10 SEO Lessons That Still Work in the AI Era

So these are 10 SEO lessons that still work in the AI era. So number one, topic clusters remain powerful. So what are topic clusters? Topic clusters are when you have, let's say you're talking about SEO for example, right? You might have a bunch of SEO topics that are tied to it. So what is SEO? You know, how to measure SEO, how to do technical SEO. Like all these things kind of combined in the cluster. They'll still work in today's day and age. SEO hasn't changed that much. We like to think that just because AEO or LLMO sounds like it's different, it's just an evolution. And a lot of the old SEO stuff is still very powerful today.

0:00

So this isn't a number, but I'll add on to Eric's and then I'll give a number. One thing to keep in mind is SEO like five, 10 years ago, whatever the timeframe, a lot it was more keyword based. You do keyword research, you figure out what keywords you want to create an article around and you do it. Now Google and AI wants to rank sites and site brands that cover topic thoroughly. So if I was to blog about SEO, they want me to blog about SEO, SEO globally, how to do research for topics and keywords, how to build links, how to fix on page.

0:35

Right.

1:08

How to optimize for multiple platforms and not just Google, how to do local search, how to gain reviews. But do you understand how I'm going really deep into a topic? That's what they're looking for. So it's, it's, it's moved from keywords to topics. Number two, before the AI era, I used to build products, give them away for free. Cause it was cheaper than paid ads and the products would go, quote unquote semi viral. Build a lot of natural backlinks and goodwill and help your rankings and get.

1:08

Your brand out there.

1:38

That works even better now in an AI world because in an AI world people still use a ton of software.

1:39

Forget what the stock prices are showing, but a lot of these software companies.

1:45

Are still generating more revenue and you can see in their quarterly earnings.

1:48

And you can build, build software cheaper and faster with tools like Lovable out there. It may not be perfect and do everything, but it still works. And it's a great link magnet. And actually even today's world. And it was also a good citation magnet. Well yeah, that was number two.

1:52

Okay, number three. So the game very much is everywhere and we'll kind of COVID maybe a couple here. YouTube, for example, is cited quite a bit. When you look at the LLMs. So SEO to me, whether you're looking at like an answer engine, you know, platform or surface, or whether you're looking at traditional SEO, it's still kind of the same to me. So these answer engines and these search engines are grounded. YouTube is a major one that they look at. Top three are actually, you're looking at Reddit, you're looking at YouTube, Wikipedia. Obviously this changes over time, but the point is what used to be able to. YouTube is number two search in the world. So it still works really well. And I think the key thing to understand in terms of old SEO lessons that still works in today's era is that you have to adapt. Because SEO, as far as I know when I first joined, it's changed. And we're not talking about every year it's changed. It's like every, almost every week it's changing. And that's very much happening in the AI era.

2:11

Do you know, actually is the real number two search engine in the world, it's not YouTube.

3:09

What is it now?

3:13

Instagram.

3:14

Is it really?

3:15

Uh huh. A lot of people search for things that you wouldn't consider search. Like searching for a friend. Yeah. So there's a lot of. I think Instagram has like 6.5 billion searches a day, but they're not the.

3:15

Same type of searches. Intent is very weeny. Like.

3:26

Yeah, the intent.

3:28

Yeah, forget weenie.

3:30

The intent isn't as monetizable as someone typing in affordable laptop. Yeah, yeah, but it's still search. It's just.

3:31

I. Yeah, but technically.

3:40

No, technically it's search, but it's just a different type of search. Search. It's not always as monetizable. But Instagram still generates a crap ton of revenue.

3:42

No, for sure.

3:49

Majority of their ads are actually reels these days. I don't know if you knew that.

3:50

That I knew.

3:54

Yeah, yeah, yeah. So I had the percentages from. I think it was like Sensor, I think it's well above 50% are now reels. That is making Instagram their money from monetization. All right, so number. What are we at?

3:54

4. Number four. Back in the day people would be like, don't put all your eggs in the Google basket.

4:07

Even though it's the most popular search engine, they would say make sure you're on, Go make sure you're on Bing and make sure you're on any other search engine. We see this pattern going more and more. SEO shouldn't just be Google, it should be Google, it should be Bing, it should be ChatGPT. It should be Perplexity, it should be Instagram, because people do search on Instagram for content and inspiration. It should be Pinterest, it should be Amazon. You know, if you're selling e commerce products, search has changed from search engine optimization to search everywhere optimization. Even if people hate that acronym in which you have to optimize for multiple platforms, they all have their own algorithms. You got to learn them, get good at them.

4:13

But it typically starts with good product.

4:48

Good service, good content, good brand.

4:51

Number five, what's old is new again. People are like, oh, especially the new people coming into aeo, they're like, yeah, you know, it's these. If you spam the listicles, you're gonna rank really highly. And then it's like, oh wait, these LLMs are catching up. Catching onto the fact that listicle spamming is pretty easy to game. Now you have to do guest posting. Oh wait, that has been happening for a very long time. Oh, internal links are very good signals. Oh, yep, still sounds like SEO. So a lot of the old stuff in SEO still very much works well today.

4:54

Yes. What are we at? 6. So back in the day I used to buy companies that had a lot of brand queries. Brand queries is someone typing in the name of that domain name or the brand name into Google and then clicking over to a website versus just going directly helps with rankings quite a bit for the traditional Google sense. Strong brands also perform really well in LLMs like perplexity and in ChatGPT. So you can either try to build your brand. The bigger your brand is, the easier it is to rank on traditional SEO. But the bigger your brand also the easier it is to rank on these LLMs. And you can build a brand by doing omnichannel. Keep pushing out good information, products, services, right? Engaging in communities. But it takes a long time to build a brand. That's why I've tend to buy em over time versus just trying to build them. But either strategy could work.

5:24

Number seven, content refreshing still matters. When you look at LLMs, for example, you look at a ChatGPT, usually 95% of their content, depending on the content type, you're looking at updated content from the last 10 to 12 months or so. You have to make sure that your content's updated all the time. So this is never actually. This has always been the case with SEO. You gotta make sure that you're tending to your garden, that you're keeping your content updated. It's no different in today's AI era.

6:20

I had an Interesting question. On my webinar that I did earlier today, someone asks, how often should I update my content? And as you know, that's a loaded question because it varies per industry. And the example I gave there, because people are probably wondering how often they should update depends on your content type. If your content is on bananas and the nutrition facts about bananas, even if it's two years old, it probably doesn't need to be updated. Even if it's the speed of cheetahs and how fast they run. That one I gave as an example, I'm like, cheetahs run roughly.

6:45

You really like that cheetah example?

7:15

Yeah. Cheetahs run roughly the same speed that they did today that they did a year or two ago. So you don't really need to update much of that. But the way I look at updating content is you go check out your Google search console, you can compare year over year, week over week, month over month, quarter over quarter. You can pick the timeframe, and you can see which pages are losing traffic, losing rankings, losing impressions. Keep track of all those metrics, because sometimes you're losing clicks because of AI and AI overviews, but you're getting tons of impressions still. And you want to make sure that you're not losing tons of impressions, because if you're losing tons of impressions, someone's usually ranking ahead of you. And why are they ranking ahead of you? Well, they're probably doing something in their content that you're not. That's when you should look to see what your competitors are doing, because they're probably doing something you're not. And update your content to be more competitive. Competitive and better than theirs. And AI can help you analyze their content and then figure out the gaps in yours. And then you can update number nine. What number are we?

7:17

I'm on number nine. And then you finish with number ten.

8:19

Well, I didn't go on the last one. I was just expanding on yours. Oh, great. Go on it.

8:21

Go ahead.

8:24

Okay, so I'm on number eight. Yep. Number eight. In the world of SEO back in the day, people write articles on everything, like the nutrition facts about bananas and. And guess what kind of conversions people get from that stuff? Little to none. And today we're seeing people lose a lot of traffic on those basic queries, and they're getting no real effect in their revenue. And they're saying, my traffic's dying. My traffic's dying. Write stuff that hasn't been seen before on the web, stuff that's not regurgitated, and you'll do well, content can still be king. You just got to write about new stuff that isn't regurgitated. That's what did well in SEO years ago and it still does well today. The problem is people only want to focus on keyword gaps and content gaps and just cover all the stuff that's been beaten to death by a hundred competitors.

8:25

Yep. Number nine from my side, last from my side. Then you go number 10. Partnerships will never go away. So whether it's partnering with a website to help, just even a simple link collaboration, that's one thing. But you want to go way beyond that because SEOs back in the day, they used to kind of just stop at the link collaboration. You want to go further and figure out, hey, can we do a webinar together? Can we figure out how we can maybe cross pollinate our audiences, maybe with our different products? Right. What else can we do? Can we co host an event together where you actually develop a business relationship? And it can start with a link collaboration first, But I think the key thing is always expanding on that relationship that will never, ever go away.

9:15

Yeah.

9:53

Number 10, links are more valuable than ever in the LLM world. People are like, oh, you don't need a link, you just need people to mention you.

9:54

Well, if you can get them to.

10:01

Mention you and you get them out of link, even better if they link.

10:02

In most cases when they link to.

10:05

You, they're also mentioning you. Not always, because sometimes a link text says something else. But in ideal world, you're getting that brand mention and you're getting that link. It helps you with your traditional rankings and it helps you how often LLMs are citing you. And the way you get those, build good content, build free tools, amazing product and service, all that kind of stuff helps. And of course, being in business for a long time, the businesses that are really old tend to do better. But keep cranking on the content marketing because it helps you get more mentions and links.

10:07

So I want to, I want us to react to this. So Pep, one of your, your best friends from CXL said that agencies are getting fired. We surveyed B2B SaaS, mid marketing leaders about where they're cutting budgets. The number one answer wasn't paid ads, wasn't tools, wasn't events, it was agencies. Here's what they told us. Here's three quotes for you. Number one, agencies have no concept of what it actually means to run a business. Number two, they assume we have endless time to implement new ideas. Number three, I need execution, not strategy. I have plenty of strategy, 58% of these leaders are under extreme or high pressure to show short term pipeline. They don't have 90 days for discovery phase. They don't want another brand positioning deck. They don't need a 50 slide audit. They need someone to do the work. Yet most agencies still sell like it's 2015, week one through four, discovery weeks five through eight, strategy development weeks nine through 12, the playbook presentation weeks 13 plus. Now you, now you execute it. Meanwhile, the CMO needed results in week two. So the truth that agencies refuse to hear, mid market companies don't hire you because you're smart. They hire you because they're underwater. When you don't show, when you show up with strategy theater instead of execution, you're solving a problem they don't have. The agencies that survive won't be the ones with the best decks. They'll be the ones who show up day one and ask, what can I take off your plate today? So then you have this one has a thousand likes on LinkedIn, 350 comments. And you have all these agencies in here commenting over here, right?

10:38

Yeah, they're probably trying to disagree or something and.

11:58

Yeah. Or they're trying to come up with like a creative way to respond where it makes them kind of stand out. That's what they're doing.

12:00

So have you seen agencies get cut from your friend circle?

12:06

No, I think we're getting more inquiries than like usually January for us is a slower month. It started off with a bang.

12:10

So January is always busy for us.

12:18

I don't know why, like the past couple years it's been slow for us, but this year it's like the pace is like really quick.

12:21

So yeah, we've always, if we look.

12:27

At most of our clients, they do require strategy help, but it's guided strategy. It's not them saying like a Deloitte or a BAIN Or a McKinsey saying, hey, can you figure out what we should do to fix our business? I'll give you example of guided strategy. They may say, let's say they're, I don't know, I'm making up a company name. Let's say they're Walmart. And Walmart may be like, again I'm just creating a hypothetical here, but Walmart may be like, we know Google's awesome, we know Facebook's awesome. A lot of people are using chat GPT now. We want to figure out how to dominate GPT. Can you do analysis and figure out how we can do better on ChatGPT? And then if we like that, can you go and execute on it? So in essence, when I say guided strategy, they already know the channel. You got to figure out how to do it, what it's going to look like, pitch it to them, and if they like it, they'll pay you to execute on it. So we're seeing most our work, and it's always been this way is actually execution. That's what we mainly get paid for. The strategy has never been for us, hey, figure out how to grow our company. It's always been, these are our channels. Can you figure out how to do them better? Or hey, Facebook or Google's not working out as well. Can you come up with a new strategy for this channel and then figure out how to go and execute on it? But it's never been blind for us where clients are just like, maybe once in a while because we have so many customers. But I'm 99% sure the majority of the time they're asking us go and execute on a specific channel. But before you do, we want to hear your gameplay.

12:31

Let me just rephrase this, okay? Agencies are getting fired. Let's replace the word agencies with bad People are getting fired, which is a constant everywhere. Yes. The thing is, I would say most agencies, and we can put bad in front of it. The reason they've survived up till now is because the work was. There's more of like a gap, like a moat. Right. So before, let's look like 10, 15 years ago, many people didn't know how to do SEO. So if you even knew, like a little more like, you're going to get hired, right. And you're probably going to be able to maintain the client for a while because your client doesn't know what's happening. Okay. But people have gotten a lot smarter since that time. Right? So the bar continues to get higher. I would just say that a lot of agencies haven't met that bar, especially with AI now, if they're not going to adapt, that's why they're getting fired. Right. Because all this comes down to is how do you run your company? What is your company about? How well have you hired? If you haven't hired, well, then, well, that's why you get fired ultimately.

14:16

Right?

15:13

So when you hear. I will say this, I think we can both agree on this. Like when you and I, we talk to our friends, the reason why there is kind of contempt or like people kind of look down upon agencies is because they've had such bad experiences with the product, because the product is so. It can be inconsistent with bad agencies. Right. You might have a really good person, that person gets burnt out, they quit. Boom. Okay, now you're thrown onto a junior person. That's what I think most. That's why there's a general disdain for agencies. But look, just don't do it that way. Just set a higher bar. Easier said than done. But then you, I don't think you need to worry about getting fired. Because if you're that damn good, they can't afford to fire you.

15:13

Yep.

15:48

So on a side note, you have in here Sequoia agent led growth is a new market. I think this is like from Pat Grady. I could be wrong.

15:50

From Sequoia. I don't think it was from Pat Grady. Let's double check. No. Okay, so this is Sonia. Sonia, Tweety Bird, Sonia Huang.

15:57

Okay.

16:08

Partner at Sequoia. So she says we're going from the age of product led growth to the age of agent led growth.

16:08

Okay.

16:14

So product led growth is. That's when you might have a product where you have a crippled version of it and you give a lot of features away for free. Like HubSpot has a freemium version, I believe. So you see this most clearly if you're using Claude code actively. It says, hey, for a database, you should use Supabase. For hosting, use Vercel. It's choosing for you the stuff you should be using. So. Exactly. We're talking about using these agent encoders.

16:14

Right?

16:37

So product led growth brought us closer to division of best product wins. But ultimately people are still lazy. Yes, people are still lazy. That's never going to change. They can't read all the reviews, nor do they want to read the reviews, to be honest. And they kind of default to just what looks cool on the website. So whereas your agent has infinite time to go and make these choices for you, it will read all the documentation, read all the user comments, and figure out what you need for your use case. So again, instead of product led growth, where you have to kind of play with the product and look, look at G2 reviews and all that type of stuff, right? We're now moving to a world where robots are just going to be talking to each other all the time. What would that water taste? Nasty.

16:37

No, no, I've already read this, but.

17:14

I have a very different take than what?

17:16

Go for it. Go for it. Oh yeah, it was packed.

17:18

You're right. Yeah.

17:20

Their whole take is marketing agents will start choosing everything. But I haven't read it fully, but I saw it on social media and one obvious thing is agents will look at your G2 reviews. Agents will look to see what is popular. Agents will see what people love and what they're saying. This is still marketing. You're optimizing for algorithm. We were doing the same thing for Google. Yes, it was a human who clicked on a link and then bought, but now you're optimizing for the same metrics, the same signals, so an agent can click and buy. There's not much that's changing on the consumer end or the marketer's end.

17:22

Yeah, I do think, I mean, I don't think it's unfair what they're saying, though. I think it is. We, we will be as marketers, we're going to be optimizing for more robots.

18:01

But you already optimized for robots. What robot do you not optimize for? You optimize for Google, ChatGPT, Facebook, Instagram, Snap.

18:10

What I mean is, you're not wrong, but I mean, the behavior of these agents is a little different than like a Google bot, for example. Right.

18:18

So they look at sentiment on Google Bar and Chad GPT for things like their AI. They look at your brand versus the competitors. The context are looking at a lot of the similar signals. Yeah, it's just an evolution.

18:25

So that's what I'm saying. I think we're on the same page.

18:37

Yeah, it's not a drastic evolution to be very drastic.

18:39

I'm like, I don't see this as.

18:43

No, drastic. I think they're trying to paint a new word and coin another word. And that's. Andreessen's pretty good with the media play, but I think they are trying to coin a new word and sometimes it sticks, sometimes it does. Like founder mode, for example. Product like or like content marketing. Right. I think they're trying to do that. I just think marketing has always been an evolution. So. So that's what's going to happen. And I, I don't know when's the last time these people were, you know, operators. But, you know, SU is a legit company or a legit firm.

18:44

Yeah, they're a legit firm. Yeah. Yeah.

19:10

And we'll catch you.

19:12