- I don't look at OpenAI as open AI. I look at OpenAI as Microsoft. It's mainly funded - Fair. - by Microsoft and controlled, so I look at OpenAI as Bing versus Bard. (calm contemporary music) - We're going to do some big debates around search, around marketing, and we have a very special guest, we've got Neil Patel who's the co-founder of NP Digital. He is general marketer and man about the internet, and we're going to talk marketing today, we're going to talk about AI, we're going to talk about a whole host of things. Neil, welcome to the show. - Yeah, thanks for having me. - Neil, how quickly you're going to change your LinkedIn headline to man about the internet. (all laugh) - I don't know, I've seen Neil Patel on the internet a lot. Man about the internet seemed like the right intro. - I wish someone would describe me as man about the internet, I love that. - Kieran, I know you had a couple of things that you for sure wanted to kick off with, so I want to hand it to you to get the debate started today. - Neil, you're one of the, like renowned for many reasons but search is definitely one of the things that you have mastered, and you've done a lot of great takes, so we were watching some of your short-form videos around ChatGPT, the impact it has in SU and advertising, so we have had the spicy takes for like three months right now that GPT is a natural language layer on top of the internet, and is showing us that like chat, the chat experience is that experience that most users predominantly are starting to fall in love with, and it moves everything back a click, so it moves software back a click through the new ChatGPT app store, it moves search back a click, and we think it's very, very disruptive for search and advertising. It would be good to start to get your takes. I think you actually are maybe on the opposite end of the spectrum, and then we can try to like, - I am, yeah. - convince each other that-- - [Kip] (chuckles) Who's right. - Yeah, who's right. - So I do see AI as being a huge part of the future. So let's go back a little bit. When most people think about, and let's focus actually on ChatGPT, forget even just AI, let's actually just ChatGPT, I know, because OpenAI is taking most of the glory right now when it comes to the press and everything related to AI versus Bard or anything else out there. Right now, ChatGPT, the way they end up spitting up answers to you is by scraping the web. In essence, they're crawling the web, gathering all this information, and then using it for output, right? The input, you know, like if you think about back in the day, there was these things called article spinners. Article spinners, you shove in a article, it shoves you out a output based on the input. You no longer have to put in the input because it's just scraping the web. The issue though, if you think about it, is Google's been around for more than 20 years. A lot of the search engines have been around for ages. Any time you guys do a search, would you guys agree with me, not any time, but a lot of times when you do a search, there's still misinformation out there. You get inaccurate information. You can type in anything. - Inaccurate and low quality, both for sure. - Correct, and they've been trying to solve these problems for, call it 20-plus years. I'm making up the timeframe, but it's been way more than 10 years since they've been trying to solve it. Right, I know people both at Microsoft and Google and engineers literally trying to fight misinformation, this is why Eric Schmidt back in the day, the ex-CEO of Google, used to talk about, brands are how you sort out the people from the cesspool, I'm butchering his quote, and what he would talk about is, brands are less likely to put out misinformation. It's not always true, but it's more realistic that a brand is going to do their fact-checking and they're going to put out less misinformation. Again, not always, but in theory, that's what it's supposed to be like. So if your inputs are off, your outputs are going to be off, because if they haven't been able to figure out what's misinformation and that's being inputted into AI, you're always going to get out, or not always, for a portion of the queries and the responses, depending on what you're looking for the AI to spit out from a answer perspective, you're going to get misinformation as well, inaccurate, wrong, whatever it may be. Now, that's just from a article standpoint. I think there's a huge, huge, huge, actually I was looking at a McKinsey study, the real value in AI from a revenue standpoint from businesses is a lot of things like business processes and efficiencies, crunching analytics from a marketing standpoint, like when you talk to a lot of the Fortune 1000, they're less concerned about creating content. Content's already cheap. The consumer though is always about like, "Oh, like you need these ChatGPT "or need these tools to spit out a contract "or spit out a answer, write a song or a poem, "this is amazing," and all that stuff, yeah, I think it's going to do wonders for. But going back to the question about disrupting advertising, yeah, I think a portion of the queries are going to end up disrupting advertising for search, but not for the majority. because the real revenue generation from Google and these search engines when people are typing is for a lot of the transactional keywords. It's not, how does Google's algorithm work. You know, when you look at back in the day, when you think about the Knowledge Graph, people were like, "Oh, that's going to disrupt search "because you're just getting the answers and less people are going to click," but if you look at what Danny Sullivan from Google has said, since every year Google's been around, there's actually have driven more clicks to websites. I think Google's going to answer more queries, just like it did with Knowledge Graph or what's the weather in Las Vegas, Nevada? I just don't see it really disrupting everything, because the majority of the ad dollars, like we manage billions of dollars in ad spend for companies, a majority is transactional keywords, like, if you want to talk about search specifically, right? This is not social, just search, it really is transactional-based keywords where the majority of the revenue's being spent on ad dollars. - I agree with you on transactional keywords. There's a couple of things here. There's AI which we should talk about, and there's also, Kieran, you kind of alluded to kind of this evolution of a chat user experience and a preference that's possibly going to be driven by individuals and consumers to have a more chat user experience versus a traditional graphical user interface or in Google's term, like just a core search box. If we move to a more chat-centric interface, isn't the ad model, like AdWords breaks a lot in that interface, like the actual experience of how you discover ads does break though, doesn't it? - It does, but for a lot of stuff, like if you're looking for cheap laptop, you're still going to just see ads. You don't want ChatGPT to spit out a answer or Bard and like, here's the best cheap laptop. You're like, "Yeah, I don't know if you're getting paid "for this or what's happening, "let me do my own research and pick what laptop I want." Because what works for you may not work for me. I do agree, it's going to change how search is and how we function with it and it is going to be a more chat-based type of model, because what OpenAI did, the most amazing part is, it's really good at understanding what you're looking for and giving you a output, and that's the amazing part about it. Yes, there's a lot more magic to it, I just don't see it disrupting too much of the advertising. Yes, you can say some of the queries, and there will be a portion that are affected, and I don't know how they'll monetize, but they can either do it like Twitter's doing, hey, we're going to charge you for a verification, we're going to charge you to start using some of this after a certain amount. The Instagram numbers were really crazy on the blue tick mark, what was it, like 40-something million people in the first 24 hours is what the user-- - Yeah, it was a lot of money, it was hundreds of millions of dollars in revenue in the first day, right? - Yeah, I think they said 660 million. I didn't see a article from Facebook verifying that, but that's what people were doing with the back-of-the-napkin math. - That's the rumor. - So I do think that Google will figure out other ways to monetize, and I do think this type of model will actually cause us to use search more, so in theory you can say, yes, a portion of the clicks will go away, a portion of the ad revenue will go away, but I actually think this will create more usage of Google and other search engines, which will cause the total volume to go up, just like what Danny Sullivan broke down in which, yeah, you may end up losing things from like Knowledge Graph or Bard or ChatGPT, but if the number of people using Google on a daily basis continually increases, there's still traffic and volume to be had, and potentially more than there was from the previous year. - Well I buy your argument that AI's going to actually increase adoption of people just like engaging online, looking and searching for things. The point that you're making that I want everybody watching to understand is the point you're actually making is, look, I believe especially, I'm paraphrasing you, I believe especially for transactional keywords, that choice matters in search results, that having a lot of choice and a lot of options for those keywords, so I could do my own refinement and research, matters a lot. I don't know if that's going to be true or not, but I think that's the point you're making, versus getting one answer from a AI chatbot, you're like, "Hey, I want to see 10 different cheap laptops "and I want to pick the cheap laptop "that has these certain specs "and things that I particularly want." Is that what you're pushing on? - That's what I'm pushing on. So like, let's say if you asked it to create a contract for you, right? It's not going to be perfect. I think we can all agree, if you have ChatGPT create a contract for you, you may probably want to send it to a lawyer first, and I'm not saying just for today, I'm also saying this for two, three, four, five years from now as well, because again, a lot of the inputs are off, and you may need things really customized to your liking, so yes, you may be able to save some money on legal expenses by having a lawyer review versus creating from scratch, but there's ad opportunities. Hey, would you like a lawyer to review this contract that you just created through Bard or ChatGPT? Click a button here and we'll connect you with a lawyer. I think they'll also start having new monetization methods that Google will end up making money from. - You know, we kind of break this down into three core components which is navigational queries, informational queries, and transactional queries, and navigational queries are like, I'm just going to type the domain because I want to navigate to that website. We can kind of do away with those and just think about informational and transactional. I think an informational, I think users are maybe lazier than you are describing them, Neil, in that I don't think the average user cares as much about the misinformation as they care about the ease of use, and I think the one thing that chat has shown is that users will always choose like ease of use over anything else, and that's why I think it's the fastest growing app of all time. Also because it just is like a new thing, right, and people love, like Clubhouse is one of the fastest growing things of all time and it turned out it was kind of garbage in the end, but I do think there's something in that which is like, when I think of the ChatGPT experience, it provides a concise, formatted answer and I don't have to do anything else because it just like gives me, it's like my assistant, it goes through the blue links, provides me the answer in a way that I can easily understand that, and then I don't have to do any other work, and there's something in that ease of use where users will always gravitate towards the easier thing, but on their transactional, I think that's a really good point on transactional which is like, most of the money on search and advertising is made in transactional queries which is like nearest to the buying step. However, I do think with the new ChatGPT store, it can start to like cannibalize those very, very quickly, because now I can actually say, "Hey, like plan out my trip to Barcelona. "Choose me the best things, "the best restaurants applicable to me, just go book them." I actually don't care how it's booking them, I don't care if it's using OpenTable, I don't care if it's using a local provider, I don't care how it actually gets done because I'm a user and I'm lazy, right? I just care that it's done, and then the other thing I, book me a flight between 400 and $450, in the background it can go execute that in any of these flight providers, I actually don't care how it does it, I just care that it gets done within the kind of price range that I've given with the seat that I want, with the time that I want, and I do think that there is something there that could really start to cannibalize search, because now I don't need to actually go and trawl through these links, and like search is predicated on the fact that you have to query and you have to like go through the blue links yourself so we can interrupt you and try to get you to click on the advertisement link, and I think that's the, either the bull case for a chat versus search, or the bear case for search. - Okay, let's go back a little bit. You mentioned a great example. Hey, I want to book a trip, let's say to Paris, pick me the best spots to visit, here's how many days, I want to stay in a hotel in this price range and I want a flight in this price range and this is the city I'm leaving from, here's the times I like to fly, here are my specifications on the seat, if I need lay-flat bed, if I'm just okay with the window seat, whatever it may be, right? You're giving the specifications. Right now when you do the searches, people are clicking on the blue links and they're making money. To run ChatGPT or run Bard, it is costing billions of dollars. There's no way Microsoft or Google, or let's look at just Microsoft, Microsoft put, what, 13 billion or something like that into the company? They didn't put that money for no reason. I don't know if it's all cache or servers, but this is just a expensive process to run, they're not going to keep burning this money without making money. They're a business. - But they'll take a percentage of the transaction. So they can pay, they could take a subscription. - Bingo, you got it right. - But what I'm saying is, that changes the way that thing is priced, because now it's priced on just, I'm the user, I only care that it gets done, but I don't need to go and actually go research myself, I just care, you, the AI assistant, go take my parameters and go, does it, so it pushes the search experience out of the user's purview and puts chat as the only kind of layer in between you and the thing getting done. - But those still make money. For them, it doesn't matter if they're charging per click or charging per transaction. If you look at Google and their history, they've tested the model of charging per transactions before in various different industries. They've also looked to cannibalize certain markets like mortgage and flights, and they've tested some of these things. - Why would Google win that? If that experience is like, not what Google has excelled at in the past which is the blue links, and it's a chat experience, I wonder why Google would win that versus OpenAI just dominating that and crushing Google in that experience. - I don't look at OpenAI as OpenAI. I look at OpenAI as Microsoft. (chuckling) because the reality is is, it's mainly funded by Microsoft and controlled, so I look at OpenAI as Bing versus Bard, right? The difference though with Bard and Bing is, Bard has a really good dataset of the whole web. Google in theory has the biggest index. If you look at the most popular social network, it's not Myspace. If you look at, you know, a lot of the early adopters, like the most popular search engine, it's not, I don't know which one was the first one but it's not AltaVista or Lycos or, you know, Yahoo or whatever it may be. A lot of the winners were latecomers, and I still think we're in the early innings of AI, and I do agree, it's going to change from that aspect of how you search. What I'm saying it's not going to change is, they're still going to make money. They're going to figure out how to monetize, and as a marketer, we don't really care if someone clicks on our website, all we care about is, are we generating the revenue? And it's just a number. We spent X on that click, here's the conversion rate, and here's how much revenue we made and here's our costs for Y dollars. On the same aspect, it's like, this is now turning into like affiliate marketing. We spent X dollars, they gave us a customer and they dealt with all of it for us, even the conversion side, and here was our profit. As long as the numbers work as marketers, do we really care? - So there're a couple things in this that I wanted to punch on. The first is, you're right, first movers don't always win, right? Google was the ninth search engine, I believe, that got started and really won and dominated the market share. The point you all are making in your travel booking example, as some dude who's just sitting listening to this debate, the point you're actually making is that right now, businesses bid on keywords to have their ads appear, and what you're actually saying is, no, they're going to bid directly on the sale, that hey, you know, I'm looking for a flight between 400 and $450, you have to decide if you want to offer this person a flight for this price, right? There's going to be different people who are going to, you know, different airlines are going to have to decide, like, am I willing to bid in at a certain price, because the chatbot is going to just give them whatever the best price is at the flight. That is very different than what's happening today. This is like, - No, it says, I-- - I'm bidding and selling this fixed cost thing. It's putting variable pricing into a lot of companies, and it's going to disintermediate a lot of sellers out there too. - But you know this better than anyone else at HubSpot, it's almost the same model right now. You are right, we are bidding on a keyword, but we're all back ending it out to a cost per acquisition anyways, right? We're running our campaigns and tracking goals and conversions, it's, yeah there's a click, but we're really looking at, what is our cost per conversion, whether it's at Zapier or HubSpot, it doesn't matter the company, we're all looking at the same thing. Here's what it costs to acquire a customer and here's our LTV and here's the profit in the long run. - What I'm saying is that, one, it forces you to be great at economics and unit economics and a lot of marketers aren't, so that's a big thing for everybody watching is like, wow, if you're not close to the economics of your performance marketing, you're going to have to get much closer because that, how those transaction happen is going to change a lot, and two, like in your flight example, it does kind of kill like Expedia and a lot of the middleware companies that are trying to basically compete on price arbitrage and like bundling, because the chatbot will do that and then the airlines can just offer, the airlines, hotels, what have you, can offer that direct, so it's going to be fascinating on the I think kind of core like transactional search side of things. - And in my opinion, it's progression. Expedia, KAYAK, I use KAYAK a lot. - I do agree. - So why do I need a KAYAK if I'm already searching on Google and then I'm searching again on KAYAK, it's like a search going to another search engine as a user, I just want my ticket. Just give it to me, and if that means Google makes more money, I know this sounds bad, people are going to say, "Oh, they're going to be bigger, more monopolistic" and yada, yada, yada, but as a user, it's more of a easier experience. I don't care. - Exactly. That's what, yeah. - That's Kieran's argument. - And Booking.com, they are like the, you know, poor man's example of basic AI, right? They are just the ability to reaggregate the aggregator. Like they just reaggregate Google to make it easier to sort through, and I think chat is the ultimate aggregator of aggregators, because now it just actually encapsulates all these things and filters it all through for you and just gives you the thing you want. The thing I actually don't know, like to your point Kip, I hadn't thought of that, which is, the auction model moves to some sort of price sensitivity in the background, so you actually, in, how do you get chosen among all of those flight providers, why are you the one that gets chosen? And that's the part I've actually been trying to think through because today, you get chosen on an ads auction model where you can like bid on the keywords and actually win that auction, whereas, when you bypass the need for the keywords, what are you, what does the auction model look like, and maybe it is, hey, every time that person has a certain set of criteria, that I can change something in the background to be the best fit for that criteria. - Yeah, you're basically bidding on what you're willing to sell your product for and trying to average it out over the long run, across a bunch of different queries, right? - And it becomes more efficient for the business and the end user. Here's a great example, let's go back to this travel one. If you're looking for going to LA to Japan and you want a window seat and you want economy with extra legroom and you want a TV, you know, on your seat, and you want these food options, I making this up, yes, there's a lot of airlines that probably fly that route, but what happens if the other airlines don't have that seat available but they have different seats? They don't have to pay for that cost. Right now, they'll pay for that click, going to someone's website and their conversions won't necessarily be what they're looking for because like, oh, we don't have that option. So as a business, this is easier. I'm only paying for what I have, I'm saving a lot of money and time getting a lot of clicks that aren't going to convert in the first place. - Well look, that's the key point here. The more you move the bidding and the ad experience closer to the product and the product cost, you're going to increase your quality of customer, right, and you're going to have lower risk that you buy all these potential customers that actually don't want your product. - And I think it's going to go one step further. I think your transaction's just going to happen on Bing, or it's going to happen on Google, and you're not going to even go to the airline to transact and they'll be like, cool, I don't know, Google Pay, is it Google Pay? I don't know what Google calls theirs. But Google Pay, - Yeah, it's Google Pay. - or one of them, right, or Apple Pay, like it literally is going to be, oh cool, click, you're done, they already got your information, and you're off and to the races. - This is the whole thing about the ChatGPT app store that's wild which is like, why do I go to any of these software provider's websites or even their apps when I can just like interface with it through the, why do I go to Instacart when I can just like ask ChatGPT to use Instacart for me? Why do I go to any of these products when I can ask ChatGPT to use it for me? Now I've just commoditized like every user face and app, because I don't actually care about any of these apps, I just care that they exist as a plugin, and I don't want to know about them, I just want to know my thing gets executed, which actually completely, like, OpenAI is now like the aggregator of all these apps and it commoditizes them, and so it now actually has a huge power over how to wield that in terms of your app getting picked over another app. - And a lot of businesses are scared, but at the end of the day, what's going to make a business win is you do what's best for the consumer. Like I don't look at Expedia as doing what's best for a consumer. I just think they're a middleman that just found a inefficiency in a marketplace, and the first result being Google where people go to or Bing or wherever, they need to fix that, right? Most people that I know, because we've done a lot of work for different travel companies, it's not like people are like, let me just go to Delta Airlines and book, because not every airline flies specific routes. Sure, if you already know what airlines fly specific routes and you only leave from one destination, all right, you're going to end up doing that, but for a lot of people, they're just looking at like, oh, how do I get to here from this location? And if you end up as a business, build the best experience for the customer, EX like Amazon, right, if you look at the behavior of young people, not like my parents' generation but really young people in their 20s and 30s, they don't go to the grocery store to buy toilet paper. They just go to Amazon, be like, toilet paper, click, subscribe, comes to their house, it's Prime, it's more convenient, they don't understand why parents go to the grocery store to buy toilet paper. Like this is inefficient and backwards. But they obsess about the customer, they provide the best experience. I believe it's going to come down to building the best product or service for the end user, and having a really strong brand. because like if you think about Nike, there's a lot of shoes that build similar quality shoes, people love the brand. If you look at, you know, let's go back in the day, I think the HubSpot founders invested in David Cancel's company, I think it was called Drift, was it Drift? - [Kip] Yeah, it was Drift. - It was Drift, all right? So I don't know if most people know this, you can do a lot of the same stuff you can do in Drift and Entercom on HubSpot for cheaper, right? So what do people ended up doing? I know at some of our companies, we just use your solution, and you could say, "Oh, you know, "Drift does this or Entercom does this," I'm like, "Yeah, HubSpot does it and as a end user, "maybe Drift has some other features, I don't know, "but for what we needed, "you guys did it all and it was cheaper. "It was more convenient." - Bundle and cheaper. Bundling matters, right, to the end user in all of this. - Yes, but not just bundling. It's all one place, so super convenient, right, to be even more specific, - Yes, convenience, exactly. - and it was cheaper. They could offer 20 other features that you don't have, but if I don't need those features, it doesn't matter. It's about delighting the customer. And that's what I was getting at, like with building a brand, if you have a strong brand and you're delighting them and you're delighting them like 98% or enough of the users, those edge cases for those 2% doesn't matter. With Zapier, you know how many competitors are out there that do exactly what Zapier does, some are even cheaper. We still use Zapier, why? It's just a brand and it works and we've been using it forever and it's just known for doing this. Like, there's something to be said for having a really easy-to-use product, clicking some buttons like I have on Zapier, and it just works. - Yeah, convenience and price. - And that's where businesses will need to go. - Yeah, convenience and price are really hard to beat.