You Ask, I Answer: Future of SEO and ChatGPT Impact?

In today's episode, Chris asks, do people still hire SEO firms? Yes, they do. Absolutely they do. Um, there is still a need for companies to help other companies learn how search marketing works. However, and this is the big however, there is a huge question mark on the future of traditional SEO right now.

And that huge question mark has to do with large language models. The one that you're most familiar with, of course, uh is the one everyone can't stop talking about, chat GPT, which is based on the GPT 3 language model. But this is not the only one. There's many of these out there GPT J6B, GPT Neo X20B, and so on and so forth. Some are run by OpenAI, some are not, and some are by um uh Ulithra AI, uh Hugging Face.

There's there's tons of these models that are available. Here's why this is a big deal. Right now, when you use Chat GPT or any large language model, you ask it a question, and it gives a response. It gives sometimes a factually incorrect response, but it gives a response that seems plausible. And there are plenty of companies that are working on factually correct versions of these tools.

When you ask it, for example, things like, what is a transformer? Or how many cups are in a gallon, or things like that, in a chat format, in a conversational format, they're not sending you anywhere, right? When was the last time you got a link to click on out of chat GPT? Never. You'd never have gotten one.

You've had to maybe Google for something afterwards, but you've got no link from it. No, you might also be interested in or check this out. And so, as these models take off, and as this style of interacting with knowledge continues to grow, that imperils SEO. That's one of the reasons why Google is having a freak out about ChatGPT because it imperils the very core of their business, right? It imperils their ability to serve ads along search listings, which is at the heart of a good chunk of their business.

Now, they do have other things, display advertising, they have YouTube ads, Gmail ads, etc. But the big money is in those search ads. And if search changes dramatically, they will have a hard time making up that revenue. Here's why it matters for us as marketers. When you look at how our large language model is constructed, particularly for unbranded search terms, you know, informational searches, it's not from any one source.

So our text that we publish on our blogs and stuff feeds these large language models, but we don't get credit for it. And as a result, for those large language models, they could take away a substantial amount of our unbranded search traffic. There will still be branded search traffic, right? Someone who's looking for trust insights will still search for trust insights. And a large language model is not going to say, oh, you should go to trustinsights.ai, right?

Um a regular search engine will do that. But if a certain if someone is searching for uh reasons why AI models are good or reasons why AI models are bad, those queries aren't going to go nowhere. They are going to go nowhere except in the chatbot itself. And so all of our informational content that we use to drive unbranded search volume is going to go away. Not completely, but it will probably go away substantially.

What will it look like? Kind of like what chat looks like now. Think about it. When you ask your Google Home or HomePod or Alexa device for something, it will say, you know, here's the answer to your question. Would you like more information?

We'll send you a link in your uh thing, or there's a link displayed on your uh Alexa device or whatever. How many links are there? One. Just one. Which means if you're not number one, you're nothing.

Right? If you are not number one, you are nothing. So SEO is going to dramatically change. It will dramatically change as long as chat-based interfaces to large language models continue to be popular. So what do we do about this as marketers?

Now is the time to be building your insurance policy. And your insurance policy comes in two parts. Part number one, your brand. You must, must build brand. You must build brand.

You must be known by name so that even asking the most sophisticated language model still points people towards your website. Hey, go check out trustinsights.ai. Go check out ChristopherSPenn.com. Go check out marketingovercoffee.com. That is number one.

You've got to build brand. And if you're not building brand, you are in danger. Should you still be building unbranded content marketing, you know, uh informational marketing? Yes, but recognize that you're at this point you're just feeding large language models. It will not benefit you.

Build your brand instead. Number two, you must build community. Right? You must have an acquisition method that gets uh people to you in what is called reliable reach. That means your email marketing list, your text messaging list, your Slack, your Discord, your Telegram, whatever means you can to get people into a place where you can reliably reach them is going to be important because you can't do it on sub public social media.

You just can't, right? Algorithms like at Twitter and Facebook and LinkedIn and Instagram and TikTok, you cannot reliably reach people when you need to reach them. They're simply not accessible. You will get some of them when you publish your latest TikTok or whatever. But it could be hours, days, maybe even weeks before enough of your audience sees a public social media post to matter.

With private social media, you get the timeliness. So that's important. Um with an email list, as long as your email service provider working correctly, you will you will get that. Um you have to be prepared for your organic search to dramatically drop. So your assignment is this go into Google Search Console and look at your performance for branded and unbranded queries.

You'll have to do some some regex selection in in there. You want ideally 50% of your traffic to be branded search. At least 50% to be branded search. And this flies in the face of everything we've been saying about SEO for 10, 15 years. You know, you need unbranded search to bring in new audience.

Once large language models take over search, you are just feeding the models and you are getting nothing from it. You will be just feeding models that will run ads and take your money, and you will have nothing to show for it. So I would focus on building brand. Now, some of it, sure, unbranded content, explanatory content and things is important. But I would start peppering your unbranded content with your stuff, right?

Things that are uniquely you, things that people can get nowhere else. Um it will read a little bit awkwardly, but can you imagine? You know, if uh if a large language model decides to vacuum up your content, it will strongly associate your names with that content. That's how those models work. Otherwise, do what you can to build brand.

So advertise for sure. Show ads to relevant people. Create shareable, share-worthy content, stuff that people want to hand out that has your name and your logo on it. Um, studies, research, unique stuff. Gate it.

I'm gonna say something that again is gonna piss off a lot of people. Gate your content. Why? Because ungated, ungated content, you're just feeding large language models. You're feeding GPT 3.

And it's gonna take your content and your knowledge, and you will get nothing from them. When you slap it behind a gate, even if it's just a gate like, you know, punch the monkey in the face, right? No, you don't have to take people's personal information, but just put up something that blocks these crawlers from getting your information from taking it and using it for themselves. Put some obstacle in their way so that your knowledge is can benefit you first, right? But that's how you're gonna prove your brand value and then build those communities.

Build those communities. You've got to build private communities that you can reliably reach. Otherwise, the large language models will invariably benefit those who can crank out the most content. Those who can crank out the most content and sway those models a little bit will inevitably be those companies that have the deepest pockets. You and me, who are I'm assuming you're not a billionaire.

Maybe you are. If you are, hi. Um, I'm not a billionaire. But you and me, we are not the beneficiaries of most AI models, right? They are created on large volume, and large volume correlates strongly with resources, with money, etc.

So that's my long take on SEO and where things are going. Take it for what it's worth. If you disagree, cool. This is how I'm looking at it for my own marketing because I know how these models work, and I know that we are not the ones who are going to be uh benefiting unless we are the ones who are running the model. If you are an AI company and you can get market share, you will have a disproportionately larger and larger amount of market share as long as your company continues to do well because people will rely more on more of those models, and pretty soon there won't be room for competitors who don't have AI on their side, who don't have models of their own that they're deploying on you know with customers and in customer interactions.

It's going to be a very interesting few years in marketing, but the writing is already on the wall that SEO is in a lot of trouble, and it's not going to get better. So hope that answers your question. Probably a little more than you were expecting, and I'll talk to you soon. Take care. If you like this video, go ahead and hit that subscribe button.