Interestana
Home/News/Gemini Local AI Search Favors Business Websites: Study
Search Engine Land3 min read

By Interestana AI Editorial — AI-drafted, human-overseen. How we report

Gemini Local AI Search Favors Business Websites: Study

Gemini Local AI Search Favors Business Websites: Study

Business websites accounted for nearly 60% of citations in Gemini's local AI search results, according to an analysis by Steady Demand. The study examined 14,472 citations derived from 1,487 distinct local search queries conducted across 50 major U.S. metropolitan areas. This finding indicates a strong preference within Gemini's AI for direct business sources over other types of online content. The research also highlighted a significant issue termed "Grounding Drift," where identical Gemini queries produced substantially different AI recommendations and cited sources. In a direct comparison, repeating the same Gemini search query resulted in overlapping cited sources only about 40% of the time. This contrasts sharply with Google's local pack, which returned the same top listing approximately 90% of the time under identical search conditions. Furthermore, Gemini recommended the same top business in its AI responses only about 7% of the time for repeated queries, suggesting that a single AI search result should be viewed as a temporary snapshot rather than a stable indicator of visibility.

Reddit emerged as the second most cited source in Gemini's local AI search, capturing 13.7% of citations. This performance by Reddit surpassed the combined total of all local service directories, including prominent platforms like Angi, Thumbtack, and HomeAdvisor. The dominance of direct business websites and the strong showing of Reddit suggest that Gemini's AI prioritizes user-generated content and direct business information when responding to local search queries. The study's methodology involved analyzing local search queries across 10 different service categories within the specified metro areas, providing a broad overview of Gemini's citation patterns.

When the same 1,487 local search queries were run through ChatGPT for comparative analysis, significant divergence in citation patterns was observed. Both Gemini and ChatGPT cited many of the same domains only 8% of the time, and they recommended the same top business just 4.2% of the time. This indicates that different AI models are not only drawing from different pools of information but also employ distinct ranking or recommendation algorithms. Specifically, Gemini predominantly cited business websites, whereas ChatGPT showed a greater reliance on Reddit and business directories. This disparity underscores the importance for businesses to understand that success or visibility on one AI platform does not automatically translate to another.

Ben Fisher, co-founder of Steady Demand, emphasized the implications of "Grounding Drift," advising that businesses should not assume that success in one AI engine will automatically carry over to others. Monitoring visibility across multiple AI platforms is crucial, as relying on a single platform could present a misleading picture of overall AI visibility. The study's findings are particularly relevant for local businesses aiming to optimize their online presence for AI-driven search, suggesting a need for tailored strategies for each AI search engine. The analysis provides concrete data points on citation percentages and recommendation consistency, offering actionable insights for SEO professionals and business owners navigating the evolving landscape of AI search.

Original source — read the full reporting at the publisher:

Read on Search Engine Land

Get the weekly AI digest

AI news + new model releases, weekly. Drafted by our agents, reviewed by humans.

Read next