The Franchise Edge | Christian Ward

This podcast episode explores the impact of generative AI on digital marketing, particularly for franchises. Christian Ward, Chief Data Officer at Yext, emphasizes the importance of well-structured data for businesses. He believes generative AI can revolutionize marketing by automating tasks and personalizing content at scale. While acknowledging potential challenges like data quality and bias, Ward sees generative AI as a powerful tool to improve marketing outcomes.

We hope you enjoy the show!

Transcript

0:00

(energetic music)- Hey, welcome back to "The Franchise Edge" podcast, brought to you by Scorpion. I'm your host, Jamie Adams, and this is the first episode of 2024. So welcome back. Hopefully you've missed us. Got a lot of great guests this year. We took a little bit of a break over the holiday, but I wanted to start the year quick. Today I'm going to be talking with Christian Ward. Christian is the chief data officer at a company called Yext. Most of you are probably familiar with Yext. They started a long time ago in the call tracking space. Not many people knew that, but eventually kind of grew into things like online listings, reputation management, things of that nature. But it's certainly been around the local and the franchise space for a really long time. Christian's got a really interesting perspective. His background is really in big data. He does a lot of research and a lot of kind of forward thinking and vision casting around things like data privacy. But he's also gotten really in tune with generative AI and how that's impacting digital marketing and marketing in general. And we spend a lot of time today talking about generative AI, specifically around digital marketing for franchise businesses. A lot of our time spent talking about service based businesses today. So if you're a service based business, I think you're going to get a lot out of this conversation. We talk about some things around data privacy as well, some things to be thinking about as regulations around cookies and data privacy become more of a thing here in the US. But anyway, I think you're going to enjoy the conversation. I really, really got a lot out of it. I appreciate Christian's time. We're going to put information in the show notes on how you can get in touch with Christian if you'd like to. And also, we're going to reference a really interesting academia paper written by a couple of Princeton students around this idea around generative engine optimization. Think of it as SEO, but for generative search, things like ChatGPT, Google, Bard, et cetera. So without further ado, thanks again for joining, and I will let you get to the episode. See you soon. Welcome back to "The Franchise Edge" podcast. Today I'm joined by my good friend of almost a decade, Mr. Christian Ward, the chief data officer at Yext. Christian, thanks so much for being here today, man. How are you?- I'm great, Jamie. Good to see you.- Yeah, likewise, likewise. Now it is almost a decade for you at Yext, yeah?- Yes. I took about a year off to write a book with my brother back in 2018.- I remember that.- But then I came back, so the actual at Yext is a decade.- Okay, got it. I'm a decade at Scorpion in March, so getting there. I know, it's crazy, time flies. Time flies. It's been really fun watching the trajectory of Yext over the years, too. I mean, I've known you and Howard and that group from the time that it was Felix. And then watching it evolve. And I remember one of the first conversations or meetings I had with you at Yext. We were talking about structured data, and I think this was like the very early impetus of what eventually became kind of your answers and your knowledge graph product. But I remember talking about that before it was even like a real product. So tell me a little bit about what you're up to these days. I want to talk to you a lot about. You've got a great perspective. You've always had a great knowledge around data, particularly as it relates to marketing and local marketing, which is a big problem that franchises are trying to solve and opportunities are trying to take advantage of. But really want to talk about your perspective on AI and generative AI specifically and how you believe it's going to influence marketing. But just give us a little bit of update on what's going on at Yext and what you've been up to in the last year, year and a half.- Sure. So as you mentioned, I'm chief data officer at Yext, and in that role, my job is actually to work with clients. So partners, data partners. So the Googles and the Bings, now the new fun AI companies, but also our major enterprise clients. So some of the major franchise organizations around the world. And when I meet with them, what's kind of amazing about my job is they all are very focused on what they're doing. But if they could all be in one meeting with me at the same time, you would realize they all have the exact same problems. And it doesn't matter if it's like HVAC, major chain, or it's quite frankly a hospital system or the World Health Organization. We work with all these. And it always comes back to that problem that you said when we first discussed. It's I need to structure my knowledge, the data, so that the customer journey is easier. That continues to be our focus. I've been an entrepreneur and run small businesses myself. Started my first company when I was 23, and I've been doing, I had three different companies. I like where I'm at now because it lets me work from the small business all the way up to the largest to really try and figure out what's the best way for them to structure that knowledge and get it out there so it's useful and it's driving engagement. And as you said, I think where we're going with this AI capabilities is just going to open up all these things we've always wanted to do as business owners and just really couldn't do it. We couldn't scale ourselves.- Yeah. So maybe just because I know the audience for this is mainly franchise businesses, and we also have a lot of independent small businesses, typically, that are service based, service oriented. I always tell people the majority of Scorpion's customers are what I classify as falling into the oh shit categories. Since this is our podcast, I guess I can say that word. But what I mean is like the consumer journey kind of starts with an oh shit moment. It's, oh my God, my plumbing is broken and I need a plumber. Or oh God, I walked into my house and it's 100 degrees outside and my HVAC system is not working. Or hey, I had a car accident, I need a lawyer, hey, my tooth is killing me and I need a dentist. So it's these kind of undesirable consumer categories. But again, that's the audience. But you use the word structured data, and I want to make sure that we kind of tiptoe in the definitions here. So can you give just a very elementary definition of structured data? Just thinking about who our typical audience is and why they should be concerned about it?- Sure. When we talk about data, we'll actually talk about the two categories because I think it really helps to make it very clear. There's unstructured, which tends to be like this. It's like a dialogue, or it could be a voice call, but often it's conversation in some form, some human conversation or a book that's got a lot of stuff. When you look at unstructured, what's amazing is inside the unstructured is this awesome structure waiting to come out. So when you say, this is our audience type, they're this type of business, and they are in these regions and these categories, all those things, your brain as a human automatically structures and goes, oh, when he said bank, he's actually talking about a river bank because he's a fly fisherman guide, he's not talking about a financial bank. And what your brain has done has added this structure of context. And it happens for us based in our own sort of neural network, it happens very quickly. We don't think about it. But for machines, they don't have that. And so you have to kind of feed it to them. And the more structure you feed to them, the more they can actually get you to the right answer. And so, for example, if I say that I have a problem with my heater, if I say that, that may not be enough structure or context for a search engine, because they might be, well, it could be your car heater, it could be your pool heater. So there's other elements. So structure is all about just trying to find, if you think of it almost as connecting the dots, where each dot is a particular entity or like a person placer thing, and then the relationship is there. That's why we always talk about structure. And a lot of times we talk about something called a knowledge graph, which is basically how Google knows what you're looking for. They piece it together in that context. And so at Yext, we've been working in structured data, and quite frankly, my whole career is in structured data. It was originally on Wall Street and financial services, but then I really got into sort of the business side of every potential business and what structure could look like. So that's really the difference.- Okay. As it relates to just kind of operationally. Let's take it down to what people know Yext for specifically. Early foray into structured data and where you guys really drove a lot of innovation was around what now is kind of a common thing. But it's like business listings, right? So you guys kind of created this really robust platform that pushed structured data like a business name, a business phone number, their website, their hours of operation, what category they were. You guys built this platform of data that would then distribute that data to Google Business Profile. Or back then, I guess it was Google My Business, your Bing business profile, your Facebook business profile. So that was kind of an early innovation, right, that Yext brought to the marketplace, or at least made mainstream and made it more accessible to the average local business. So that's a fair kind of descriptor of a very basic category of structured data.- Yeah, absolutely. So things like an address in Ireland look very different than an address here in the States. There's different structure, their codings are different. And so getting all that to work in any jurisdiction for any business was the beginnings of Yext, I would say. And that's very well said. I didn't found it. I wasn't at the very beginning, but it was early enough. But what I'll say is that issue has only gotten bigger. So since 2015, I want to say, the number of data fields that Google and Bing and Uber and every platform out there, right? The number of data fields that they want structured and fed to them about each of your businesses has gone up 700%. So that's kind of mind blowing, which is, let's say in 2015, they said you can fill out in an old GMB, Google My Business profile, you might be able to fill out 100 fields. Now they're asking for 700. They want to know every holiday hour. They want to know if there are certain service areas you won't service. They want to know if you do kitchen remodeling in addition to emergency plumbing. It's all of these things where. And for a lot of your audience, yes, it might be we meet through oh shit, but the reality is, through a great experience, then you become the trusted advisor, you become the person they go to for all of those subsequent or even additional things. And I think especially that happens a ton with electricians. HVAC, which is you then take over the servicing contract. And so that's the beauty of it. But you have to first tell Google and others everything about you because you want to show up. Because how a human asks the question, we've all lived for 25 years living in this keyword, weird keyword speak, where people are like, HVAC, broken, cold, Texas. That's not English. That's horrible, that sort of process. I think we're being freed from that now. But you still have to have the structure so that the AI and the advanced search systems can understand. Oh, I know what you meant. What you're looking for is Paul. Paul's great. Yeah, he's right in your area. Let me get you Paul. That's the opportunity ahead. But that structured data still drives all of it, because if you're Google and these other platforms, crawling the web doesn't actually give you the knowledge to answer perfectly. Crawling the web gives you a way to understand and to talk to a human meaning as an AI. But you still don't know if Paul is working tomorrow or not. Only Paul knows if he's working tomorrow or not. And so that data is what you feed into Yext and scorpion, our platforms, to basically get that information out.- You kind of tipped your toe into the water of where I really want to take the conversation, which is around generative AI. So we all know it's crazy to think about just two, three years ago, people were using words like AI and machine learning as part of some obscure way to describe what is probably more of an algorithmic driven approach to product and optimization of certain activities, be it ad management or things of that nature. But then open AI comes along a year ago and all of a sudden it's like AI is real, right? Like it is in the hands of every consumer. And it's quite simple to use. It went from this very obscure, I don't understand it, to, wow, I can actually use this every single day of my life. And it just makes sense. I think we've got a perspective at Scorpion of our customers from a marketing perspective, they lean heavily into search engines today. Again, because most people don't think about a plumber or an HVAC. That zero moment of truth is still very real to the types of businesses they need to be present when the consumer is in that moment of need. So search is still kind of the place where that happens. What is your perspective on how search is going to change? How is generative AI going to change the traditional search experience that people go to today at Google? Or maybe they're going to Bing and maybe they're using like copilot or Bing chat, maybe. Yeah, exactly. But I'm just curious, what do you see this play out in the next twelve months to three to five years?- Yeah. I always end up drawing kind of the same continuum when I'm meeting with clients, which is, I think the world sort of splits and I'll explain exactly what I mean, which is if you were to draw a line left to right, and you sort of said, this is a branded question and this is an unbranded question, but this is an objective question or a subjective question, you end up with four quadrants where there's basically unbranded objective questions. What is HVAC? I didn't ask about a brand, I'm just trying to learn. Those questions are going to be really well answered by AI. And similar to that's what SGE search generative experience and Google is really focused on right now is those sort of. I didn't ask you about Nike or Paul the plumber. I actually asked you about a topic. The next quadrant is a subjective question about an unbranded thing. So that would be like, who is a good HVAC provider? Who's a good attorney? That's unbranded. You still don't know who you're talking about, but you're looking for an opinion. Google and all centralized search systems will still dominate there. Why? Because you're really kind of looking for the aggregation of a lot of thoughts. You're looking for like rankings, or you're looking for websites that have earned the respect that they've been around for a long time. There's other signals and it's a different kind of thing. And then you move up to subjective branded. And this is where I've actually picked Paul. I'm going to stick with Paul. I don't know why, but we're going to stick with Paul. Paul the plumber. Is Paul the plumber any good? That's a subjective. Now, this is interesting. Again, Google and centralized search systems will do really well there because they have ratings and reviews and information. They've gathered a ton. And I think that works really well. But the last quadrant, which is I ask a question like, will Paul the plumber come out to my house at two in the morning? The only one who can actually answer that is Paul. And so these objective branded questions are where the business owner has this big advantage. Now, there's not as many queries like that, but those queries convert at probably 10 to 20 X. Trying to rank for plumbing in Texas. So what you end up with is you get a really close look. And where I think AI goes is Google. Absolutely. And other major search engines are going to take the first three quadrants and still have a ton of influence on them. But the last quadrant, I think it's very possible that you will start to see someone in their house. They come home, the kitchen sink is leaking, and they literally go, hey, assistant or R2D2, I got a plumbing problem. Can you immediately find someone for me that is open now that will come here and find them? Now, the reason why that is, is it's not necessarily about a copilot at that point. It's about an emergency and only I need a brand, I need an objective answer. I think that's where the AI steps in really clearly. So, for example, if I want to know what time a law firm practice opens and I ask that question of Google, unfortunately, I get 9 billion search results, right, with a ton of information. And half the time, that's not really what my next question, that's not really on my path. What I'd prefer is just someone immediately, an AI, literally just going,"They open at 9:00 AM." That's it. That's all I want. And the AI is really good at that, right? And that's where I think you start to see where I'm in my car and I ask my Tesla a question, or I ask whatever car, all cars are going to have AI. Now, Volkswagen just announced a deal where they're putting ChatGPT into every vehicle. So then it's sort of like, if you can just talk to it, that starts getting to where those branded objective questions really help consumers find the answer they need. I think the Internet as we know it kind of splits a little bit, and I don't think that's years away. The adoption curves of people baking this stuff into everything they do is all over the place. I think you're talking in the next 18 months, you are going to see that this is in a lot of places in your life very quickly.- You mentioned you're talking about that split. You wrote an article in"Adweek" sometime last year, and you made a couple of points around AI and generative AI's impact on the traditional kind of search funnel and you. One of the subheadings, I believe, was something like dialogue over monologue. We actually wrote a little bit of a whiteboard, not a little bit. We wrote a white paper here at Scorpion where kind of point number one was consumers will move from search to conversation. So dialogue to monologue. Search to conversation. What's your perspective on dialogue over monologue? And just that whole idea there. And we'd love to riff on that a little bit. I think I've got a perspective there, too, but love to hear what you meant by dialogue over monologue.- Yeah. So centralized search has forced all of us to put out as many monologues as possible and also the evil marketer monologue where it's sort of like, hey, I'm going to put this out there and then I'm going to track you, and then I'm going to target you, but you're going to step in my monologue as often as possible. And that's actually not ideal. When I think of the ideal customer journey, it's not that I found it via search, it's that when I'm talking with the business owner or with the attorney or with the plumber, whomever, I'm having a good interaction. See human trust. And this is not me talking. This is sort of just years of academic research. But we tend to trust because as I'm having a conversation, I'm learning from the other party, and they are staying in my flow, meaning I'm in control of what I'm trying to learn and they are responding well. And so what that does is we now have AI that can mimic that on behalf of the business, because most business owners, you don't have time to answer every call and talk for 20 minutes. There used to be an old, I think they're still around, but Zappos was a foot shoe company, if you recall Zappos. And they had a policy of basically bonusing people in their customer support division for talking longer to customers, which is absolutely crazy when you think about it. Cost perspective. Most people actually measure the benefit of their assistants or their call center by how quickly they can get rid of the person so they can pick up the next call. They took a totally different approach, and I think AI actually gives us that scaling. So in a dialogue, you're actually building trust. So if I could go to an AI and say, listen, I'm having trouble, I might have rear ended somebody, I need an attorney. And the AI says, well, let me tell you what we do and say, okay, well, is it going to cost me anything? No, you can come in for a free consultation. Everything that other than legal advice, this thing can help the person get to. That's where I think this changes. And Jamie, it has such big changes for marketing because what it means is marketing is about awareness, but the consideration and action phase really becomes the dialogue. It's almost like we put up a ton of web pages to make people land on the page. But most digital websites are just going to be a conversational interface so that you're capturing zero part of data that's freely given to the business and that helps you then pick how to proceed. And I think that's a huge opportunity.- Yeah, I agree. It's funny. We're talking a lot about, I think in that particular piece, you're talking a lot about the experience of a consumer can have through generative AI on a business's website, right? Where that business can apply AI through a chat or something like a GPT interface on their website that can answer questions in a very human like way, specifically about that business, right? So it's funny you mentioned that we haven't even talked about this, but the first kind of big AI centric product that we've invested in here at Scorpion is an AI chat product. And it was quite difficult for us because again, the nature of the businesses, we work with lots of attorneys, right? So you can't give out legal advice. Huge no. But then also, just like simple questions, know, working with a lot of HVAC contractors and getting the dialogue right, like, you don't have boilers in Florida, right? You got boilers in Boston. So actually building the AI in a way that could learn based on location, based on a business's website content and answer some of those questions. I think that's a unique value proposition. As a matter of fact, just from our perspective, when we've started layering that product on our customers websites, we're actually seeing conversions increase about 30% higher than they were before the chat product. So it's been a great way to your Zappos example a moment ago. What we see is a lot of the dialogue starts with that chat where it's kind of a long form question, right? It's like, hey, can you guys do this type of work? And then it kind of goes from that to like, well, could you come at this particular time and fix that problem? And then the chat pushes them into a call to action, like schedule an appointment online or pick up the phone and call us. So interesting that we're talking about it through that lens. But let's take it back before that, right? How is generative AI going to impact the Google search experience? Because it's one thing once they're on the business's website, but in most cases, unless businesses just have a ton of brand awareness and the consumer just goes directly there. Which, again, in the case of most plumbing and HVAC companies, most attorneys, you don't know the brand, right? You go to Google because you want to. Do you see that interface changing at all? Like Google?- Yeah, I know a lot of search engine optimization specialists, like SEOs, they're predicting Google will never launch, where it just starts talking to you. And I'm like, oh, yes, they definitely will. To me, that's very ostrich head in the sand. The reality is, I think Google starts to transition a little bit. I think it's a welcome transition, which is right now we bid on keywords and we optimize keywords. And that's how many of your businesses and our clients track that top of the funnel. But realistically, think of a dialogue versus keywords. Keywords, I type in plumber. Plumber now. Okay. Yeah, there's not a lot of information in that, but I've been taught that if I go too much deeper than that, it could take me off the rails or I'm just going to get ads anyway. Now, the opening salvo like you're seeing on your chats is literally like kitchen sink ruptured, can't use. I need someone here now. They freely tell you way more knowledge. And then the chat bot comes back and says, I understand completely. Do you know where your main valve is? And they answer, no, I don't. And you go, great. Okay, that changes my next question. What happens in one, two, three, back and forth would have taken the old school of people hitting your website going to Google. It would have taken that five months of tracking and it would still be wrong data here. It's like, no, it's such a better interaction and it helps you. Let's say you're the business owner, you can prioritize. Actually, I need to call this person back right now because it's flooding their house versus someone who knows where the main is. And the chatbot explains to them, go turn off the main if you know how to do that. So there's these elements of, you can scale yourself. And for Google, what I think happens, right, is if they can power that initial dialogue and really help people, I don't think they can shove ads in there. I think it's a really bad thought process. What starts to happen is after two or three, back and forth, let's say, about an attorney, Google can then say, hey, listen, based on what you've told me, I have four attorneys in the area. I've reached out to each one. These two will give you a free consultation right now. Would you like that? That's an offer that's different than an ad. And so I think what happens is no business is better positioned than Google to start to really make personalized offers based on the chat. And what's absolutely bonkers about that is they didn't have to track me and follow me around with creepy cookie technology. I literally had a chat and told them what I needed. Yeah, that's good for everybody involved because the offer is empowering you, the business owner, to step up and say, hey, if you're having a conversation like this, here are my offers that you can choose from, and the AI will be smart enough to say, this is a good offer, I should present that. But that's different than shoving ads into every visual sort of space on the page. This is really allowing for the dialogue to identify the right information and the right offer at the right time. And I really think Google is well positioned for that. That's what I see happening next.- You can see Google specifically kind of tiptoeing towards that with an ad product, like local service ads on the search side, where that ad product is not available to every category of business today, but it is available to most of the businesses that we work with that are in the services. Like your attorneys, your home service businesses. And for those of you that aren't aware, these local service ads are cost per lead ads today, they're not cost per click. What's really interesting is with these local service ads is that you don't pick your keywords, you don't place a keyword level bid. You tell the business the type of business you are, the types of work that you do. The geography where you'd like to show up, and then you just kind of trust it. You trust Google to display your ads, right? And you do have some control over your reviews and how reputable you are as a business. Those play into can you show up in those ad units or not? So the experience you're delivering in your service delivery, but the actual control over, it's again, very different than traditional ad words. So I could see a world where Google takes that ad unit a step further and says, hey, what types of offers are you willing to provide at a certain period of time for this type of query? That would be. That's interesting. I could absolutely see.- But Bing, I believe, has already started to put in ChatGPT so you can actually ask the ad questions. So that's again with an offer because hey, I can offer this for you, and you might have more questions. See, the problem with all of this is for the last 20 years we have recreated driving past a billboard at 60 miles an hour and then making you pay for it, for looking at it. It just didn't make any sense. What this is, is I'm not buying a legal, I need to talk to somebody or I need to have enough knowledge to know I would like to pursue and have continue this conversation. And I actually think that's better for all of us. If you type in right now something like Nike men's shoes, size nine, the entire page that you see on Google is ads. And it has no idea if I'm looking, if I'm looking for my son or if I'm looking like it has no context. And I think what we're driving towards here is context is king. It's also more privacy focused, which is if you tell me what you're looking for, I don't really have to guess, I don't have to follow you around, I don't have to read your garbage. That's a big change. But marketers are going to have to get comfortable with this idea of scaling themselves as businesses through using tools like this. Like you had pointed out, the AI is getting there. I wouldn't say anyone has cracked the nut. You might have seen some guy convince an AI to sell him a Tahoe for a dollar. Actually, people should look it up.- I've not seen that, but I have heard about that, yes.- It's very funny. And there's a few. There was an AI in Europe that the customer convinced it to always swear at them when it responded. And those responses are absolutely hilarious. But that's the part you can't have that happen, and we're still working through that. But I think it's good enough, certainly if you lock it down, and I know you guys as well as we, that's a lot of what the focus is, locking it down. But if that dialogue starts to really work, attaching a series of potential offers as something the AI can leverage, totally, totally going to happen. Like, definitely going to happen soon.- Yeah. Couldn't agree more. I think something else you also hit on a couple of minutes ago is you talked about just the impact that Google, the blank Google page with the simple search box, like the impact that that alone had on how we interact with the Internet as humans. And again, Google's kind of predictive search certainly helps. But you go as a consumer and you type in plumber, and you're going to get, like predictive near me. Like, oh, yeah, that's what I'm looking for, plumber near me, right? Or plumber Dallas, Texas. It's like, I would never ask, if I'm having a conversation with you, Christian. And I go, "Christian, plumber, Dallas, Texas." You'd be like, "What in the hell is wrong with you, Jamie? Did you smoke something this morning?" And I would say, "No, I didn't." But I would say, "Christian, who's the best plumber in Dallas that services my zip code?" That's how I'd ask you that question. But if I go to Google, it's plumber near me. I think what's interesting, though, is ChatGPT, and these interfaces are teaching consumers that we can actually have a conversation with the Internet like we have a conversation with each other. How do you think about how that changes typical SEO today, right? Because today people optimize. You mentioned SEOs. SEOs optimize around these things like plumber near me. Well, as over time, I'm going to learn that I can actually ask more specific questions to my Internet, to my Google, to ChatGPT, or whatever the interface is. So how do you think SEO should be thinking about things like that and the changes that this is having on how people search again? Because I think people are going to move from this idea of traditional search, plumber near me, to a conversation with an interface.- The search is actually just going to come up also in a broader conversation. So you might be talking to your AI about remodeling and it's going to recommend or have offers from five, a contractor, a plumber. So it's not just, I need one thing. It should be in context of an hour long conversation about remodeling the kitchen before you put it on the market. Think of the number of professionals that are needed in that dialogue, but they can't access it because it's all through search. Now, I will never bend against Google, they built, and what they built was probably the most important innovation that we've seen, but I will say that it's highly inefficient. And so what I think that you're seeing is, my wife works in pediatrics as an occupational therapist, and we often debate sort of the concept of, there's something called Piaget's stages of development of children, and it's basically like, before two, it's just babbling. Well, that's keyword search. Like, we're just in the baby, baby stages of finding information where it's literally like, baba, I want my bottle. But that's all they got. That's key. If you actually read what you type into search boxes, you'd be amazed at how childish you sound.- It's true, though. Yeah.- The next stage in childhood development, really, from two to eight, is really where a conversation starts to work. And if you have children and you're listening, you're aware of this, it's amazing, the uptake of what they understand. And then after eight, you start getting into much more complex thought. So not just describing what you're seeing in your interaction, but complex structures. We're now moving into this second phase and it's really kind of exciting to watch, which is we're going to have those dialogues. But I think saying goodbye, if I'm an SEO, I offer this. I think it shouldn't be search engine optimization, it should be search experience optimization. That way they can keep all of the conferences and T shirts the same. But it's an experience, right? Because OpenAI is not a search engine, and ChatGPT, and these are completely different mechanisms, and the way they process knowledge is completely different. And the way humans engage with them is a new experience, but that is going to be the experience going forward. Like, for example, I want everyone here, just ask yourself this question. Do you think you're going to be talking more or less with machines in the next five years? You have no idea how fast it's going to ramp. And so every decision now, again, I'm way out in geek land where I'm using this stuff more than I use Google now, but the reality is it's opening up these doors, and you said it really well, which is in the past, computer savvy humans had an advantage. They could get the computers to do what they wanted. But now that computers are human savvy, that advantage is open to everybody because you could just tell it, hey, I want to build a new website or I want to do a campaign. You'll be able to talk your way through things, and that opens up all new possibilities for business owners to scale themselves where they never would have been able to do that before. And I think that's really very likely.- Yeah, I think you're right. I think the other thing that you mentioned, computers are becoming more human savvy. I think that that's going to give humans the confidence to ask questions of certain business types that they just didn't ask in the past. So one area that I always kind of advise residential service businesses to think about from a content perspective is, look, that's been a business where pricing transparency has always been very much closed off. Like, oh, well, I can't give you a price to fix your plumbing until I send someone over to do an inspection or a diagnostic. And there's some truth to that, right? Because sometimes a drain clog is not a drain clog, is not a drain clog, right? Toilet paper getting stuck versus a root growing through is a very different problem with different pricing. But I do think that there are some questions that could be asked that I think are very common on a consumer's mind. And at some point there is going to be enough content on the web that these models have access to, that these generative AI models have access to where someone can go and ask ChatGPT,"Hey, I've got a clogged drain. This is what's going on. What's the average price to fix that in my area?" Someone is going to start to publish content and answer that question, or at least provide some guidance. And I think they're going to have advantages in these GPT interfaces because GPTs want to get a response someplace and they're going to say, hey, going back to Paul, your Paul example, Paul the plumber says he's fixed over 500 drain clogs and the average price is between $500 and $1200, depending on X, Y and Z. Do you have an opinion, or do you have a perspective on, again, just how GPT is like Google did? It influenced how we search the web. Do you feel like GPT or these types of open AI or, I don't know what we call them, generative AI search experiences are going to impact the types of questions that consumers are asking that maybe give businesses an opportunity to think about how to get in front of those queries differently than they get in front of a traditional Google search?- Yeah, there's two things here. Number one, I actually think we've been doing it wrong in marketing for a very long time. And I'll get a lot of crap for saying this, but basically, most of the signals that we analyze when we're helping a business rank and show up, they're horrible with something called survivorship bias, which is we're asking these questions, they land on your website and 90% of them or whatever, bounce away back to Google because they can't find what they're looking for. But then the people that do stay and they click around, they're hunting for something. You're getting those signals and, like, we should optimize for these signals. This is amazing data. It's not amazing data. It's literally the bombers that made it back right in the classic survivorship bias. That's wrong. What we should be doing is what Google's been doing, which is every time I land on their site, even though it's keyword, I tell them what I'm looking for in my own words. So what I think, to answer your question, is the way we get to which plumber is going to bring price transparency or which lawyer is going to explain what they do and what they can say. What you're actually going to do is once you have on your site, let's say you have the Scorpion chat, and on that site, I go in and I start talking to it. Your team at Scorpion then analyzes what are people asking on this page, and what do they then ask next? Then you take that data and say, you know what we need to do? We need to produce monologues, web pages that answer these. Because if they're asking the question on your website, they're actually asking it 10,000 fold on Google.- 1000%, man, that's so true.- You just don't know what's happening because you're like, no, I like my drop down menus. I'm like, your drop down menus are terrible! Get a search bar in a chat box because it will tell you what they don't understand. And that's the goal. That's where the dialogue teaches you what to create the monologue about, so that you capture more people to have more dialogue and it creates this wonderfully smart optimization window.- Yeah, I could not agree more. That's exactly how we're thinking about and what we're using our chat product to do is actually influence the types of content that we would go produce for our customers on their website. Again, just to optimize around the questions that we already know are being asked because they're being asked through the chat interface. So one of the thing that I picked up from you and following you online is this paper written by a couple, I think it's Princeton guys, around this concept of generative engine optimization. And again, I'm not as skilled in the world of structured data as you are, just probably enough to be dangerous. But I read that as kind of a framework for how to approach optimizing content for generative AI, for machine learning, search engines or indexes or however you want to call that. Is that kind of spot on on what the purpose of that paper is?- Yeah. So that paper is a really good example of, number one, it's not peer reviewed, right? So it's not like a true academy. It's called archive. And the reason why I shared that paper with everyone was it's actually a good framing of the problem. It's not like it has an answer. It's more looking at it. And I think to a lot of SEOs, it's going, oh boy, this is so different. How are we going to understand how to optimize for this? And so they propose some things, but it's still way too early to say this is how you win for it.- Yeah, of course.- I'll posit a different scenario, which is I think they frame the problem well, which is I don't know if buying this ad helped. If those companies start showing ads and offers, I don't know if writing up a long form blog post or a video, it's not as well trodden in terms of a path. The other thing is when people start talking to AI because there's such variability. So one of the beauties of keywords is, plumber near me. Everyone can target, right, because they can agree. But if it gets to being much more open, I think what actually starts to happen, and again, I'm going to anger some people and other people are going to be like, hallelujah. It's going to change how conversion really happens because we've all been taught, once again, I'm going to say I think we're doing it wrong. We've all been taught, I want to rank for plumber, Texas. That's the key. But when you do rank for that, with the money you're spending, there's no conversions necessarily at plumber, Texas. The conversions are plumber remodeled modern kitchen, Dallas. Those longer ones are where the conversions happen. Because I've given you enough that you really can nail it with good content and show up. The problem is to write the long tail you need AI. You and I don't like our plumbers and lawyer friends. We just don't have enough copywriters to put out enough monologues on the stuff you actually do. So it's that same idea of the research is going to keep coming out about how to optimize. But I actually think it's much simpler than that, which is ask customers if you can record talking to them. Ask customers when they're on the phone, get permission in all cases, even when you're sitting with someone, go get an application like Apple voice memos, just ask them to record it. Then hand it to a team like at Scorpion and be like, can you mine this for all the great structure we should be talking about? Because that's what's going to work best in these dialogue based systems.- Yeah, yeah, I agreed. I mean, it's interesting, we just, again, just not to get too much in product talk, but we released another kind of AI driven product that when we're doing marketing, be it organic or paid advertising, obviously a big part of our value proposition is attribution tracking to the best of our ability. So call tracking leads and recording, and of course there's always the call out. This call may be monitored or recorded for quality assurance purposes, but one thing that we started doing recently is we launched this product where we're actually leveraging AI to actually listen to and transcribe the phone calls and actually score and rank the quality of the call. Because again, one of the problems that small businesses have particularly, it doesn't matter if you're a franchise or you're an independent, that small business owner and operator doesn't have time to go listen to every phone call that comes into their business and say, hey, did my team handle this well? Or hey, like you said, is this an emergency situation that would be high dollar that we should prioritize now. So we started doing that. And again, it's interesting mining that data and it's seeing the questions and the types of things that consumers are asking for and how businesses could take that information and leverage it to produce, like you call it, monologue content. Content on their website, that again, these are already questions that we know, we've recorded the call, that are being asked.- And also, you actually brought up before what a lot of people don't realize is that Yext was a company called Felix before it became Yext.- Yeah, that's right.- And that process they were trying to do back 15 years ago where they'd listen to the call and then they used the IBM speech recognition software, which now I look back, it was like the baby before we have what we have today. But that was the idea. And look, I think a lot of what marketing platforms have been trying to do for the small business owner for over the years have been spot on in sort of, you look at what you have available to you and you try and optimize for it. That search. In this case, I'm actually saying something different. I think we're optimizing the entire business, which is every dialogue, every interaction can be stored and understood for its value, and it can tell you what you're not really doing well. I'm amazed at how many small businesses or franchise owners or anyone that is in market. I'm amazed at how the real beauty of most of these people and most of these businesses is their storytelling, of why they're an attorney in this area. And so much of that is never captured, but it's the thing that the customers love about them. I think this is a little bit of, I would say, I think getting back to the roots of having more dialogues with customers and capturing the essence of that, versus we kind of went down this path of, hey, if I just spam out enough stuff, I'm going to show up for plumber. And I'm like, I see what you're doing there, but I'm not sure that's actually the point of marketing. I think the point of marketing is to create a human connection with your audience, your chosen, happy customer. And we've lost the journey a bit. And I actually think this could bring a golden age of rethinking, a renaissance of what does it mean to engage with customers.- Yeah, I agree. I would say that, again, we work with enough businesses, and we've got somewhere in the range of like 16,000 or so customers today. Our customers, they're paying us to pretty much run their marketing, right. Most of them rely on us. We're providing them with the tech stack that they do all this stuff on. But we're actually pulling the levers, right. We're doing a lot of the work for them. And I would tell you, going back to the storytelling piece, the customers that actually invest in telling their story, like video content and engaging on social platforms, right, head and shoulders, they convert more leads, they convert more customers than the ones that don't do that, that just rely on, let me just blow a bunch of money in paid search and just spend, spend. So your thesis on storytelling and kind of getting back to the heart of what marketing is, we see that play out all the time. That's why we always push our customers. Hey, you should invest in not using stock photography, like, simple things like that. Like, don't use stock photography on your website. Hire a professional to come out and take some real photos. Do a bio video on why you founded the company and the impact that you make on your customers lives and the impact that you make on your employees lives. Those things matter, even to people that have a plumbing problem.- Yes. Yes.- So I think that's spot on. I want to shift the conversation. By the way, this has been awesome. But I want to be mindful of your time. But I know that you mentioned, you said that you took about a year off from Yext, and I know one of the projects that you were working on around that time was around data privacy. And I know that's something that you've got a lot of knowledge around. So I'd love to get your perspective on how you feel or how you believe kind of everything from legislation and just common good sense, practical business practice that we should be thinking about as it relates to online privacy and cookies and things of that nature. What's your perspective on where things are and where things are headed?- Yeah, it's a great question. I will partially channel my brother, who's a data privacy attorney. That's why we teamed up.- I remember you reached out to me around them because you're like, hey, can you help us out with some marketing things? And I was like, man, we're really more B to C, not B to B. I don't know if we'll be a ton of value there, but I appreciated the reach out, regardless.- Of course.- But channel away. Plagiarize the hell out of your brother, by all means.- Yeah. This is actually what Thanksgiving dinner discussions are for us, which is sad, but basically, the way we look at this is, I don't think data privacy ever had the momentum to have the impact here in the States than it did abroad. And there's actually really good reasons, historical reasons, why, which is if you consider World War II and you consider everything that happened then, privacy of who you are and what religion you are and where you were born, that is ingrained in their cultures. And it's different here. Here we went almost the other way, which is something called surveillance capitalism, which is if the product is free, then the product is me, which is every time you use Facebook, Instagram, all those things, they're selling you to a marketer. And so it's a different approach. And the reason why AI legislation didn't work here, now we have, I think there's 18 states that now have privacy laws. But the reason why you can't get it over is it's too hard to understand. I feel like everyone listening to this, I don't know, and I do this for a living. I don't know all the stuff that's tracked. I know that I talk to my neighbors and suddenly there's an ad for what we were talking about outside on my phone. And it's scary. That stuff's bad. But what Jay would say, and I completely agree, is AI is likely to be the thing that finally brings privacy home. And here's why. In the law, the way laws are set up today, when you try and explain what audience segmentation is and targeting based on third party, first party, and zero party, the lawyers and the judges, their eyes roll back. They have no idea what you're talking about. But on the other hand, let's say I say I hired a young person at a college to stand in the front of my law office as the sort of the greeter, and they'll answer any question someone has, but they don't give legal advice, but they'll answer any question they have. And someone comes and they convince that person, and the person who hired is not good. They lie, they're abusive in their content. They mislead people. Well, that's AI. If you're using AI to scale dialogues, it can't get this stuff wrong because it opens up what judges and legislatures totally understand, which is if you hire someone to be your representative and they go off the rails, you're at fault. They are your agent. And so it doesn't mean don't use these tools. It means you have to have the right plan of how to use these tools. And so I'll put it very succinctly, which is you should be building a lockbox of the data that you're okay with the AI representing. You do not feed the whole website you have to the AI and then say, go for it, talk about this stuff. And the reason why is if you feed, like, I actually tried this. If you feed, like, Viking websites, you know, Viking stoves are, like, super posh.- Yeah, of course.- Really nice. If you feed their website, which has thousands of pages to an AI and start asking questions about Viking stove, it does okay. But then if you ask how did the Vikings use stoves to cross the English Channel and attack the Irish, it will write you an amazing conversation about how that happened, because there's too many words. You've given it too big a leash. What you really want to do is be able to say to regulators, and the way to prepare for AI is, for everyone listening, your AI strategy to leverage it is really just a data strategy, meaning you need to put data aside just like you do with Google profiles or with all this structured data we talked about. You want to be able to say to a regulator, I didn't let that thing have access to anything but knowledge that I said, I'm okay with the consumer knowing this. I've set it aside and it's very specific to answer a question. That way, when new AI tools come out, you've got your little module here of perfectly structured data. That's much easier to manage when you unplug GPT-3 and you plug in Claude or whatever five, you can swap them out and you're not beholden to letting the AI do too much. So it's very much about having this almost like lockbox mentality of this data over here is okay for consumers to engage with. Do not hook up AI to your whole data stack. I think it's a massive mistake, and it's going to lead to a lot of areas where lawyers will be able to prove you didn't do the due diligence in training or hiring that AI to prevent it from doing the thing it did. It's very focused on how you control your data that you make available to the AI.- Yeah, that makes sense. And it seems like, we've got to be heading in that direction one way or another. I mean, even if you look at some of the lawsuits that have come out with some of your big publishers, I think "The New York Times" sued OpenAI, right, for accessing and referencing a lot of their data and how they formulated answers. There were some big suits. I think there is going to be some world, right, where, yeah, like you said, if you're a business, franchise or otherwise, it's almost like a robot.txt file, right?- Yes.- It's like, hey, you can access this part of my website to formulate answers or information, but this over here is mine and you can't touch that for now.- I do think there's a way to approach it. The reality is this stuff is, I mean, literally, you and I have talked for whatever, 40 minutes-ish. Do you know, probably by the time we're done talking, there's going to be three new AI platforms that are created? It's moving so fast. As a business owner, the most important thing is to work with your Scorpion rep or whomever. But say, I want to start thinking about what knowledge to separate so that know there is an error, an issue, or why did it give Paul this answer but Sally, this answer, you want to be able to tell regulators or anyone, hey, no, we're really careful. This thing is not off the leash reading whatever. It only uses this knowledge. So if a mistake was made, you can actually figure out why the mistake was made. If you feed something the Internet and 10,000 pages from your website, you're going to have a problem about what it comes up with. So it's having a very particular approach.- Last question I just want to kind of bring back though, we talked about privacy, we talked about impacts to search. One area that I think that a lot of local businesses particularly, again, going back to service based businesses, your oh, shit categories, one of the areas that they just struggle to really make sense of and work for their business at scale are your social platforms.- Yeah.- Right, again, do you really want to get blasted by Facebook ads from your plumber or your attorney right now, retargeting and things of that nature? They've been on my website. I know they're interested. Therefore I want to target them and stay top of mind all day long. But it's really difficult to just spend a bunch of money on Facebook if you're a service based business. If your goal is not brand awareness, if your goal is, I want to hit a certain cost per lead metric or cost per customer acquisition metric. My hypothesis, however, is that some of these social platforms, in particular Facebook, are super well positioned in this ChatGPT interface world. Because if I think about the number of people that I know that I go log into my feed and say, can someone recommend a plumber near me, or can someone recommend a great criminal defense attorney? Hopefully there's not many of those in my network. You get the point. I feel like Facebook is sitting on so much data. So much Q&A data. You talked about the dialogue type conversations that happen online. Do you feel like there's a world? And I know we're just kind of riffing here, and this is just totally hypothesis, but do you feel like there's a world where Facebook is thinking in the back of our mind, say, this is a place where we can really get into that search/dialogue conversation with our user base and not just treat our users like the product, which is what you mentioned earlier, because today that's what we are, right? We are the product for Facebook. Do you feel like that there's a place where they maybe can get an advantage here? What's your thought on that?- So first off, the promise of Facebook was always the human graph, which is you and I know each other. And if I know you just remodeled your place in Dallas and I know I liked it, of course I want your opinion and other people's opinions. It never really panned out. It's partially because they tried search and other things that never really quite did it. I think they could get here. So Facebook, if you are aware, they launched their own LLaMA models. They're launching them with free commercial rights. In a way, they're kind of ruining it for everybody else, like OpenAI, because they're like, ours is going to be good enough. See, what's going to happen this year is language models, the things that we're talking to, they're going to reach, like asymptote is what you call it, but they sort of going to reach a point where they're all pretty good and they all kind of understand everything enough, right? And so Facebook, I think, is doing this as part of a plan, which is if they have that, they can build it into the metaverse, where you can talk to anyone about anything or any object, because you need that for people to understand how to use the metaverse. It's too complex for most people, so you need the AI to be able to guide them. But also if they give it away to free to all of their clients, like your clients, then they could power some of the chat, they could power the search. So I think this is a bit of an arms race in terms of who gets there. But I would look at it this way. There comes a point where I say, you know, Jamie, you and I are on the same page. Okay, to a machine it's like, are you guys standing on a big sheet of paper? What do you mean you're on the same page? It doesn't understand some of the idioms, and it's getting so good. That was that aha moment, for someone like Facebook. Bing has been brilliant with this. With Microsoft partnering with OpenAI, Google will get there. You don't bet against them. They have everything they need to do this. They just have to execute you. Get to this world where depending on what you're trying to achieve, awareness is one thing, but if I'm on a platform and it could be Instagram, whatever, and just say, hey, does anyone have a good plumber? I have a problem right now at the house, but I'd love it if someone told me. The AI actually could go back in and say, it looks like these three or four people are really, you should ask them particularly. I think that is absolutely going to happen. It's so funny. Technology is the whole thing that gives humans back time. The time it takes me to figure out which plumber to call for a remodel. That might be 5 hours and 82 tabs at different times over the course of a week and a half. This thing literally is just going to walk and talk to you and use things like your friends in the Facebook graph, but also what offers are out there from contractors. It's all going to start coming together to where you can say to your AI, "You know what? Can you just figure out who I should hire based on what other people have done and get the best quotes and just make me a dossier, send it to me at the end of the day." I know people think that's like "Star Trek." It's not. It is very close. The way you win is the more knowledge you're willing to share with these systems about who you are, what your story is, what services you provide, the detail, structured knowledge. Then you're enabling these tools to take advantage of you. And look, it's been this way for a long time. If you search for plumber and there's no photos, there's no details on the Google list, those listings always underperform. The more knowledge you'll let people have about who you are, and it's factual, the better you're going to do. And I do not think that changes with these systems. I think it's the exact same model. The more knowledge, the better, and that's probably where it goes.- I don't think that we should end on anything other than that. I think that's a great wrap up. Like you said, these new advanced technologies like OpenAI and these generative search products that are going to come out of this, it's going to enable a more efficient pathway to the same data that consumers want access to today. It's just a matter of learning how to take advantage of those things and what pathways you have to feed that data to those systems so that they can actually surface your business relative to the guy or the gal down the street. So, man, Christian, this has been awesome. We're just over an hour of talk time. I'm sure it'll get edited down a little bit. But man, I've always really appreciated you personally. I know we've had a lot of good times out and about together over the years. We've had a lot of really great philosophical conversations like we've had today, and then also just want to say on behalf of Scorpion, we really appreciate the relationship that we've got with Yext. It's going on I think probably eight years or so now that we've worked with you guys. So we really appreciate the partnership as well. And, man, thanks so much for joining and taking hour with us today.- Thank you so much. And obviously, I'm thrilled with the partnership and the friendship over the years. So if there's any follow up questions or any other information, I'm happy to jump in and certainly speak with anyone.- Yeah, we're going to share your contact info, your LinkedIn profile, all that good stuff in the show notes. We'll also reference the generative engine optimization white paper that we talked briefly about. Again, I think, like Christian mentioned, it's not peer reviewed, it's not factual, but it does kind of lay out the problem. And there's some interesting things. I think if you're interested in all these things, you'll find value reading that as well. So, Christian, thanks a lot and again, till next time, thanks for joining this episode of "The Franchise Edge." We hope you enjoyed it, and we'll be back again real soon. Take care, everybody.(upbeat music)