Numbered Titles Dominate AI Citations, With Wide Model Variance
The study provides concrete evidence that the SEO convention of numbering titles transfers to generative AI citation behavior, but the wide variance between models indicates that one-size-fits-all optimization strategies may not work across all AI platforms.
Reporting from 1 source: ASCII.jp.
A study by Optyino.ai analyzed 59,221 AI responses and found that 57.93% of cited web page titles contain numbers. The tendency is strongest in Claude (79.78%) and weakest in ChatGPT (35.53%). The results suggest that the longstanding SEO practice of numbering titles remains effective in the AI search era, but the impact differs by model.
The analysis, covering 60 prompts across purchase consideration, comparisons, and general queries, examined 542,251 cited URLs. After excluding URLs with unknown titles, the dataset of 395,396 titled URLs showed that 229,066 (57.93%) had numbers. Responses containing at least one numbered title reached 64.42% of all responses. When broken down by AI model, the gap between Claude and ChatGPT is 44.25 percentage points, implying that the content strategy for AI visibility may need to be tailored to the dominant model in the user's ecosystem.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.