# GEO: Generative Engine Optimization

**Authors:** Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik R Narasimhan, Ameet Deshpande  
**arXiv:** [2311.09735](https://arxiv.org/abs/2311.09735)  
**Date:** 2023-11-16  
**Summary page:** https://www.summarizepaper.com/en/paper/2311.09735v1/

## Key points

- Large language models (LLMs) have revolutionized the search engine landscape, leading to the development of Generative Engines (GEs).
- GEs utilize generative models to provide accurate and personalized responses to user queries, replacing traditional search engines like Google and Bing.
- This shift enhances user utility and increases traffic to generative search engines but poses a challenge for content creators who have little control over how their content is displayed in these black-box systems.
- The solution proposed is GEO, a novel paradigm that introduces impression metrics such as "Word Count" and "Subjective Impression" to measure the impact of citations on user attention objectively and subjectively.
- GEO offers generative engine-agnostic strategies for website optimization, aiming to enhance website visibility through stylistic alterations and structured content modifications.
- A general algorithm is presented within GEO to measure impression metrics of a response, along with a benchmark called [Benchmark Name] comprising diverse user queries across multiple domains.
- Through rigorous evaluation, it is demonstrated that GEO can boost visibility by up to 40% in generative engine responses, with effectiveness varying across domains necessitating domain-specific methods.
- This work opens up new possibilities in information discovery systems with implications for both generative engines and content creators, providing tools like GEO to navigate challenges posed by generative engines and ensure fair representation in the creator economy.

## Summary

The emergence of large language models (LLMs) has revolutionized the search engine landscape, leading to the development of Generative Engines (GEs). These GEs utilize generative models to provide accurate and personalized responses to user queries. As a result, they are rapidly replacing traditional search engines like Google and Bing by synthesizing information from various sources with the help of LLMs. While this shift enhances user utility and increases traffic to generative search engines, it poses a significant challenge for content creators who have little control over how their content is displayed in these black-box systems. To address this issue,  is proposed as a novel paradigm to assist content creators in improving the visibility of their content in Generative Engine responses. GEO introduces impression metrics such as "Word Count" and "Subjective Impression" to measure the impact of citations on user attention objectively and subjectively. Additionally, GEO offers generative engine-agnostic strategies for website optimization, ranging from stylistic alterations to structured content modifications that aim to enhance website visibility. Furthermore, a general algorithm is presented to measure impression metrics of a response, along with a benchmark called  comprising diverse user queries across multiple domains. Through rigorous evaluation, it is demonstrated that GEO can boost visibility by up to 40% in generative engine responses. The effectiveness of these strategies varies across domains, highlighting the need for domain-specific methods. In conclusion, this work opens up new possibilities in information discovery systems with profound implications for both generative engines and content creators. By providing tools like GEO, content creators can navigate the challenges posed by generative engines and ensure that they are not disadvantaged in the creator economy. The detailed framework and methodologies outlined in this study pave the way for future advancements in optimizing content visibility within generative search engines.