AI Visibility Forecasting
Citation concentration, and why it matters more than volume
Whether the set of sources engines draw on is widening or narrowing is the single most consequential forecastable condition.
What concentration means
When systems assemble answers, they draw on sources. Concentration describes how narrow that set is: whether a small number of sources supply a large share of what gets cited, or whether citations are spread widely.
It is more consequential than the total volume of citations, because it decides whether being a good source is enough to be used.
Why the direction matters commercially
Concentrating. A shrinking set of sources supplies more of the answers. Being outside that set becomes progressively harder to overcome regardless of quality, and the value of being inside it rises sharply. In a concentrating environment, earning a place in an authoritative source is worth more than publishing anything.
Spreading. More sources contribute. Specific, well-supported material from a smaller party has a better chance of being used, and publishing something genuinely new pays better than it does otherwise.
Why it is forecastable
Concentration is an aggregate property that emerges from many interactions, and aggregates have inertia even when the underlying product decisions do not. That makes it one of the few things in this field where a directional statement is defensible.
Timing remains far less tractable than direction, and any statement about when is labelled low confidence for that reason.
What it is not
Not a claim about which specific sources will be favoured, which is a product decision made privately. Not a prediction about a business own citation count. And not a reason to chase whichever sources currently dominate, since that is a lagging response to a condition that may already be turning.
How a business should use it
As weighting on where effort goes. In a concentrating environment, effort moves toward earning presence in authoritative independent sources. In a spreading one, effort moves toward publishing material specific enough to be worth citing.
The mistake is treating either as permanent. The condition changes, and a strategy chosen for one and never revisited becomes exactly the kind of decayed alignment that is expensive to notice.