How We're Actually Using AI

The way we're going to run this one is probably 35 minutes of content or so, and then we'll leave 10 minutes for questions, and we'll have mic runners towards the end, I believe. But let's go ahead and get started. So we're going to be talking about how to actually use AI to drive revenue. I know all the sessions for HubSpot, like AI, AI, AI. We're going to hopefully share a couple of case studies, things that we're trying right now that are working for us, and maybe some things that aren't working so well for us, just to share real-life experiences. But quick introduction, Neil, if you want to do a quick one-liner, I'll do one and we can get into it.

Sure. My name is Neil Patel. I've been in the marketing space for 20-something years and I run an ad agency called NP Digital.

Yep.

And my name's Eric Siu. I've been doing marketing, what, 14 years or so? We both have our little YouTube channels and we do this little podcast called Marketing School. Just quick show of hands, who here actually listens to the Marketing School podcast? Okay, see what I said? I said less than 30%.

That's great.

I was right. But that's good. That's good for us. Please subscribe. So I think the first thing to call out, so Neil and I both have SEO backgrounds, and those of you that are familiar with SEO are probably used to seeing this traffic pattern now in your Google Search Console. This is known as an open jaw pattern. And thank you. Open jaw pattern or alligator pattern. And what this is basically saying is that your impressions might be going up on your website, but your website traffic is actually going down. And so there are a lot of implications around this and we're going to spend kind of the next couple slides talking about it and what we can, how you can actually counteract this. So Neil, I'm going to turn it over to you.

But before you think about the alligator chart, if you click the back one, with the alligator chart, a lot of companies are worried that, hey, I should stop doing what I'm doing. I should stop blogging. I'm going to be screwed. I'm just going to waste a lot of money and get no results. If you stop what you're doing, what you'll find is you won't get cited by the LLMs a lot because ChatGPT and Perplexity and even Google are leveraging blog posts, which is a primary way that most of you guys in here are getting a lot of organic traffic. So if you stop doing that because you're like, hey, I'm no longer getting the Google traffic I used to, what you'll find is you won't do as well on ChatGPT and Perplexity and all the other LLMs. So just keep that in mind. That's really important. And just because traffic goes down, that doesn't necessarily mean revenue goes down. And if your revenue is going down, there are ways to counteract that.

Yeah. And before we move to the next slide, what's working out well from an SEO standpoint right now, from a website traffic standpoint, is more people are creating more bottom-of-the-funnel transactional content and people are getting away from more top-of-the-funnel. We're not saying don't do top-of-the-funnel content, but more of the bottom funnel, that type of content tends to convert well on ChatGPT. So I was looking at some of the leads that we've been getting for Single Grain the last couple weeks or so, and some of these are, we're talking 9 or 10-figure companies, so significant companies, and a handful of them are coming from ChatGPT and are going to very high-intent pages. So for us, that's Reddit marketing agency or SEO marketing agency in Los Angeles, top digital marketing agency. So that type of stuff works. Again, that's just one of the actionable things that you can do, and more to come, but I'm going to move to the next slide so Neil can jump into his beautiful graphs here.

So when you think about AI search, when you look at generative search just in general, there's a lot of reasons that people end up using them and it's shaped by things like personalized journeys or conversational search. And when we really look at the search trends right now, if I had to give something that's super actionable is what we found is when you create really detailed content, and I'm going to use HubSpot as an example because everyone can relate, I was on a podcast yesterday with Kip, and I don't know when it'll be published. He told me on the podcast that roughly 13% of their revenue is from LLMs like ChatGPT and stuff like that, where people are coming and signing up. This is for like new revenue, I'm assuming, right? Because they've been established for a long time. It has to be new revenue. When we look at how to get traffic from these sources, we're seeing patterns change on how people search. It used to be where they searched for 1 or 2-word phrases years ago, then it started getting more specific with 5 or 6-word phrases, and now it's getting ultra-specific where people are typing in sentences. So the HubSpot example is best CRM for a construction company with less than 100 people, right? You see how that's super specific? Because people will ask, I'm a construction company and we're small. And I'm looking for a CRM, what's affordable and what's a good fit. So then they'll create an article that's super detailed about this that LLMs will pull from it and it gets their brand out there and it's driving revenue. You won't get the clicks that you're looking for, but it drives revenue. And at the end of the day, clicks don't matter. It really is revenue. But when you get very specific with your content like that and you publish it at mass scale, what you'll find is You'll get pulled into these LLMs a lot and you'll drive quite a bit of revenue. But the key with doing this, and a lot of people don't do this, they'll just create the content like top CRMs for a construction company that's under 100 employees or midsize or small size, and they won't include comparisons. So what we found with the LLMs is everyone here has competition. Even if you're like Uber, like we're revolutionizing the taxi, you know, industry and you're starting from scratch and we're going to, revolutionize how people have transportation, you're still competing with taxis. Waymo, in theory, is competing with Ubers, right? You still have somewhat of competitors, even if you're like creating a brand new company. And if you do comparison tables and you break down how you're different, but more so better for that specific vertical, and that's in the comparison table. Most people with the comparison table, they just talk about how they're better, but they don't loop it right back into, for that industry, that main topic of the article. If you do that, you'll start generating way more citations in these LLMs and it drives a lot of revenue.

Yeah. So speaking of citations, the one concept that we've been talking about quite a bit is the concept of being everywhere. And so if you think about organic reach 15 years ago, Google was very much the only game in town. There was no Instagram, there was no Snap, YouTube was, was very much in its nascent stages, but today that's changed quite a bit, and so everyone that was big on SEO back in the day was very— they were very tunnel vision focused, right? But nowadays you do have to think about, yeah, YouTube is the 2nd largest search engine in the world. Reddit's getting a lot of credit. I think it's the number 1 cited source on LLMs, which we'll show a chart in a little bit.

Depends which LLM, but yes.

We're talking Wikipedia. Wikipedia's number 2 on that one. Number 2, which one? What do you mean?

Which LLM?

ChatGPT.

Oh, okay.

So you have Reddit number 1, you have Wikipedia number 2, and then YouTube is number 3. And so this doesn't mean we need to try to boil the ocean necessarily, but it does mean that marketing is getting harder and we do need to look at these other channels, which is why when people come to us now, whether it's for NP Digital or for Single Grain, a lot more people are looking for Reddit. A lot more people are looking for YouTube because they know the game is changing, right? And so we're not saying don't stop doing traditional SEO, but you have to look at the other side of things. Otherwise you're going to fall behind very quickly. Um, and the good news is we did a podcast this morning. We did a recording at a studio. Um, and yes, very, even the Google search liaison, um, former liaison Danny Sullivan was saying that— Is he former? I thought he's still— Former. It said, that's what I read. It says former.

Oh, he's no longer at Google.

I don't think he's there anymore. Yeah. So what he said was, look, and we agree with this, um, but a lot of GEO, LLMO, AEO, whatever you want to call it exactly, Is a lot of the traditional SEO stuff still comes into play. Now, the problem for us is that when, when people come to us, whether it's large enterprises or mid-market, they just want the new stuff, right? And so just keep that in mind. A lot of the patterns are still the same. That's not saying it's going to be the same in the next 12 months, but it is very much the same right now.

And I know some of the stuff that we're talking about, it seems basic and boring, but when we talk to people at Inbound and we've been going to Inbound for a while, and we know a lot of companies here. This is the stuff that you guys aren't doing that really drives revenue. I get getting mentioned on Reddit isn't rocket science, but look around this room. How many of you guys focus on getting mentioned on Reddit, your products and services? Raise your hand. Be honest. Look at, look around the room. Everyone look around the room. You see how very few people are doing this? It's not—

By the way, I want, I want to talk to you all after. Any of you that do Reddit, please come up to me.

Okay. So it's less than a percent. It's these basic things that drive revenue versus a lot of companies in here. And check this out, how many of you guys use AI to create like cool videos and images? Raise your hand or have done it. Look around the room, a lot more people. And I bet you a lot of those cool videos and images aren't driving revenue. And I'm not saying that they can't drive revenue. We want to get back down to the stuff that's actually going to move the needle. And after you start growing your revenue and profitability, Then go do all the fun stuff that may not drive ROI anytime soon.

Yeah, speaking of ROI, we do have a chart that will show you the ROI of the new stuff and the old stuff in quite a bit, so stick there. It's like a, by the way, that's an open loop just like YouTube, but I'm going to keep us going here. So Neil, you've got another beautiful chart here, go ahead.

This one's pretty simple. Everyone's just like, ChatGPT's won, and yes, they may have public mindshare, especially if you're going to places like Latin America where they don't understand what, most CMOs don't know what LLMs are, but they know what ChatGPT is. But when you look at trust that people have, people trust things like Copilot. You know, if I asked you years ago, you'd be like, yeah, Slack is way better. Who uses Teams? Now, if you ask any large corporation, you're like, you're on Slack, right? They're almost all on Teams. Just because someone is later to the game doesn't mean they won't win. Google was not the first search engine. There was AltaVista and Lycos and all these other solutions and Yahoo beforehand. Facebook wasn't the first social network, right? MySpace was before them. Peter Thiel has a book, what is it called? Zero to—

Zero to One.

Zero to One. And he talks about most of the winners tend not to be the first player who enters the space. The second, the third, the fourth learn from the first. They avoid the mistakes and then they go from there. And Eric was talking about this earlier today when recording our podcast. Google has released some amazing AI technology.

Yeah, so those of you, show of hands, anybody use Google Nano Banana? Anybody? Okay, good chunk of people.

You should check it out, it's all over X.

Yeah, VO3, yeah? Okay, yeah, so Google has good products. We were talking about this morning, $115.6 billion in net income in the last trailing 12 months. They're number 1 in the world.

Yeah, in other words, he's saying that's the biggest company in the world from a profitability standpoint, and they're growing, what was the clip, was it 18% or something?

Something like that, but I think the thing is, like, we're all talking about AI, we're talking about ChatGPT, but I think we're underestimating Google. And they're not quite going to go anywhere for a while and they got some pretty smart people there.

Yeah, and when he said 118 or 115 or whatever the number is, that's profit. That's not revenue, right? That's a big number.

All right, so this one's a big one. We've been talking about this quite a bit on the podcast too. So whether it's Ahrefs, Semrush, or Sierra Interactive, we, LLMs convert a lot higher. So Ahrefs is seeing 23x, which is pretty high. Semrush, 4.4x higher than Google organic. And then Cyr says ChatGPT traffic converts at 15.9% versus Google's organic at 1.76%, which is 9 times better. And so I think this is worth talking about because again, we're not saying traditional SEO is dead, but it's time to start thinking about the surfaces because when you're using a ChatGPT, when you're using AI mode or you're using Grok or Perplexity, you're doing a lot more research and you're, at least for me, I'm using Super Whisper and I'm dictating what I'm saying. And sometimes when I dictate, Neil, it's like 5 minutes of dictating. Yeah. It's not a search query of like buy red shoes online or something like that. So we're spending a lot more time researching. Obviously if we're spending a lot more time researching, we're probably going to convert a lot higher and that's what we're seeing here. So this one is, do you have any thoughts on this one before we move into our overall thoughts?

So we see the data range a lot from SMB to enterprise. For enterprise, we don't see it convert as high of a rate as a lot of people are saying. When I say high of a rate, like some people are saying 23x or 13x, every business is going to vary. But with enterprise brands like Amazon, they already have strong loyalty where the conversions are going to be higher than normal. For SMBs, SMBs being like companies generating less than $100 million a year in revenue, we're seeing the conversions higher than typically for enterprise. And when I say higher, whether it's enterprise or SMB, we're typically seeing it convert anywhere from like 5, 6x better than organic traditional SEO traffic to on the high end, call it like in the high teens. But for most businesses, we're seeing it into the high single digits or low double digits better than traditional organic traffic.

Yeah, and what we'll talk about in a second, how we, how you can better rank on these LLM services, but I think I'll move us to the next section actually. So how we're using AI to drive real-world business results right now. So this is us nerding out on our little podcast over here. But when we think about ranking on the AI surfaces here, so structured data implementation, so you think about how would you get on rank 0 a couple years ago? Well, you have to have a good H1 tag, right? That is now, that rank 0 is now replaced by AI overviews, right? And I think it was Ahrefs that reported that AI overviews are eating what, 34% of clicks, something like that, right? You remember?

I don't know what they reported. I didn't see that.

34, 37%, something like that. And so, you know, also double E-A-T, right? We're talking expertise, authority, trustworthiness. What was the other E again? Experience, expertise, authority, and trust. Yeah, and so the— what tends to happen is whoever does— you guys do a lot of research on your blog. There's a lot of custom research that's published. That stuff tends to get cited really well because the LLMs or even Google is looking to eat up this type of content. And so, Yes, this is still a lot of it's still very much traditional SEO, but earlier we talked about citations. And so the reason if you search on ChatGPT right now, this is not to brag necessarily, but if you were to say, okay, what's the best marketing podcast? If you go to Grok or Perplexity, I've tested this a couple of times and usually Marketing School will show up, not because we're amazing or anything like that, but we've done over 3,000 episodes. So that's a lot of citations right there in terms of number of pages. But we also have a lot of reviews as well. We've been doing this for what, 9 years, I think, for the podcast?

Something like that.

Yeah, and then we're on YouTube. You have the content you publish, I have the content. So there's a lot of content on the web where it's pretty easy to see that, okay, there's some trust established. Who's often there as well? Gary Vee, because he's been doing it for a very long time as well. And so when we think about AI services, yes, traditional SEO still, but earlier we talked about search everywhere optimization, and that's where you have to think about, yes, YouTube, and YouTube gets cited top 3. Uh, Reddit, where it makes sense. I don't think you and I personally are optimizing for Reddit right now, if I'm—

We are.

You, you are, for your own brand.

Yeah.

Neil— NP Digital or Neil Patel?

Both.

How are you doing it for Neil Patel?

No comment.

Wha— oh!

How are you doing it for NP Digital?

No comment.

Okay.

We can—

you'll have to listen to Marketing School for that.

Um— No, I won't say it on that either.

Okay, I'll, I'll, I'll let you speak. Go ahead.

Uh, so when, when we look at getting citations in AI, When we look at all the data, everyone's talking about like, oh, answer engine optimization. Just create a direct answer that's short and to the point for whatever people are typing in. Everyone's doing that. Good luck trying to get, you know, tons of AI overview citations from that, or even AI citations from that. What we found is the number one way to get cited in like these AI overviews is having facts, stats, and data and unique research. If you start pushing that in to whatever way, you can survey people, you can partner with other companies. Like if you looked at a16z, Andreessen Horowitz, I believe it was on TechCrunch, they had this whole article about AI and had all these charts. A lot of the data was from SimilarWeb and they put like footnote chart from SimilarWeb, but it looked like a16z with their brand on there. The point I'm getting at, and you should of course cite your sources and ideally use multiple sources so you can, and try to add in your own opinion or your own proprietary data, but something's better than nothing. And then when you cite a lot of, or when you're creating a lot of stats and putting it out there, I'm telling you, that is what's really getting corporations cited a lot.

Yeah. And one thing we talked about this morning too, so one person posted, remember the greatest SEO play for this year? So one guy posted something about how if you create a custom GPT, those tend to rank really well right now because OpenAI has strong domain authority and Google's ranking it pretty highly. So not only can you do that, you can create custom GPTs, but we like creating tools and giving them away for free because the tools tend to get a lot of links as well. There's free calculators you can make. You can make a custom GPT for that. What else is there that you can do? You can buy a GPT plugin. So now I'm talking about distribution, right? But these plugins tend to rank highly. And if it drives— one of my friends has a— I'm not going to disclose the niche, but he's driving a ton of users right now. And this one plugin, on ChatGPT ranks number 1. And I think he bought the top 2 plugins. I think he spent, I don't know, $200,000 or $300,000, but it's driving millions in revenue for him. So that's something you can think about as well. You can go acquire a plugin that's already ranking highly. It's no different than buying a website.

So I have this chart on my X profile. I posted it 6 days ago and it says ChatGPT loves free tools. It's not in this slide deck. And I break down on neilpatel.com our referral traffic from ChatGPT. Homepage, 1.1% of AI traffic. Or ChatGPT traffic, blog post pages, 2.1%. Tools, 96.9% of my traffic from ChatGPT goes to free tools.

Wow.

Super effective.

AnswerThePublic, Ubersuggest?

This was just data for Ubersuggest.

Wow. Okay, that's a lot.

96.9%.

Yeah. Yep.

And here's the cool part. You don't have to spend millions of dollars like I did to create free tools. You can just use tools like Lovable and Cursor and all these things to help you out. You still may need a developer.

It's not that easy.

No, I know, that's what I'm saying, you still may need a developer to help you out, but what you'll find is it's cheaper to just release free tools than it is to pay for the ads. It takes a while for them to get popular, but they naturally get shared because people love anything that's free that they're used to paying for. Even if your free version, I know this is going to sound bad, sucks compared to the paid versions, a lot of people will accept free that's mediocre versus pay that's amazing.

You're not saying that about your own tools, are you?

Yes, sometimes compared to the other players, our features aren't as good, just being really upfront.

Yeah.

They've been doing it longer than us, but at least we're now releasing a lot of the AI visibility stuff. I think we'll be able to beat them because they don't have a 10-year head start and we'll be able to release a lot of them like on the 15th or 16th of this month, release all the AI visibility stuff like that Ahrefs, SEMrush, Scrunched, Profoundoo, and there's free versions of it. We're like, screw this, it's just too expensive though.

Yeah. And what Neil's talking about too, it's a nice business strategy, right? Blue ocean strategy. So it's very much a blue ocean. It's kind of new where you're trying to attack right now. It's not as competitive and that's a good thing, which is why you're trying to give it away for free, right?

Well, everyone else is trying to charge like a few hundred dollars minimum before—

A few thousand.

A few thousand, yeah. Well, most of the solutions that we saw are like $5,000 to $7,000 a year and we're like, why? This data's not that expensive. We'll release a version for free and then you can upgrade for like $20, $30 a month.

Yeah. All right, so I want to move over to social and we've got some ample time, so we'll have ample time maybe for questions too, but winning at social in the AI-driven world. So earlier, again, what I mentioned was in— looks like the graph is coming up in a second, but YouTube, Reddit's number 1, okay, for ChatGPT citations. I think it's like 40%. And then Wikipedia is up there too. Then you have YouTube. And I saw another graph today. LinkedIn seems like it's coming in number 4. And you had a graph yesterday during your talk on how I think number 2 in terms of view duration, LinkedIn is number 2.

Yeah.

After YouTube. For long-form videos, I believe it's LinkedIn, which is incredible to me because I don't watch crap on LinkedIn.

Dude, I watch a lot on LinkedIn.

You do?

Yeah.

It's a different, different behavior.

All right. So— But I also don't really use Instagram.

You don't?

No. People message me.

Why don't you view my stories all the time?

Yeah, that could be my team. No, but I do. I actually view his stories, but I probably log in like once every 3 days and like view a few friends' stories and that's it. Yeah. All right.

So look, I'm going to show you a couple of charts here. I think the next one's more interesting to me, but 54% of social browsers use social media to research products.

Okay.

That's, that's nothing really new. What's more interesting to me is that if you look at this one, and I'm going to back up in a second, what I called out earlier. So Reddit number 1 here, top 10 web domains cited by large language models in June 2025. So actually, Leo, this is Google's AI mode, AI overviews, ChatGPT and Perplexity based on 150,000 citations for 5,000 randomly selected keywords. So it's not just ChatGPT. So you got Reddit number 1, 40.1%, Wikipedia 26.3%, YouTube 23.5%. LinkedIn is not even here, so I must have been looking at something else this morning. And then you got Google at 23.3.

And so the problem with this though, just being really blunt, this is just a general good rule of thumb though, but when you're starting to crawl millions and millions and billions of queries, the data starts changing and that's the reality of it. Like if you guys use any LLM and you ask a really detailed question that can recommend a product or service and you type it in 5 times, on like different devices or right away, you'll actually start seeing differences in the responses that you get back, even if you type it in like a few minutes later on a new device incognito.

Yeah, so I want to get to in a second what you can actually do to win on social, but I think this graph's worth calling out because old marketing still works today. So you, when you look at email, and you were talking about email yesterday during your talk, so average ROI multiplier comparison, old versus new marketing channels, you got email at 38x, You got SEO at 22x, good old SEO, and then you got direct mail at 12x, which is interesting to me because a lot of people like to talk about the new things all the time, but sometimes what's old is new again.

Yeah, and we get with email, people don't just opt into your email list. They're coming because you wrote a blog post and it was SEO traffic or social traffic or paid ad traffic. But generally, like, focus on collecting more emails. It just converts really well. It still is the digital ATM.

I mean, it's very much an ATM for you because you had so much organic traffic before and everything, all that organic traffic was converting to email. And that's what, that's a big chunk of what got the business going, right?

No, it's free tools.

You register for a free tool.

Before that.

No, I'm telling you. You didn't have free tools in the beginning though.

That's true, but our email list exploded with free tools. The blog barely collected any after.

So, but your email was still a couple hundred thousand people before the free tools.

Uh, yes. The tools made it worth it well into the millions. We scrub our list like every day because I don't want to pay. I'm kind of cheap on certain things. I'm like, man, these email providers charge me with all these contacts. So I scrub the people who are inactive, but we always have, you know, multiple millions of people on our list at any given point.

Cool.

All right.

So actually, just to speak on that point for a little bit too. So Neil's got the, the, the, he had the blessing of a lot of free tool traffic, SEO traffic from that. But a lot of our mutual friends, they're spending an arm and a leg to buy emails. When I say buy emails, paid email, paid ads, sorry, for emails. And it works out well for them because if they're making $40 per email subscriber, then they'll pay $5, $10, $20 if they need to. And who do you know that crushes with emails? We know a handful of people that are doing 8, sometimes up to 9 figures just off their email list.

WPBeginner does really well with emails.

That's our, our friend Syed's company.

Uh, Smart Passive Income. Pat. Pat Flynn.

Yeah.

Huh? He does well with emails.

Yeah.

But corporations?

Yeah, corporations. We think Bulgari back in the day.

Yeah. Uh, I know Legion Athletics, they crush it with emails. They continually run, uh, promotions. It's like a supplement company. Yep. But most companies, I, I don't even know, everyone in here probably does well with email.

Yeah.

I don't know about that.

Um, Probably. How many of you guys do well with email and make good money?

That's a good chunk.

Okay, that's probably only like 10, 15%, but still, 10, 15%.

That's a good chunk. Yeah. All right, here, this one's yours.

So, but yeah, when we look at traffic sources that make up a website's traffic, it's still a lot of the old channels. And everyone wants to focus on what's new and sexy. And I'm not saying don't, I'm not saying you should ignore the future, but what I'm saying is don't let your eye off the ball. Because when you do that, you start losing a lot of revenue. And a lot of the old channels like paid advertising, you know, organic, like this is all making up a lot of revenue. So many corporations spend money on organic social and just social content. I think it's a great way to get your product discovered, but it's one of the least, one of the lowest ROI channels that we see.

Yeah. And then you got this one.

Yeah, and this is AI platform growth over time. ChatGPT is crushing it. Doesn't mean that you should ignore the rest, but you should optimize for all of them. We don't know who's going to be the winner, and people are discounting Google. I really think Google's going to take a big slice. The reason I say that, a lot of you guys are young, some of you guys are middle-aged like me, very few of you are my parents' age. If I ask my mom and dad about ChatGPT, they're like, yeah, Google gives me the ChatGPT results. ChatGPT is not Google, right? But you know what's unique about people like my mom? They bought their homes back in the day, and they invested their pensions in stocks like Apple back in the day. They made a killing. They have money. They're not technology savvy, but they are spenders, and a lot of them use platforms like Google and Facebook. You can discount them, but they drive a crapload of revenue. And a lot of the young people use TikTok and they don't— they go to their parents for money. And a lot of you may discount Snap, but it's us who is on Snap because we were the younger generation that thought it was cool. And now we've grown up and we have jobs. So don't discount Snap either.

Yeah. By the way, just a couple of practical things I think you can do from a social side. And we're doing this right now. So Remember when MrBeast said that podcasts were overpowered a couple years ago?

I do not know that, but I disagree with that.

Yeah, okay, go ahead. Why do you disagree with that?

Look at the election. Look at how many views Trump and Harris got from just being on podcasts versus the views they got on being on TV. I think podcasting is one of the most underrated channels. Salesforce has a lot of stats and data on this. The average income of a podcast listener is much higher than the average US household income.

I didn't say underpowered, I meant overpowered. So when I say MrBeast said podcasts are overpowered, it's like the stack that we're running right now. So this is being recorded right now, and I think everybody here has seen this in action, right? So this is a long-form podcast, not long-form, like half-form, 45 minutes or so. We're going to chop this up. This is going to go on our respective YouTube channels as well. Remember the speech we did at Germany last year?

Yeah.

That one got 100,000 views on it, right? We're just trying to extend the LTV of the content that we create, but you have the long form and then obviously we chop it up and there's different tools that we use for that, which I'll talk about in a second, but you start with the long form. Okay, great, this goes on YouTube. Now you chop it up for short form, it goes on LinkedIn. Maybe you have the, you have the, what, the horizontal version for LinkedIn, right? So you're chopping it up for Shorts and we're getting—

Or you can use a horizontal version even on Instagram. We do it, it still gets eyes.

And, or you take those ideas and then you, convert them to written form for your blog or for your Twitter or whatever it is exactly. But the thing is we have a whole entire funnel. We have, this is our pillar content right here and then it spreads out. But now with all the AI tools that you have available, you can do this a lot more efficiently, which you couldn't before. But this is how you can start to stack views because sometimes we'll be doing the podcast, I look at it as a slot machine. Sometimes we get lucky. Sometimes you'll talk about something, millions of views. Sometimes you and I argue about like a discount at Shake Shack, 8 million views, right? Or I'll talk about like a discount store at, uh, like, uh, remember that Laurel Piano store in Italy? 4 million views.

But I never redeemed my Shake Shack free food.

'Cause you're too slow. But, uh, but the point is, um, you never know what's going to hit. And like, the question is, does that content even convert? I'll tell you what, I was at a, I was at a business conference and 4 people came up to me saying they saw that one video, right? Did that convert to anything? No. I was about to say, no.

I don't think you drove any revenue.

But, but the key thing here is like, Neil, yesterday you were talking about the rule of 7. I think it's the rule of 21 now. You need at least 21 hits for someone to take an action because attention's so disparate nowadays and people are so distracted.

Yeah, I agree with Eric. You need multiple shots. Don't try to make anything perfect. Perfection is overrated because when you make something perfect, it still may only get one view. And a lot of my stuff that we don't try gets a lot more views. So just if you try to make everything perfect, you're going to just be pissed off when it doesn't get the views. The data that we see, the average person buys after interacting with a brand 11-ish times. I had a chart yesterday that breaks it down by price point and B2B versus B2C, but it ends up averaging out somewhere around 11-ish. But he's spot on. It's going up where people need to interact with brands more and more. So you got to be in the ChatGPTs, the social medias, and all these places. So, and most of the people in the room, check this out. How many of you guys see a good chunk of your revenue? When I say a good chunk of, let's say 5% or more of your revenue from organic social, not paid, organic, raise your hand. Look at everyone in the room. I may see 1 or 2 hands.

Maybe they just don't want to raise their hands anymore.

No, they were raising their hands earlier. I'm going to go with this. Okay. How many of you guys spend the majority of your free time when you're going to do something, Like on Instagram or TikTok and you're mindlessly scrolling before you sleep or you wake up? Don't be shy, raise your hand. Okay. That's majority of the room is my guess. All right. So this is why you got to be on these channels. They may not convert, but you learn and discover products and brands and even B2B. And I had a chart yesterday, I think 6% of the usage on Instagram for people viewing content is actually B2B. Right? Like 6% of the time they will. It may not get the most likes and shares, but people view all types of content on these platforms. So just make sure that you're on 'em. And most of the people in this room are on B2B, and you guys almost all raised your hand that you guys are on Instagram. So just be everywhere so you can go after the rule of 7 or 21 or whatever you want to call it.

By the way, one, one more thing before I move over into our predictions for the next 18 months or so. I've been talking about this to Neil, hopefully the third time it'll land, but If you guys are looking to hire great people for YouTube specifically, there's a great website called ytjobs.co. So youtubejobs.co. And this is a great site because you can actually see all these strategists, all these editors. It actually, they'll list the channels that they worked on and those channels will verify that those people have actually worked on those channels and you'll see testimonials. I see YouTube as a massive struggle for a lot of companies right now. And I don't think that's something that agents, I think Reddit for sure, agencies can help well with paid media, SEO. I think YouTube's a different beast where, you know, the brand oftentimes they're going to have to do a lot more work with the agency. So if you go to YT Jobs as CEO, they're not paying me to say this, but it's just, it's a good resource.

And YouTube is the most cited source in Google AI overviews. Sundar, the CEO of Google, when he did a talk, he said, because of all of our AI overviews and releasing AI-related content that helps people out for all those types of queries, they're seeing 10+% more in usage and searches. Cool.

All right, so we are on to our predictions for the next 18 months. So this is the next phase of marketing, and this is my favorite one over here.

So this has to be you.

Oh, I don't play video games. Yeah, so okay, I used to play a lot of online poker in college, and so on the right side over here, this is a guy named Elke. He used to be like a StarCraft player, and you can see him multi-playing a bunch of tables here for online poker. I actually think the next 12 to 24 months, maybe even longer, marketing is going to start to look like a video game. And for my session tomorrow, actually in this room at 11 AM, I'm going to be going into the nuts and bolts behind how I'm using Cursor, Claude Code, MCPs. And by the way, I'm not technical at all, nor is Neil. I just, I'll show you how we, I found a way to create like a business operating system or a battle station around this. And I feel like it's like magic. So the, what you're seeing over here, like, yes, Cursor is an IDE. Yes, a lot of people use it to code. I use it to pull reports really quickly. I no longer have to wait days, weeks, or I don't know, even sometimes up to a month to get reports that I want to get. And that allows me to move a lot faster. I'm connected to the Ahrefs MCP here, so I can pull keyword data. I can have it come up with new A/B test results based on how things are performing on Google Analytics. And so I'm able to just open up a bunch of tabs and just ask it to do things. And this is the worst it's ever going to be. And I'm so excited. Like, let's see.

You got it back.

Look how fast I'm talking about it. Let me finish. Let me finish. Let me finish. So, so, so When I want to do programmatic SEO, for example, I can publish drafts to my, to my, my, um, my, my WordPress draft folder, right? And so all that to say is that we're going to be all here. We're going to be able to move a lot faster. And now I'll slow down and turn it over to you.

Okay. Are you going to talk about this tomorrow in your presentation?

Absolutely.

Break it down to people in a way that they can understand because you lost half the room.

Okay. Thank you.

So— Hence the laughs.

Thank you.

Thank you.

So—

No, but it's really good stuff. It's just, Tomorrow he'll break it down and he'll break it down so everyone can understand in layman terms.

Yeah, I think if when we think about pulling data from HubSpot, for example, or when we think about pulling data from Gong, right? Whatever sales intelligence tool that you're using or any analytics tool that you're using, sometimes to get the data that you want, it takes forever to get there, but with using this stack of Cursor, Claude Code, and MCPs, You can get that data quickly. Not only can you get that data quickly—

Go back, what is MCPs for everyone in the audience?

Model Context Protocol. It allows your apps to talk to each other.

Thank you.

When you do this, Claude Code allows you to spin up subagents. So let's say I wanted to create an SEO employee or a reporter employee. I can do that just by dictating the Claude Code, and within 2 to 3 minutes, it creates its own prompt, and I have a complete prompt, and that's the employee. running the task that I needed to run. Then I run these tasks, these different employees in different tabs, and I'm doing 20 different things at once, right? Maybe I'm exaggerating a little bit, probably 5 to 10. But my key takeaway here is that when I say it's going to become like a video game, you're just going to be able to move a lot faster. No longer do I need to say, if Neil is on the finance team, oh, Eric, it's going to take 7 days. Screw that. I'm just going to pull up myself and I'm going to verify that it's correct with that person, and then I'm going to move on so I can make decisions faster. And then if I need to run an analysis quickly, or if I need to pump out a bunch of programmatic pages, if I want to do that, I can do that now. No longer do I need to wait, and now I have people helping me at scale.

So what this really solves is, how many of you guys work in an organization with multiple employees? Raise your hand. Okay, almost everyone. When you ask people to do stuff, A, they don't know where to get the data from. B, they don't look everywhere, even if you tell them where the data is. C, they don't know how to look at the data. And then D, they don't know how to execute on all of it for some of the basic stuff. I'm not talking about high-level stuff. I'm talking about lower-level employees. You're just getting them to do grunt work. Okay. What this does is all the apps are talking to each other and you can tell the software, you give it a prompt, say, go do this. And it'll just go pull from all this, all these places and help you with basic marketing stuff. The advanced stuff, it doesn't do a great job on. The basic stuff, it does a great job on, and you won't have people messing up as much, and it's much cheaper.

Yeah, it's— it will get there on the advanced stuff. So you can hook it in with your Meta MCP, your Google Ads MCP, you'll look at what's performing, what isn't, and then you can hook it in with a Figma MCP, look at your creatives, and it can create new variations. And now that you have Google Nano Banana, You can have image consistency, right? That's, that's, we're going a little too deep into it, but I'm just saying there's a lot more. If you can, my key takeaway for this one is if you want this type of setup, looking at Clog Code, Cursor, MCPs, and it's like, crap, I don't know how to set this up. I didn't either. And so you can go onto Upwork, literally hire, I hired a coach from China and this guy, literally, I just share my Zoom with him. Sometimes he'll take control of my screen. He doesn't steal anything. Like, don't worry, it's fine. But like, I'm Chinese too, by the way. But, um, So, and he sets it up. It probably took like 3 to 4 hours of doing a couple sessions with him. And then after that, I'm just like, hey, I want to do this. Can I do this? Can I do this?

Right?

And now I'm actually having him help my entire team. That's the next step. And then if my entire team knows how to do this, then how much more leverage are we going to gain? If your entire team knows how to do this, how much more leverage are you going to gain?

All right.

So this one's good. So this is the last slide over here and then we have time for questions. So earlier, Neil, I'll let you read off on this one because we talked about this on the pod. Um, but what's old becomes new again. So we said Snap, people shouldn't sleep on Snap because according to Triple Whale here, Snap provides a 3.87% return on ad spend. A lot of people are sleeping on Snap. What I think is like 600, 700 million users right now. I have no idea. Something like that.

I just know Snap converts really well. It doesn't drive the volume as Meta and Google, but you know, it works really well.

Yeah. So look, what's my point of calling these out is what's old When you look at the next 18 months, what's old might become new again. You look at these MCPs, all these things that we're talking about. So it's going to be exciting. I think I would just encourage everyone, maybe spend 20% of your time, 10, 20% a week looking into the stuff, playing around with it, and then encouraging your team to play around with this, and then you'll get a lot stronger.

Cool. Let's go to questions.

If you want to run personalized LinkedIn ads at scale and get these ads up in minutes instead of weeks, you got to check out Carrot. Karrot, K-A-R-R-O-T dot A-I, will allow you to create personalized ads on LinkedIn, personalized landing pages. We are also adding in different automations to spin up these ads quickly. We have different creative options as well that will integrate even cool things like Google Nano Banana. A lot of cool things coming down the pipeline. If you are serious about B2B marketing, you're serious about LinkedIn, you gotta check out Karrot, and we'll see you on the other side. Oh yeah, they got 2 microphones over here. Yeah.

Get up to the microphones in the aisle. Sorry for making you guys walk.

Hi.

Thanks for being here. We once had a, just a huge fan of our company go on Reddit and was constantly talking about how great we are. And we got a note from Reddit accusing us of self-promoting.

Yeah.

And we didn't know the guy. So you said optimize for Reddit. Can you, since they don't allow self-promotion, do you have any ideas about how to do that?

So first is most companies try to use AI responses in Reddit. We did an experiment with 100 messages that were AI created, and I don't know the exact percentage, but I know it was more than 70, but less than 80. So more than 70% of the messages got deleted. So Reddit is really good at detecting AI content. That's the first thing to know because most companies try to create their Reddit responses that way. The second thing is you actually got to participate in the community and be helpful. The third thing is, is every once in a while your company may get mentioned in a sly way, but it has to be objective. It can't be too promotional. And if you do that, you can do well on Reddit over time, but it's slow and steady. A majority can't be talking about your business or pushing it. And if you are talking about your company, you typically got to talk about other alternatives as well, and you have to do it in a neutral way. You can't be like, I use HubSpot and Salesforce. And HubSpot was 10 times better. Like they don't like content that is pushing one direction or another versus people just showcasing stats or data like, oh, HubSpot's more friendly for SMBs and startups. Here's the price, average Salesforce. It's a great tool for large enterprises, but you have to pay consultants, it costs this much, and then you keep going down that path, right? That's more objective. That kind of stuff works better on Reddit.

Thanks.

Hello. Hi, hi, Neil. Hi, Eric. Um, my name is Alex Gross. I, I work for Kiln. Um, I'm the head of AI and product. First of all, I just wanna say thank you guys. You guys have been putting out content basically consistently for over a decade. I've been following you all. Thank you so much. It's been incredible.

Thank you.

Actually, let's just give it up for these guys.

Seriously, it's incredible. Thank you.

And like, you're so candid here. It's really fun to watch your friendship blossom, um, online, but then like here in person, it's like very very genuine.

Oh, we're real friends.

Oh, I believe you.

Yeah, no, it's awesome. I think it's part of what makes your content great. Um, you both are running sizable businesses. I mean, like, and you— but you both sound like practitioners. So I'm trying to understand, like, how do you actually structure your, your operations, your week? Like, you're messing around with MCPs. Like, that's surprising to me that you would actually be doing that. Why aren't you just passing that off to some dude in China, right? Like, so I'm kind of curious, like, Both of you, Neil and Eric, like how are you actually structuring your time? Are you spending 5% of your time on content and this is actually the highest leverage activity and then you're dumping it off to your team or how's that work?

Yeah, so for me, and I definitely want to get Neil's day too because it's really interesting how he structures it. Mondays are a mess for me, it's all meetings, but then the rest of the time, especially in the evenings because I'm not married, don't have kids like Neil does, so I have a lot of time on my hands. And so That's where I spend my time on MCPs and AI in general. And I think, so how I've defined it before on the podcast was that you have fanatics in your company, you have the converts. So the fanatics are crazy about AI and they're on everything. The converts are, they need direction, they need, they're down for it, but they don't know where to start. And so the fanatics will kind of bring them along. And then you have the anchors. The anchors are resisting. They're like, oh man, I don't want AI to take my job type of thing. And so for me, I feel like it's my job as a fanatic and the leader to have it come from the top down. If it's not coming from me, We were talking about it this morning, even Neil's playing around with this stuff quite a bit. If it's not coming from us, the most incentivized people to drive it, it's not going to happen. And we see this as pretty existential for us. And I don't know, I'm personally excited about it, but I'll transfer it to you, Neil, because Neil's structure is interesting because he doesn't have any direct reports.

Yeah, I don't have direct reports. I'm a terrible manager, so I never manage. I'm a terrible CEO, so I'm never the CEO. I just go around the business. And look for the biggest opportunity of the week that needs to be solved that would help us make more revenue and profit. And I just focus on that and I go disrupt and make a mess and solve a problem and probably irritate or piss people off.

Oh, probably for sure. Yeah.

And then I go to the next one the next week and I just keep doing that every week.

Yeah.

And I just hop around.

I think the big thing is, um, we want to focus on the, the, the new things, the bold things and the, the big things. Um, they said big, bold, new. Those are the things that we want to focus on.

I'm actually different. I just focus on what's going to make me the most money.

I like that. Um, we got time for one more.

Yeah.

Friends watch other friends' reels.

You know, I know. Well, I do watch his every once in a while, but not necessarily the reels. I like his stories.

Cool. Um, Eric, you're talking about how some of this stuff is, you're saying existential. Uh, you're clearly playing around with this stuff, getting deeper on it. We've also heard in this conference in multiple sessions about how leaders and companies are not seeing good return from AI spend, AI investment, both in time and in resources and cost. I'm curious how you balance these things because, Neil, you're also talking about not missing the mark and focusing also on traditional marketing things that are working. How do you balance that difference there?

Yeah, so for me, and then we talked about it this morning too, I want to hear Neil's answer, but I do focus on what's already working. So what's working for us is still very much on the traditional SEO side. So how do we use AI to enhance that? Whether it's on programmatic or, you know, other, like using the Ahrefs MCP, I can come up with new pillars or content we should be consolidating, deleting, what's cannibalizing. So that's what I'll use those MCPs for. So I use it on what's already working and what's already driving ROI. So for me personally, It's driving ROI. I don't try it on VO3 or Google Nano Banana. Those are fun to play with and I play with those things, but Neil actually made a good point this morning because you actually asked a question yesterday to the crowd on who's driving ROI from AI and almost no one raised their hand.

So the way we look at it is most people don't focus on KPIs internally. So what are you— everyone has a job. In your job, what do you do that actually makes the company more money or saves them money? That's, it's really simple. Then you look at AI and how can you actually use AI to do those things better? And you should see more revenue growth or savings. And what we like is a lot of this is experimentation because it's so new. Take a portion of your budget, very small, tiny percentage. It's in the single-digit percentiles as a business, and you just try a lot of new random stuff. Some of it will work, some of it won't, and you keep the majority for traditional stuff that is producing ROI. And then you adjust over time.

Cool.

So everyone, thank you so much. Neil and I are getting kicked out, but we're going to probably hang out over here right after. But thank you again.

Take care, everyone.