Mention Intensity
- Intensity is the depth dimension: how much of the answer is about you, not whether you appear.
- Count span-aware and alias-deduplicated, or intensity inflates faster than any other mention metric.
- Read it with Brand Rank: deep and early means the brand anchors the answer; shallow and late means filler.
- Report median and threshold-share, not means — intensity distributions are skewed.
Consensus definition
Research on generative engines formalized early that whether a brand appears is too coarse a signal: the founding GEO benchmark proposed position-adjusted word count and subjective-impression measures precisely because answers allocate attention unevenly across the sources and brands they name.1 Mention Intensity is the practitioner form of that insight at brand level: within one answer, count the span-aware brand occurrences (or the share of answer text devoted to the brand) instead of recording a binary hit. Where Mention Rate asks "in how many answers do we appear?", intensity asks "how much of the answer is about us?"
rhinegold operator refinement
Rhinegold reads intensity as a co-signal, never a target. Two disciplines keep it honest. First, the count must be span-aware and alias-deduplicated — a brand with three name variants is not three times as present, and substring hits inflate intensity faster than they inflate Mention Rate. Second, intensity must be read against answer style: list-shaped answers distribute shallow mentions across many brands, narrative answers concentrate depth on few. Comparing raw intensity across providers with different answer styles measures the providers, not the brand.
Operational use
Use intensity to separate two situations a presence metric cannot distinguish: broad shallow visibility (mentioned everywhere, never substantially) and narrow deep visibility (few answers, but built around the brand). The pairing with Brand Rank is diagnostic — deep mentions in early list positions indicate the brand anchors the answer; shallow late mentions indicate filler.
Measurement boundary
Intensity says nothing about direction — a brand can be intensively criticised (that is Negated Mention territory) — and nothing about sourcing (that is Citation Rate). Absolute intensity values are tool- and provider-specific; only trends on a stable prompt set and comparisons within the same provider are meaningful.
Distinct from
Against Mention Rate, which is binary per answer and measures breadth across a prompt set. Against Share of Voice, which aggregates mentions relative to the competitor set — intensity is an absolute, within-answer measure. Against Brand Recommendation Share, which weights presence by exclusivity rather than by depth.
Operational note
Common mistakes
- Summing alias hits without deduplication — multi-alias brands show inflated intensity that disappears under span-aware counting.
- Comparing intensity across providers with different answer verbosity — longer answers mechanically allow more occurrences.
- Optimizing content to raise intensity directly — repetition is not depth, and answer-stuffing signals are exactly what providers tune against.
Where consensus is missing
There is no standard unit: occurrences per answer, brand-text share, and position-weighted counts all circulate, and tools rarely disclose which they implement. Alias handling and parent/subsidiary attribution are unstandardized, so cross-tool intensity numbers are not comparable.
Sources & deeper reading
- rhinegold span-aware, alias-deduplicated intensity counting across providers — methodology shared with engaged clients and partners.
- rhinegold Insights, Episode 04 — How to Measure What You Can't See (GEO metrics for B2B)