Learn
Read the board properly
Every measure here is a constructed one. Knowing what each measure excludes matters as much as knowing what it counts.
“Velocity”
How fast attention is accumulating right now, relative to this entity's own recent normal.
It does not tell you how big something is. A small thing can have high velocity.
“Spread”
How many distinct communities are carrying the conversation.
This is the fake-trend filter. Loud in one vertical is not a trend.
“Sentiment”
Valence and intensity, held apart rather than averaged into one mood score.
Mock is not the same as negative, and it is the more expensive one.
“Commerce pull”
Whether attention is converting into demand, measured through public rank movement.
Rank is ordinal. It shows direction, not units sold.
“Durability”
How quickly attention decays, fitted as a half-life in days.
A long half-life with rising velocity is structural. A short one is a meme.
“Half-life”
The number of days for attention on an entity to fall by half.
Compare it against your production cycle before committing budget.
“Mock share”
The proportion of the sampled corpus that is mocking rather than adopting, rejecting or debating.
Above 40% it becomes the leading brand-risk indicator on the board.
“Confidence”
How much source coverage stands behind the number, not how certain a model feels.
Low confidence means sources disagree, not that the signal is weak.
“White space”
100 minus spread. How much room is left before saturation.
High velocity with low spread is the creator's arbitrage window.
“Lead time to demand”
Median days between an attention inflection and a commerce-pull inflection.
This is the operational warning window, not a forecast.