underneath it. Not every mention is one unit of reputation, so none of them are counted that way. Each source carries a weight the engine applies before anything is scored: the credibility and reach of the outlet, whether a story is original reporting or a syndicated pickup of it, how far a social account's audience actually extends, how much authority a YouTube channel holds in its category, and how often a given page is cited by the AI engines themselves. A trade title read by two thousand buyers can legitimately outrank a general daily read by two million, and forty reprints of one wire story stay one story rather than becoming forty. Those weights sit inside the calculation, not in a filter applied to the results afterwards, which is why the same number means the same thing whichever channel produced it.
What it costIt meant we could not simply resell someone else's sentiment scores and ship a year earlier.
