← Marketing Trends
Marketing Trends
Spotlight: AI Context Windows Will Kill Forrester, Gartner, and G2 (Yext CEO Explains)
00:15:25
Best practices told you to optimize reputation across all 10,000 locations equally—but granular competitive data just revealed that 7,000 of those stores don't need it, and you're wasting millions on the wrong priorities.
Loading summary
Transcript14 lines
- [00:01]
A
When I talk to a lot of our clients, I do position what we do as we create content that goes everywhere. And we've got YouTube channels, clips, everything on LinkedIn, Apple Podcasts written. So our content goes everywhere and you want your brand showing up alongside actual thought leadership content. Oftentimes, if a company is saying I'm the best, it's not going to be as heavily weighted as other conversations that your buyers are actually following and they're saying things are the best. Is that an accurate way to think about how to actually position being found from a content perspective is like being everywhere?
- [00:33]
B
Yes, I think it is. I think it's been a best practice. Right. And so I think what's going to happen though is that we're going to quickly move into what I would call like a post best practice world. I'll give you a different example, but similar. It has been sort of conventional, sort of obvious that managing your reputation online, the number of reviews you have, the star ratings of those reviews, the commentary in those reviews, that has been a best practice for, you know, just about every business and particularly for any kind of local lens business, what that leads to is that leads us to expend a lot of energy on a broad based best practice. So if I, if I have 10,000 taco joints, I've been trained and taught as a marketer that like I need to be constantly optimizing the reputation of each one of those 10,000 taco joints. That's best practice thinking, right? And it's, it's been really useful because we've lacked the granular data to understand that probably there's some subset of those locations that actually need more reputation or better reputation, right? And that, that all comes back to the sort of granular competitive environment. So if I have a global, global chain of 10,000 taco restaurants, I need to understand which ones of those brand visibility are suffering because the, the star rating isn't high enough or because they don't have enough reviews, maybe because they're new, maybe because they don't solicit, maybe because the store manager isn't doing what they're supposed to be doing in the store. I could give you a hundred other versions of that same story. And what it comes down to is that best practices go out the window when you can deliver like really bespoke brand visibility optimization mechanisms across each of those individual stores. And that gets into a lot of operational detail about like how do you actually do it, right? So if I have 3,000 of the 10,000 stores, desperately need to go from 3.5 stars to 4 stars in order to be the number one sort of visibility taco store in their market. You know, as a marketing manager, that's a huge challenge because how do I get 3,000 stores? It was easier when it was like, look, I'm just going to improve the reputation across all 10,000 stores. But I spent a lot of money on 7,000 stores that don't need it. Right. And so you start to get to this world of you need better tools and better ways of doing that. So, for example, marketers, brand marketers with a local lens have long resisted the notion that we should put some of the control over this marketing into the hands of store managers because we lose centralized control. And you don't really want your store managers writing the responses to the reviews because they are going to use different voices and things like that. But you do need to be responsive. And so there are ways where we can start thinking in terms of. Once I can identify across trillions of data points, what is the thing that I need to do in this specific store, I can start pushing tasks to that store manager. Hey, what I need you to do today is respond to these three reviews, and the responses have been drafted for you. You just have to submit them. What I need you to do today is take a bunch of pictures of the store and post them to the Google profile or to these other. These other things. And all of that is a form of data content creation that winds up being bespoke to the individual. You know, it could be product, it could be service, it could be location, it could be any of those things.
- [03:57]
A
So fascinating. I was just thinking as you were talking too. I'm sorry I keep shifting back to B2B, but that is my world.
- [04:03]
B
So I'm like, yeah, no, that's our world too.
- [04:05]
A
Yeah, yeah. So these. And I don't want to call names out, but it's just easier if I do. So like the Foresters, The Gartners, the G2, what do you think? Like that world, to me, I think could get very disrupted because there's going to be so much data out there that is now pulled in very different ways. You can actually see if consumers are happy. You can see if things are getting used. You can see probably the sources of what's being used. And I wonder if these pay to play platforms that have thrived for a while aren't really going to be around because, like, they're just. At least, I think it's hard to trust what they say, like what they Say anymore. Like, is that actually the best you're in some Gartner thing? Do I care? I don't know. So what do you think about that?
- [04:46]
B
Yeah, I mean, I think it's going to get easier for marketers to understand the value they're deriving from all of these kind of software and services capabilities. Right. So when we think about, when I think about this, today you've got kind of the software market's a half a trillion dollars a year. And the services business, the services marketing and things like that is like almost somewhere between one and a half and two trillion dollars a year. So companies are spending somewhere on the order of two to two and a half trillion dollars a year on some combination of services and software. And historically we've drawn really clear lines between services and software. And we've said, look like software companies shouldn't do services because it devalues the, the sort of pure software product. And services companies shouldn't do, shouldn't do software because they need to be agnostic. Right. I think you can dynamite that whole notion and I think you can say in a world of AI, where services become easier to deliver and more of it's automated, every marketer, every brand, every business should expect a blending of software and services that's really keyed around like the job that needs to be done. The way we look at our mission is we sit with a marketer and we help them understand kind of the competitive data landscape for their brand, for their business. And then we will apply the right set of software technology services around making sure that the job gets done. And so we get out of the business of buy a reputation management platform and just manage your reputation and go through the motions and we'll give you some proprietary scoring mechanism that lets you declare victory and go down to like, that's one of 100 tactics that you need to be able to deploy. And all of those tactics should be geared back to the core data set. I don't know what that does to the like sort of, you know, the, what I would call like the sort of endorsement community for these software and services. But I think it gets a lot more confusing and I think we're already seeing it. I think we're seeing the two sort of, you know, highest, two of the highest value software companies on the planet, Palantir and ServiceNow. Like, we can't tell whether they're selling software or services. This is, this is, you know, Palantir gets knocked for this because they're basically like, I don't know if they're selling services or they're selling software, and yet they trade for 105 times like the next 12 months revenue. So obviously people see, you know, somebody sees value in what they're, in what they're doing.
- [07:07]
A
Yeah. I mean, Palantir's got that long term thinking, though, that I think if more leaders could also encompass that kind of thinking, it'd be a very different world of taking longer term bets and not how do I think of like, what will I see within the next three months? And am I seeing demand? Nope, not yet. Okay. Not worrying about that or I'm not going to worry about brand. So Palantir, watch.
- [07:28]
B
I think there's, there's like another vector to this whole thing that is just beginning to kind of hit the radar and that's this. There's a fundamental shift in this idea of context. The way I try to describe this is, you know, if you have, you know, it might be your, your partner or your spouse or your, or your longtime assistant, the, the amount of things that those people need to know about you to do something for you. I've been married for 25 years in October. My wife and I have been together for 31 years.
- [08:02]
A
Wow.
- [08:03]
B
When I talk to her, you know, if she offers to bring me lunch, right? She's a restaurant entrepreneur, so she offers to bring me lunch a lot. You know, our communication is, would you like some lunch? And my answer is like, yeah, would love some. Right? And what arrives in front of me is going to be something I like. It's not because she's omniscient. It's because she's known me for 31 years and she knows that there are certain things, you know, on the menu of her restaurant or another restaurant that I, that I'm going to love. And certain things I'm not going to love. My assistant, you know, you know, knows things about the way that I want to structure my day the way that I want, you know, how I want breaks, things like that. It's all context, what we're evolving into as a world. And if you're listening to this and you haven't done it, go into your ChatGPT and explore the memory that exists inside your ChatGPT, because that's the context. And it turns out that actually it remembers a lot about what you talk to it about. And there'll be things in there that you'll be surprised. Like, wow, I never told it that, but it's figured out that I don't like to eat at places that Use seed oil to cook. Or I always ask about sort of nut free options because maybe somebody in my circle has a nut allergy. That, that's like, that's the, that's where we're headed. And what it makes is it makes every prompt and every query and every conversation completely bespoke. And I can teach it, I can tell it like, I like to do business with people who play golf because I like to play golf, right? And so if, if, and this, this goes back to that idea of like, what's the content that marketers are putting in the world? If I have a thousand or five thousand or ten thousand insurance agents, I want them to distribute through me as much individualized content as they can. So that inside that context window I can, you know, that helps the AI identify. Well, Mike likes to do business when he's looking for insurance and he likes to do business with people who play golf. Well, here's a guy who's across town who has played college golf. In his biography, that's a match. But I'm not writing a query that says I'm looking for car insurance. Near me, not more than 20 miles. I want somebody who, you know, you're not telling it that. It's just nosy.
- [10:11]
A
So what are some big bets you all are taking right now? Because you're operating, I think, in a space and your company is pretty far ahead. And like you said, you've been kind of pitching this vision for a while and it seems like now people are definitely going to be betting in this. But what big bets are you taking within the company or with your team that also might look a little crazy from the outside?
- [10:30]
B
Yeah, so a couple things there. So it's really interesting. I think we have huge advantages in this world if this is the way that it plays out. I think we have one significant disadvantage which, you know is what it is, which is, you know, the kind of, I think the history of the company like is not as clean as like the pure Play startup that's saying, look, what we do is AI visibility optimization, right? And there's a bunch of those. And you're watching them get funded, you're watching them show up. The good thing for us is that we've been doing this for a lot longer. We'll have more R and D budget to spend than, you know, than the sum total of those companies will have in revenue. So that's, you know, for a foreseeable period of time. So we can move really fast and we can. And then I think the biggest advantage we have is just our capacity to gather and store and utilize huge data sets. So we have a new product, it's called Scout. It basically is the most granular, in my opinion, competitive intelligence product ever built for traditional search and AI visibility. And what we're doing is effectively using an army of AI crawlers and bots and scrapers to go out and gather hyper granular data around rooftop level businesses. So I'm going to use my wife's store, Organic Crush Organic, Very beautiful, healthy organic. She has five restaurants on Long Island. Yeah, so I'm plugging here. I am plugging, yes, I'm going. She, she brings me free lunch and I plug. But so you know her, what she cares about is our terms like healthy organic food near me, healthy lunch near me, things like that. This new product allows us to, to send this kind of army, army of AI into the world, do highly localized searches across traditional search and AI and pull back trillions of data points. So for every one of her five locations, every time we do this, we pull down her rank and her top 40 competitors or as many as are, as many as are delivered by the AI because they typically deliver less than the traditional search. And then we collect somewhere between 150 to 200 non performance attributes for every one of those. So you can imagine the data file on a single location, single data poll is something like one store, 40 competitors by, you know, so it's, it's kind of 40 columns wide and 200 columns deep. And then we can compare all that data so we can basically say for this store you are doing great on number of reviews, your review response time is too slow, you don't have enough Google Photos. The schema on your page is, is suboptimal. You don't have your category or you know, in, in in the tags. Your page load speed is too slow, there's not enough words on the page. And, and we're not just making those things up. This is how we move away from the best practice. We're actually looking at all your competition and how they're ranking and how their brand visibility is and we're showing it to you that way. And it's, you know, it's, it's kind of mind boggling how much data is being gathered and stored and, and, and what's going to be really cool about this is being able to kind of time series all this over time. So for each individual store in a 10,000 or 15 or 20,000 network of retail stores or hospitality or financial advisors, you're going to be able to basically look at it over time and say, what's changing about my brand visibility and all of my competition so that I can get away from best practices and into individual bespoke recommendations.
- [14:00]
A
Yeah, I don't have a local company, but if I did, I'm in. I would buy this. It sounds awesome.
- [14:06]
B
It's really neat. What's funny about it is like the worst reaction we get to it is, wow, this is everything I've always wanted. I can see everything now. I don't know what to do about it. I don't know how I'm going to operationalize the, you know, you've shown me that I'm going to always have the ability to understand what's the key thing for each of the entities. Right. Or for my brand in this granular sort of minute locality. That's that other case. Right. Brand at the local level, how do I action it? Right. And that's where the AI story is going to be so good here because you can feed this data into AI and then ultimately, and you asked what differentiates us? It's, it's the fact that listings, social reputation, page content, social content, data, all these things, we have all the tools to take the action. So you fast forward. Eventually what we're going to be doing is we're going to be automating a lot of the endpoints and the action into the endpoints here and then being able to track the ROI over time. It's going to take a little bit of time to get to the point where like, I think it'll be a little like self driving, like it's going to be safer before marketers are willing to sort of unfetter the machines to do more of the work.