Recently we've been learning a lot about how ranking in large language models like ChatGPT is different than ranking in Google, and so I just wanted to talk about one of those. Every web page has something called a meta title and a meta description. These are invisible on the page, but they're typically what you'll see on Google's search results page.
Blue title is the meta title, and the text below it is the meta description. The meta title is one of the most important ranking factors in Google, but we've known for a long time now that the meta description does not impact ranking. However, we now know that the meta description can have a huge impact on your visibility in large language models like ChatGPT.
And this is because of the way the large language models retrieve information from search engines. When you ask a question to a large language model, it'll perform maybe a dozen different searches and return hundreds of search results. But they don't want to spend the resources crawling every single one of those pages that's returned.
So instead it reviews the title, URL, and meta description of each page to determine which pages it should review. I'll make another video talking about how you can optimize your URL structure for large language models, but let's talk about the meta description. For now.
The best advice seems to be to spoil the answer in your meta description. So what does That mean well, consider the question that the user is asking the large language model before they review your page. You want to answer that question in your meta description so that the large language model is confident when they see that description that your page will have the answer to the question.