Hi there, this is our guide for ChatGPT Search Engine Optimization (some call it already Generative Engine Optimization (GEO) and how to rank with your content and websites on ChatGPT.
Introduction to ChatGPT and its role in SEO
So, you’re probably thinking: Is ChatGPT a relevant search engine for me and my audience?
Developed by OpenAI and trained on extensive datasets, ChatGPT represents a significant step in the evolution of AI chatbots and virtual assistants.
ChatGPT per se was not (at least in my understanding) intended to replace Google Search, however, due to the nature of the product, many consumers these days rather use ChatGPT as an all-in-one chat experience – including SEARCH. Meaning they also search on ChatGPT and use it over Google.
Sure, Google still is – /and will be for the short term – the main number 1 search engine.
As a software, ChatGPT can be applied in diverse areas such as generation tasks, customer service enhancement, content generation and classification, language translation, and even creative fields like writing fiction or poetry.
With that said, ChatGPT became famous for reaching 100 million users within 2 months. Right now, chatgpt.com has about 2.4 billion visits per month (according to similarweb).
SearchGPT coming soon….the impact on SEO?
With that said, OpenAI recently announced SearchGPT – a new search engine built on top of its own LLM. SearchGPT is OpenAI’s AI-powered search challenging Google Search. Key features include a conversational interface, real-time web access, and clear source attribution. While promising, it remains to be seen how users will adopt it….right now you can sign up for a waitlist.
Step-by-Step Guide to Keyword Research Using ChatGPT
Finding Keywords and Prompts
Hey, so when you’re diving into SEO or GEO, the starting point is always keyword research.
You can obviously use tools like SEMrush or ahrefs here. Alternatively, I would recommend you two tools:
- ChatGPT itself: Just type “Give me 20 keyword ideas for [your niche]” and don’t forget to sprinkle in some local keywords to reach your nearby peeps.
- Otterly.AI: Keep in mind that modern AI search engines like perplexity.ai rather are used in a conversational way than pure keywords. With Otterly.AI’s Keyword Research Tool, you can generate ideas for prompts based on certain keywords.
Monitoring Keywords & Prompts
Before you start creating content clusters or articles with the goal of being found on ChatGPT, you need to perform an audit to even understand:
- Which websites are ranking for my particular keywords?
- Which text is being generated by AI for particular keywords?
- Is my brand mentioned for those particular keywords?
Since most traditional SEO tools do not support ChatGPT (or its LLMs such as GPT-4 or GPT 4o) for monitoring, we built Otterly.AI in the first place. With Otterly.AI, you can monitor any keyword and prompt and get answer to those questions.
Let’s dive in.
How to track your brand on ChatGPT
Otterly.AI allows you to monitor how visible your content and brand is within ChatGPT. In addition, Otterly.AI will give you all the links mentioned on those AI-generated answers. In other words: You will now be able to know if your websites (or any other URLs) are mentioned for any prompt / keyword you’d like to monitor.
If you are new to Otterly.AI, you can sign up for free right here.
After signing you for Otterly.AI you can add your first keyword that you’d like to monitor.
Wait a minute or two and Otterly.AI will provide you direct insights into the following questions:
- What URLs are mentioned on ChatGPT for this particular keyword?
- Which brands (if any) are mentioned on ChatGPT for this particular keyword?
- What’s the brand sentiment (if any) within the AI-generated answer?
What URLs are mentioned on ChatGPT for this particular keyword?
As a marketer, I’m all about that website traffic!
I’m curious to see which URLs are getting some love on ChatGPT. I’m especially keen to know if my page is in the mix and where it’s placed. Just like in regular search results, the higher the link, the more clicks and traffic I’ll snag. Otterly.AI lays it all out for you with the links, anchor texts, and positions. You can even click on them to scope out the pages. It’s a cool way to see how your URLs stack up against the rest. Check out the screenshot below!
Which brands (if any) are mentioned on ChatGPT for this particular keyword?
Otterly.AI helps you see which brands are talked about and how people feel about them when it comes to business-related keywords. In the screenshot, the search query “Best Hubspot alternatives” gives us a brand sentiment analysis as well as some mentioned competitors for ChatGPT.
With a long list of brands mentioned, the marketing and SEO team can now start working on the optimization part. Is my brand correctly described and do I feel comfortable with the brand sentiment How can I increase my visibility by being better positioned on ChatGPT?
Overview of the Ranking Algorithm: What are the key ranking factors for ChatGPT?
The ranking factors for ChatGPT responses can be understood in terms of several key criteria that influence the quality, relevance, and effectiveness of the generated output. While OpenAI does not disclose all specifics of its proprietary models, the following are generally considered important factors for ranking the responses produced by models like ChatGPT:
1. Relevance to the Query
- Contextual Understanding: The response must be contextually relevant, directly addressing the user’s query or statement.
- Accuracy: The information provided should be factually correct and up-to-date.
- Specificity: The response should be as specific as possible, focusing on the user’s request without unnecessary generalizations.
2. Coherence and Fluency
- Logical Flow: The response should be logically structured and easy to follow, with ideas connected clearly.
- Language Fluency: The text should be grammatically correct and free from awkward phrasing or spelling errors.
3. Completeness
- Thoroughness: The response should cover the key aspects of the query comprehensively without leaving out critical information.
- Detail: While being concise, the response should include relevant details that add value to the user’s understanding.
4. User Intent Understanding
- Intent Recognition: Correctly interpreting the user’s intent, including recognizing when a query is ambiguous or when multiple interpretations are possible.
- Tone Matching: The response should match the tone or style appropriate for the user’s input, whether formal, informal, technical, or conversational.
5. Engagement and Interaction
- Engaging Content: The response should be engaging, potentially prompting further interaction or encouraging the user to explore more.
- Proactivity: Anticipating additional information or clarification the user might need and offering it proactively.
6. Bias and Ethical Considerations
- Neutrality: Avoiding biased or controversial statements, ensuring the response is neutral and fair.
- Ethical Adherence: Respecting ethical guidelines, such as privacy, non-discrimination, and avoiding harmful content.
7. Efficiency
- Conciseness: The response should be as concise as possible while still providing a complete and clear answer.
- Time-Sensitivity: If relevant, the response should be timely, acknowledging any time-sensitive factors in the query.
8. Consistency with Known Data
- Alignment with Facts: Responses should align with well-established facts and knowledge, avoiding contradictions or errors.
- Up-to-Date Information: Using the latest available information, especially in fields that evolve rapidly like technology, health, and science.
9. Adherence to User Preferences
- Customization: Respecting any specific instructions or preferences set by the user, such as format, length, or style.
- Learning from Interactions: Over time, adapting to the user’s preferences based on previous interactions.
10. Safety and Appropriateness
- Content Moderation: Ensuring that the response is safe, appropriate, and avoids any language that could be offensive or harmful.
- Guardrails Compliance: Adhering to guidelines that prevent the generation of inappropriate content, misinformation, or harmful advice.
These factors collectively help determine the quality of a response generated by ChatGPT, influencing how it ranks in terms of usefulness and user satisfaction. The interplay of these elements is critical for delivering high-quality, effective, and contextually appropriate responses.
By focusing on these factors, SEO professionals can better align their strategies with ChatGPT’s unique approach to ranking content.
How does ChatGPT’s approach to ranking differ from Google?
ChatGPT’s approach to ranking responses differs from Google’s search ranking in several fundamental ways, reflecting the different goals, underlying technologies, and use cases of each system. Here are the key distinctions:
1. Purpose and Goal
- ChatGPT: The primary goal of ChatGPT is to generate coherent, contextually relevant, and engaging text-based responses in a conversational setting. It’s designed to interact with users in real-time, providing direct answers, explanations, or creative content based on the input it receives.
- Google Search: Google’s goal is to index, rank, and retrieve the most relevant web pages in response to a user’s search query. It aims to provide users with a list of web pages that are most likely to satisfy their information needs, often directing users to external content.
2. Ranking Mechanism
- ChatGPT: Ranking in ChatGPT is internal and implicit, driven by the model’s architecture (like GPT-4), which generates the most probable continuation of text based on input. The “ranking” is essentially the sequence of tokens (words or phrases) with the highest probability of being relevant to the context and query, as determined by the model’s training on vast datasets.
- Google Search: Google uses a complex and explicit ranking algorithm (e.g., PageRank, combined with many other factors) to determine the order of search results. These algorithms take into account hundreds of factors, including content relevance, site authority, user engagement metrics, backlinks, and freshness of content, to rank web pages.
3. Data Sources
- ChatGPT: The responses are generated based on a large pre-trained model that has been trained on a diverse dataset, including books, websites, and other text-based sources up to its last update. It doesn’t access live data or the web directly when generating responses.
- Google Search: Google accesses and indexes billions of web pages in real-time, pulling from current and dynamically updated content across the internet. It uses web crawlers to constantly update its index and reflect the most current information available online.
4. User Intent and Interaction
- ChatGPT: Interaction with ChatGPT is conversational. The model attempts to understand the intent behind each input and generate a response that directly addresses the user’s needs within the ongoing conversation. It adapts to the context provided in previous exchanges within the same session.
- Google Search: Google interprets user queries, often inferring intent through search patterns and user data, and provides a ranked list of web pages or direct answers (like featured snippets) that it predicts will fulfill the query. Users then interact with this list by clicking on results to access more detailed information.
5. Personalization and Learning
- ChatGPT: ChatGPT doesn’t retain information between sessions unless specifically programmed to (e.g., in some fine-tuned applications). It can adapt to the tone and style of the conversation within a session, but it doesn’t learn from previous interactions to influence future ones.
- Google Search: Google personalizes search results based on a user’s search history, location, and other personal data. It learns from user behavior over time to refine search results, offering a more personalized experience that can vary significantly from one user to another.
6. Transparency and Control
- ChatGPT: The process by which ChatGPT generates responses is opaque to the user. There’s no direct insight into how the model arrived at a particular response, and users cannot see or control how it ranks potential outputs.
- Google Search: While Google’s exact algorithms are proprietary, users have some insight into why certain results are shown (e.g., due to SEO practices, user history, etc.), and Google provides tools for users to control aspects of their search experience, such as filtering results, setting preferences, or using incognito mode.
7. Output Format
- ChatGPT: The output is typically a single, well-structured response that attempts to fully address the user’s query in one go. There’s no need for the user to sift through multiple sources or results.
- Google Search: The output is a list of links to web pages, each potentially offering different information or perspectives on the query. Users must browse through multiple results to find the most relevant information.
8. Bias and Content Moderation
- ChatGPT: The model is trained to avoid biased or harmful outputs and has built-in guardrails to prevent generating inappropriate content, although it is not perfect. The ranking of responses is influenced by these safety protocols.
- Google Search: Google applies content moderation and ranking adjustments to de-prioritize or filter out harmful or low-quality content, but the user may still encounter biased or controversial content depending on their search terms and the results’ rankings.
In summary, while both ChatGPT and Google aim to provide relevant information to users, their approaches to ranking, personalization, and interaction are tailored to their respective use cases: real-time conversational response generation versus indexing and retrieving web-based content.
Summary: How to Rank on ChatGPT with SEO
I hope you found this article helpful and provided you with some first insights on how to rank in ChatGPT by conducting proper SEO. BTW: Please check out Otterly.AI for monitoring your SEO efforts ongoingly – we support Perplexity.AI, Google AI Overviews and ChatGPT.