The Social Trend Score is an internal signal we use to surface topics that are rapidly gaining attention. It is a discovery and ranking tool, not a scientific or objective measurement of a story’s real-world importance.
What it uses
Where data is available, the score is built from a weighted combination of signals: how quickly coverage of a topic is accelerating (velocity), how many independent sources are covering it, discussion volume on public social platforms, search interest, how recently the topic emerged, and how geographically widespread the coverage is. Conceptual starting weights are roughly: 40% velocity, 20% independent-source coverage, 15% social discussion, 10% search interest, 10% freshness, and 5% geographic spread — normalized against the data actually available for a given topic.
What it doesn’t use
The score does not incorporate editorial judgment about a topic’s significance, political viewpoint, or any paid promotion. A high score means a topic is receiving fast-growing, broad attention right now — not that it is more important than a lower-scoring story.
Data gaps
Not every signal is available for every topic (for example, search-interest or social-discussion data may be limited for very new or narrow topics). Where a data source is unavailable, we do not substitute a fabricated number — the methodology normalizes across whatever verified signals exist, and we label a metric as unavailable rather than guessing.
Anti-manipulation
We apply checks to reduce false positives caused by spam, duplicate syndication of the same wire story, a single publisher repeatedly updating one article, or coordinated/bot-driven activity. Editors can manually override, merge, split, or suppress a trend that the automated signals get wrong.
Update frequency
Trend calculations refresh on a rolling basis, roughly every 30-60 minutes, as new source data is ingested.
Limitations
This is a first-party, transparent-but-imperfect signal. Treat it as “this is rising in coverage right now,” not as a certified fact.