With the rise of large language model (LLM) driven search engines like Bard and Perplexity, traditional SEO strategies are becoming less effective for content creators. To address this challenge, Princeton University and the Allen Institute for Artificial Intelligence collaborated to develop GEO, a new content evaluation metric focused on generative engines. GEO aims to help content creators better understand the performance of their content in generative search engines and provide optimization strategies to improve the visibility and effectiveness of content.
With the rise of LLM-based search engines such as Bard & Perplexity, robots directly output answers, which makes it increasingly difficult for content creators to improve their websites through SEO. To help content creators better understand how their content performs in generation engines and provide strategies for optimizing this content to increase its visibility and effectiveness in generation engines, Princeton University and the Allen Institute for Technology present GEO. GEO proposes an impression metric specifically for generation engines. The principles of GEO include multi-modal understanding, content comprehensiveness and semantic understanding. By implementing the strategies proposed by GEO and participating in the GEO-BENCH benchmark, content creators can improve the visibility and effectiveness of their websites and content in generation engines and better meet the search needs of users.The emergence of GEO provides valuable guidance for content creators in the new search environment, helping them better adapt and optimize content through key elements such as multi-modal understanding, content comprehensiveness and semantic understanding, thereby achieving a new goal in generative search. Gain greater visibility and influence in the engine. In the future, GEO and its benchmarks will continue to improve, bringing more possibilities to the field of content creation.