Brand Mention Persistence

What is brand mention persistence?

Brand mention persistence is how consistently a brand mentioned in one measurement is mentioned again in the next, for the same question and the same AI service. AI answers change from run to run, but which brands make it into the answer does not reshuffle at random each time. Where a brand has started being mentioned for a question, it tends to stay; where it is not mentioned, it tends to stay absent. This concept treats that tendency as a measurement metric.

Why it matters

A study published in September 2026 (Tannenbaum, arXiv:2609.23162, preprint) analyzed 34,960 unbranded prompt observations and compared models that predict whether a brand will be mentioned in the next run. The author is the CEO of a commercial measurement vendor and the data are operational records, so the results should be read as associations, not causal effects.

When RanketAI re-aggregated four rounds of measurement records for 12 Korean B2B SaaS companies (144 cells of brand × AI service × round), 97% (68/70) of cells mentioned in the previous round were mentioned again, versus 11% (4/38) of cells that had not been mentioned. These are judgments grouped at the question-set level on a sample of well-known companies, so the magnitudes do not generalize.

Measurement points

• Count round-to-round transitions — repeat the same question on the same AI service under the same conditions, and compute the "mentioned before → mentioned now" and "not mentioned before → mentioned now" rates separately. A single measurement cannot produce this.
• Read it separately from own-domain citation — even with a strong history, the mention rate can fall when the own domain drops out of the current answer's sources (82–86% → 29–32% in the study). For the four states that combine mention and citation, see Used vs Cited and Ghost Citation.
• The first entry is the hardest — high persistence also means the "not mentioned" state rarely changes. The starting state of a brand with no history is covered in Cold-Start AI Visibility.
• Question-level and grouped judgments differ — judging several questions together dampens round-to-round fluctuation, so persistence looks higher. Match the unit of aggregation before comparing.

Related terms

• Used vs Cited
• Ghost Citation
• Brand Mention Share
• Cold-Start AI Visibility
• Query Fan-out
• Source Gap

Further reading

• AI Mentions the Brands It Mentioned Before — A 35K-Observation Study and Korean SaaS Data — study figures, our re-analysis, and actions for the four mention × citation states
• Monthly Change Was Smaller Than Same-Day Measurement Noise — a monthly record in which mention rate and own-domain citation rate moved independently
• One Follow-Up Question Wipes 62% of AI Brand Picks — how recommendations hold up in multi-turn conversations