An experiment conducted from July 19 to November 11, 2026, analyzed how generative search models cite company content. The study used a fixed set of 61 questions related to franchises and apparel production, excluding branded queries, to track citation frequency across four search surfaces: ChatGPT, Gemini, Perplexity, and Yandex.
The results showed a significant increase in the share of answers citing the company's pages, rising from 3% to 51% over the three-month period. Data was collected across six specific dates, ensuring a complete response set from all four surfaces for each question. The analysis excluded errors, such as API failures or empty responses.
The study revealed that ChatGPT and Gemini often utilized the same external search provider, effectively acting as a single source of citations. Perplexity and Yandex demonstrated independent search behaviors. The author noted that while the volume of published articles increased during the period, the correlation with citation growth remains observational.
The research highlighted that citation patterns are volatile, with over half of the sources changing between consecutive measurement dates. The author suggests that a 10-percentage-point difference between measurements falls within the margin of error, while the overall growth trend from 3% to 51% is considered statistically significant.
