Most content tools give you a number and ask you to trust it. ViceRadar gives you a number and shows you where it came from, because a score you cannot interrogate is a score you cannot act on.
The six components
Every topic is measured on six things. Five combine into the score; the sixth adjusts it.
Momentum — how fast views are accumulating across the topic right now, and whether that rate is rising or falling. Measured from repeated observations of the same videos, not from a single snapshot.
Outlier strength — how far the topic's videos are running above what those channels normally do. This is the component that separates "a big channel posted" from "this subject is genuinely pulling". A 400k-view video from a channel that averages 2 million is underperforming.
Demand signal — views per video, not views in total. Ten videos splitting 50,000 views means creators are interested. Three videos pulling 40,000 each means viewers are. Only the second is an opportunity.
Early-mover advantage — how much of the topic's useful window is left, decayed against a per-category half-life. A leaked-detail topic is worth hours; a guide topic is worth weeks.
Competition — how many channels are already on it and how large they are. This never zeroes a score. A genuinely explosive topic should be downranked when contested, not erased.
Confidence — how much evidence sits behind all of the above: how many videos, how many observations, over how long, and how fresh the data is. It multiplies the final result, so a thin topic cannot score highly on a lucky reading.
Why the parts multiply instead of adding up
These are gates, not contributions. A weighted average would let a topic with zero momentum score respectably because competition happened to be low — which is actively misleading advice, since a dead topic in an empty niche is still a dead topic.
So the components multiply, then take a weighted geometric mean. A weak dimension drags the whole score down rather than being averaged away by strong ones.
Two components can pull a score to zero and two cannot, and the difference is deliberate. Zero momentum or zero demand means nobody is watching, which genuinely is not an opportunity at any price. But an outlier multiple at or below 1.0 only means videos are performing as that channel usually does — ordinary, not absent — and a closed early-mover window means late rather than worthless. Treating those as annihilating once produced a topic with maxed momentum, 45 videos across 45 channels and 90% confidence scoring zero, displayed next to a HOT badge.
What it refuses to do
When a topic cannot be measured, it reads UNSCORED rather than scoring zero.
That distinction is the whole product. Zero means "we measured this and there is nothing here". UNSCORED means "we cannot measure this yet, and here is exactly which input is missing". Most topics read UNSCORED early in a topic's life, usually because the channels involved have no performance baseline yet — and a baseline needs observations of a channel's earlier uploads at matching video ages, which accumulate over days.
We could fill that gap with an estimate. We don't, because a number built on nothing is worse than no number: it looks identical to a real one.
Reading a score in practice
A score is a ranking signal, not a verdict. Two things matter alongside it:
- Status tells you what phase the topic is in. RISING and BREAKING are worth acting on; SATURATING means supply is climbing while demand is not.
- Confidence tells you how much to trust the score at all. A 60 at 30% confidence is a guess; a 45 at 90% confidence is a measurement.
Open any topic and every component is listed with its raw value, its normalised value and its weight. If a score looks wrong, you can find out why in about ten seconds — and if it is wrong, we would rather you could prove it.