@tjrobertson52

There's a lot of misleading information coming out about how to get recommended by ChatGPT. I wanna talk about a recent blog post from hrefs that shares some really interesting data about ranking in ChatGPT, and then talk about how it can be a bit misleading. And then finally, I'll give you my take on what's actually been working for us.

So in the title of H R s article, they say that 67% of the top 1,000 pages cited by ChatGPT are off limits to marketers. And by off limits, they mean that you, as a marketer, are unable to change or influence the content on those pages. And you might be thinking, why would I need to change the content on those pages?

Well, it's because that is the primary method people are using to get recommended by ChatGPT. After someone enters a prompt in ChatGPT, it will typically do a search, and the response it comes back with will typically just be a summary of all the pages it cited. So if you can get your brand or business recommended in the pages it citing, it will recommend you in its response.

So this article from H Ref seems to be implying, uh, good luck doing that, because 67% of the citations aren't able to be influenced. And that's what's very misleading about the article. So first let's talk about how they're getting that 67% number.

That number is made up of Wikipedia, which is about 30% of all citations. Homepages and app store pages. So since you're unable to influence what Wikipedia says, or the homepage of websites or app store pages says, they consider these pages dead.

And that's all well and good if your intention was to show up for most queries in ChatGPT, but most queries in ChatGPT are people looking for information. If you're a brand or a business, all you really care about is being recommended when someone's looking for the product or service you offer. We're currently tracking around 15 prompts across 30 industries, the kind of prompts people type when they're interested in a recommendation for a specific business.

And I can tell you, Wikipedia is not even in the top 10. Far from 30%. Wikipedia is well under 1% of the citations for those kinds of prompts.

And while homepages are a good percentage of the citations, it's typically the homepage of your website and your competitors websites. Obviously, you have tremendous influence over your own homepage, and while you probably can't get your competitors to recommend you on their homepage, it does give you a lot of data you can use to influence the recommendations. And that brings me to the second reason that this article is misleading.

The implication seems to be that the only way to influence ranking is to influence the content on these pages that are being cited. But that's not true. You also have the option of replacing the Pages that are being cited.

If you know what kind of information ChatGPT is looking for before it gives a response, you can create pages optimized for those terms. So, for example, if you know that ChatGPT is looking at known competitors, you can create content targeting those competitors. Name free pages like your brand versus that competitor or that competitor alternatives.

Check out this kind of content and it's likely to surface when they're looking for your competitor. If ChatGPT is looking at blog posts, which is very common, you can create similar blog posts on your website, assuming you have about as much authority as the websites that it's citing currently. If ChatGPT is citing LinkedIn articles or medium articles, which is really common, you can also write articles on those platforms.

The truth is, for the props that you care about, the citations are highly influential. So while hrefs is sharing some great data, I really do recommend that article. We're still in the early days of this industry.

Without the full picture, it's easy to be misled by data like this.