WEBVTT 00:00:00.000 --> 00:00:01.000 de influence. Okay, let's go 00:00:01.000 --> 00:00:04.000 de-influencer Alisa Sharf. 00:00:04.000 --> 00:00:05.000 Don't 00:00:05.000 --> 00:00:06.000 Hello 00:00:06.000 --> 00:00:12.000 Hello, and welcome in. Alright, I think we're live 00:00:12.000 --> 00:00:14.000 We live 00:00:14.000 --> 00:00:17.000 Okay, let's do this. 00:00:17.000 --> 00:00:18.000 And you know 00:00:18.000 --> 00:00:25.000 I have this. Yeah, I've missed having you here, Will. We've had a couple episodes without you, and now you're back. This is very exciting. 00:00:25.000 --> 00:00:26.000 For real. 00:00:26.000 --> 00:00:27.000 Yeah, I wasn't on our time zone, like, half of these, so, like, I'm back. 00:00:27.000 --> 00:00:28.000 Yes. 00:00:28.000 --> 00:00:30.000 All right, let's get right into it. 00:00:30.000 --> 00:00:32.000 Let's do it. People got questions, Joe. 00:00:32.000 --> 00:00:40.000 Yes, we had a bunch of people submitted questions ahead of time. That was great. So we've got a bunch queued up. 00:00:40.000 --> 00:00:58.000 Also, you know, go ahead and submit questions now. You can do it through the… through Zoom. We've got Hannah, who is kind of doing the run of show. She's going to be in the chat if you have any questions, but drop your questions in the chat or through the Q&A functionality 00:00:58.000 --> 00:01:02.000 And I think we're going to just jump right in because we have so much to talk about. 00:01:02.000 --> 00:01:17.000 So here's what we're going to focus on today. I've got a few little stories I want to start with, a few little thought starters. We're going to spend the bulk of the time on the AMA though. So again, if you've got questions, get them in 00:01:17.000 --> 00:01:34.000 And we'll be happy to answer that. So I'm going to start off talking about the power of being believed. Many of you know, many of you have heard us talk a lot about this, the value of AI search and the way to think about AI search, we think, is with this framework 00:01:34.000 --> 00:01:51.000 Be seen, be believed, be chosen. I am increasingly thinking there is more than one way to win in Geo, and I think we're going to add this as an industry in a really kind of micro focused way. I think we need to be able to step back and look at the 00:01:51.000 --> 00:01:56.000 Other ways that there are surfaces in which we can drive value through this channel 00:01:56.000 --> 00:02:02.000 So I'm going to talk about a little test I just ran to help illustrate that point 00:02:02.000 --> 00:02:20.000 We're also going to talk about the KPIs we trust. Will had a fantastic blog post a couple of weeks ago. Hannah, if you can drop that in the chat. That is a really good resource. It should get your gears turning on KPIs for AI search and probably some metrics you have not yet thought of 00:02:20.000 --> 00:02:38.000 I've also got some slides. First of all, I'll just call out, even Will and I are not totally aligned. That's how loose, that's how open this space is. There is not, if you're looking for objectively true answers, we are not in a period where we have those yet 00:02:38.000 --> 00:02:53.000 And so what I try to encourage everybody to do is think about how success is measured with your organization today and try to apply metrics that already fit that mold where possible. So we're going to give you kind of a wide variety of KPIs and metrics 00:02:53.000 --> 00:02:56.000 And then we'll get into the AMA 00:02:56.000 --> 00:03:17.000 All right, so I ran this little test and it's in part because I think I get about 15 text messages from Tallarico's campaign a day asking for money. And I was like, you know what? I don't want to give you money, but what else can I do? What else can I dig into? How can I better understand what's happening with this very influential Senate race 00:03:17.000 --> 00:03:24.000 will stay bipartisan here. I'll just say Election Day is coming up. Make sure you get out and vote. Very important. 00:03:24.000 --> 00:03:39.000 But what I did is I wanted to see for the stated positions of the candidates in this race, how often was ChatGPT giving an accurate representation of their position 00:03:39.000 --> 00:03:49.000 And if you're like me, you start to think about it and you start to get adjective, right? Because it's like, oh no, what am I going to find? Is there misinformation? I was really pleasantly surprised, though. 00:03:49.000 --> 00:04:07.000 So what we did, and I'll walk you through this, we ran our brand canon, we pulled down all of these facts from the official campaign information about these candidates positions, and the facts all held up. ChatGPT had 100% accuracy in matching both 00:04:07.000 --> 00:04:15.000 What the press coverage said, what the official campaign position said, and what the ChatGPT response was. 00:04:15.000 --> 00:04:20.000 Why am I sharing this? We're not in politics, we're in marketing. Okay, here's the here's the rub. 00:04:20.000 --> 00:04:35.000 This brand canon skill that was made on our team by John Lovett. So shout out to John Lovett, did such a fantastic job. And I want to make sure folks understand the power of something like this. A skill is a transferable 00:04:35.000 --> 00:04:42.000 you know, think of it as a Markdown file or a piece of text. It's instructions that tells an agent what to do 00:04:42.000 --> 00:04:59.000 Right now, all the major models use this technology, right? So you can create a skill, you can use it on ChatGPT, you can use it on Claude, you can use it on Gemini. And what this skill did is go off and do the research, look for the official documentation 00:04:59.000 --> 00:05:08.000 Look for what is objectively true. Put together that brand canon and made it really easy for me to pull down the data I then pulled from Peak 00:05:08.000 --> 00:05:24.000 to tell the story of, you know, are these responses accurate? So again, I don't think, you know, maybe we have a couple, but most of us are not working in politics, we're working in marketing. The application here, though, works for all of your brand websites 00:05:24.000 --> 00:05:44.000 So if you don't have something that exists today that tells the story of how you are being believed, I think this is a huge opportunity. We'll share out the instructions for this skill, but it's something you can build yourself as well and make sure it fits within your data sources. 00:05:44.000 --> 00:06:00.000 What this bill does pulls official sources. It builds this canon. It writes the claims, and then AI goes through and scores it. And it works really well. It's a great use of AI because it's looking for these objectively true or false statements. 00:06:00.000 --> 00:06:16.000 Okay, so here's the interesting thing I found. So again, ChatGPT got everything right, so that was good. But it did not cite Tallarico's official page on data centers. What is your position on data centers? Again, it got his position right 00:06:16.000 --> 00:06:19.000 But it didn't cite his official page 00:06:19.000 --> 00:06:36.000 That's interesting and putting on my SEO hat, I started to explore why I think it is because the data center position was not on a page specifically optimized for data centers. It fell within his issues that include 00:06:36.000 --> 00:06:45.000 issues related to social media, AI technology that could be why it didn't pull for the official from the official source. 00:06:45.000 --> 00:07:01.000 The fact that he has this plan and this documentation is very interesting. ChatGPT used none of it. Again, luckily, ChatGPT still got the information correct. It was still accurate information, but as a marketer. 00:07:01.000 --> 00:07:22.000 Your gears might be turning of, like, wow, what are the implications of this for me? And so, this is all early days. I would be curious if you drop in the chat if you'd like to see data studies like this, if there's a different cut at the data, the next thing I'm running with Peak is the same analysis, but I want to see what Grok says, and I'm a little worried about what that may or may not hold. 00:07:22.000 --> 00:07:40.000 But in terms of testing ideas, some of the inspiration this gave me is, number one, give big questions their own page. And this is kind of an SEO fundamental thing, right? We think in terms of pages. Also, publish what you announce. If you have information, if you're going to market with something 00:07:40.000 --> 00:07:58.000 But you have a press release, and you don't have anything corresponding on the website. Consider getting something corresponding on the website. You know, we talk about receipts a lot. I think Will will probably pepper that into a lot of his answers. And the receipts that we own and we hold up, I think, are critically important 00:07:58.000 --> 00:08:14.000 Also, as you are checking, as you are tracking prompts about your brand, see what sources AI is citing. Is there an opportunity to build rapport, to make a connection with a specific publication, with a specific journalist 00:08:14.000 --> 00:08:22.000 Are you finding websites that you had no idea would be influential to your audience? That's really good to know too. 00:08:22.000 --> 00:08:41.000 All right, so I want to switch gears here, talk a little bit about the KPIs we trust. This should be kind of a springboard for the conversation in AMA we have from here. If you saw us at MACON, you've seen some of this information. But it's the kind of thing that I really think bears repeating until we're all on the same page. 00:08:41.000 --> 00:08:59.000 So my background, my thesis here is that geo looks like SEO. So we measure it like SEO, and by the way, nobody's driving a ton of traffic through this channel. So it's like, maybe let's just wait and see when it becomes bigger. We optimized it like SEO. We focus just on our our own content 00:08:59.000 --> 00:09:02.000 And we staff it like SEO 00:09:02.000 --> 00:09:14.000 And the job of an SEO today is has really blossomed from what it was a year, two, three years ago. There are so many more surfaces to worry about, and even if you're not worried about other surfaces 00:09:14.000 --> 00:09:32.000 Google itself has evolved so much and is a very challenging channel to to operate within now. So the idea that you can just give this to your SEO team and have them figure it out, that's a stretch. That takes a really well resourced and a really capable team 00:09:32.000 --> 00:09:51.000 My PUV with Geo and the opportunity with AI search is that it is not SEO and we should not be thinking about it and measuring it like SEO. And the statement I would give to your board, to your CMO, if they push back is that at the end of the day, this is your brand on a screen in front of your customer 00:09:51.000 --> 00:09:59.000 And so why is it worth you investing and paying by impression for a YouTube pre-roll ad 00:09:59.000 --> 00:10:05.000 But if an LLM recommends you, that's worth nothing. To me, that that makes no sense, right? 00:10:05.000 --> 00:10:22.000 So I'll go through an exercise here, a prompt sequence that I think will ground us in the type of data we are looking for when we're talking about measuring visibility and success in this channel. So let's imagine we are the buyer here, we're running marketing 00:10:22.000 --> 00:10:37.000 at a 40-person company, we're B2B, we're looking for the best CRM for a small to mid-sized business. We want to be AI native, not AI bolted on. We don't want it to cost more than $100 per user per month. 00:10:37.000 --> 00:10:55.000 What happens here that is very interesting in this first response is that the model recommends Adio and HubSpot. And if you are just tracking, let's say we work for Zoho and we're just tracking mentions, we see in our AI search tracker, okay, we're mentioned 00:10:55.000 --> 00:11:08.000 Right? The brand is mentioned here, that's good. But it's in the context of, hey, I would skip this. I treat it cautiously, because yes, the AI, it's cheap, it's capable, but their AI is bolted on. 00:11:08.000 --> 00:11:24.000 I don't know anything about these brands, but I think if I worked at one of these companies, I would not want my brand described this way. So all of a sudden, like, the spidey senses should be going off of, wait, there's a lot more to the story than what is appearing in my rank tracker. 00:11:24.000 --> 00:11:45.000 The other element here is how did it choose Adio and HubSpot right away? How did it make this decision? In the past, we saw these models doing web search and doing queries that said things like, what is the best CRM for a 40 person team? What is the best CRM for an AI native team 00:11:45.000 --> 00:12:01.000 Because of all the listicle spam, we are seeing less of those queries in query fanout. And here you see specific brands are mentioned. This gets into this idea that Will has been talking about a lot that you want to be one of these brands 00:12:01.000 --> 00:12:23.000 where there is this hardwired connection relating your brand with the product or service you sell. And so even though Clarify and day.ai and lightfield.app were not recommended and mentioned in the wrong context, this to me is still a win. This is a good sign, because you can do something with this 00:12:23.000 --> 00:12:40.000 If you do deserve to rank for this query. By the way, if you're wondering how do I get this data? There's a few different sources for it. I like Fan Out Fox, which is a Chrome add-on, a Chrome extension, keywords everywhere also has a Chrome extension 00:12:40.000 --> 00:12:57.000 Also, this data is generally available via API, so if you're trying to put together more of a recurring poll, work with your instance of Claude or ChatGPT to try to build out a prototype and get something working there. 00:12:57.000 --> 00:13:07.000 All right, so that's just about being seen, right? And there's value in that. What about being believed, which is the example I was going through with the Tallarico campaign. So 00:13:07.000 --> 00:13:25.000 As we prompt more in this conversation, we start to see another side of the story. And yes, HubSpot has good ratings, at least on G2, but on Trustpilot, not so much. And all of a sudden, these anecdotes are being pulled in to the experience 00:13:25.000 --> 00:13:41.000 What really blows me away with this response is how it's one single analysis and one single reviewer is being pulled into the response. Note there's 16,000 reviews on G2 alone 00:13:41.000 --> 00:13:51.000 And so maybe these one-off reviews are representative of the bulk of negative sentiment on G2 or on Trustpilot 00:13:51.000 --> 00:14:06.000 I might doubt that, though, and there's nothing within this response that suggests that this is, you know, a statistically significant complaint. If you're a HubSpot user, you might look at this and say, yeah, that's pretty that's a bit spot on. But 00:14:06.000 --> 00:14:23.000 The point here is that there are, you see all the chips, all the citation chips here. None of these are citing HubSpot, right? And so there's a missed opportunity here potentially for HubSpot to get involved 00:14:23.000 --> 00:14:36.000 Write the objective truth about their brand, about their services, about their weaknesses, right? And be able to be cited and influence this answer or just rely on what the reviews are saying. 00:14:36.000 --> 00:14:52.000 All right. And then the last one is were you chosen? And within this prompt, I go back and forth and I say, well, my CFO is interested in Salesforce. Should I consider that? And finally, HubSpot gets the recommendation. This to me is inherently 00:14:52.000 --> 00:15:12.000 Hugely valuable. We know, despite what we say about how much we trust or don't trust AI, when the AI tells us something is the right choice, more often than not, we are selecting that. And so we see that the AI is telling me my pick is HubSpot. Start with these starter seats, and here's what you should look at 00:15:12.000 --> 00:15:21.000 That's a win for HubSpot. And remember, Adio was the first brand that the AI started to consider for this query 00:15:21.000 --> 00:15:25.000 They're not in the mix any anymore at all, right? 00:15:25.000 --> 00:15:31.000 All right, so the KPIs, I think, map back here and we'll talk about this in the AMA 00:15:31.000 --> 00:15:50.000 These are six KPIs, six metrics that your exec team, your board, your leadership is probably already aware of and maybe is already measuring other marketing or advertising investments in this way. Share of voice. How often is the AI naming you for the prompts that matter to your audience 00:15:50.000 --> 00:16:08.000 AI shortlist rate, this is what you would get from those query fanouts. How often is your brand hard-coded right into the response? Not only is that a great metric to look at for diagnostics, that to me suggests that we are winning in some degree, right? 00:16:08.000 --> 00:16:25.000 Then we have the be-believed metrics, and I think these kind of map to frequency when you think about MediaBuying. So what about message pull-through? Is AI telling your story? First year, one of the things we're tracking is how often are these models pulling through the language 00:16:25.000 --> 00:16:43.000 of be seen, be believed, and be chosen. When somebody asks about Seer and our approach to geo, that's an interesting statement for me, because we've gotten a really good response from our audience that, yes, this framework is helpful, it makes sense, it helps position us as a thought leader in this space 00:16:43.000 --> 00:16:53.000 So the more we are associated with that messaging, the better off for our our pipeline growth. At least that is the hypothesis 00:16:53.000 --> 00:17:05.000 And then brand accuracy, this is what I showed you all with the Telarico example. It's got to start with a brand canon, but that can be created a little bit more easily than maybe you might have imagined. 00:17:05.000 --> 00:17:20.000 All right, then we have sales proxy. So at the end of the day, we're trying to acquire customers, right? And no matter how difficult the measurement is, all of this activity has to result in more people buying the product you sell or the service you offer 00:17:20.000 --> 00:17:39.000 And so pipeline contribution, however you're currently measuring this for the impact and influence of awareness channels, consider how you can use the same formula for geo recommendation rate. This is something that I have not seen in any tools out of the box, any AI search tracking tools 00:17:39.000 --> 00:17:58.000 But if you choose a provider that has an MCP, you can easily connect to your data to Claude, to ChatGPT, or to Copilot, and you can pull out the recommendation rate. AI is very good at being able to do that and model the data as you need. 00:17:58.000 --> 00:18:13.000 All right, so we always like to leave you all with a little bit of homework. Here's the three things that we recommend you do this month. First of all, write your brand canon. And it's an interesting exercise to have AI go off on the internet, go to official documentation, official sources 00:18:13.000 --> 00:18:30.000 And do that itself, because you might find there's some things that got wrong, which will prompt you to update those official sources. Once you have that brand canon, count how often AI is citing your own site and check your highest stakes facts daily 00:18:30.000 --> 00:18:45.000 So whether this is a pricing conversation, an eligibility, legal claims, what are the serious errors that are incorrect and are going to cause you to lose customers that you otherwise were this close to acquiring 00:18:45.000 --> 00:18:51.000 Make sure you've got a system in place so you can detect when that's happening. 00:18:51.000 --> 00:19:00.000 All right, so we breezed through that in 20 minutes. That means we have about 20, 25 minutes to get to all of your questions. 00:19:00.000 --> 00:19:01.000 Will 00:19:01.000 --> 00:19:02.000 Yep. 00:19:02.000 --> 00:19:05.000 I'm going to I'm gonna start with you 00:19:05.000 --> 00:19:08.000 All right. I got a question on prompt tracking. 00:19:08.000 --> 00:19:16.000 So what guidance would you give someone who's just setting up AI search visibility tracking for the first time 00:19:16.000 --> 00:19:28.000 Especially as it relates to selecting which prompts to track 00:19:28.000 --> 00:19:38.000 Okay, I would ground everything in actual customer conversations. So the first thing I would say, especially if you're B2B, if you're recording sales calls 00:19:38.000 --> 00:19:50.000 take those sales calls, and there's a few places where humans leave clues on what kind of words they might leave in prompts, and they're largely ignored by everyone in the industry. The first one is your, if you get sales calls 00:19:50.000 --> 00:20:06.000 Great, right? Like, sales calls, good, because you can actually check your prompt list again that you made up or that you were given against what your customers are actually saying their pain points are. First thing. Second thing I would do is paid search data. So paid search data, when you pull in a search term from paid search. 00:20:06.000 --> 00:20:21.000 You get to see words that the customer was using to try to solve their own problem in their language. So I would bring that data in. You also have the grounding queries from Bing Webmaster Tools, which also gets you a feel 00:20:21.000 --> 00:20:36.000 on like, how is Bing turning prompts into searches, which I think is very helpful to try to understand people are putting whatever they want in. Bing is like Bing's webmaster tools is like truncating that down into a series of searches, which helps you to understand that 00:20:36.000 --> 00:20:45.000 And the other thing I would do, there's a few other things I would do. One, if somebody came to my site, if I was e-com, and somebody came to my site through an LLM 00:20:45.000 --> 00:21:00.000 and converted on my thank you page or the email that I sent out, I would try to incentivize those people to be like, do you want to join, do you want 20% off your next order? Cool, I need you to join my, group, whatever that is, that tells us, like, what were you searching for before you did this? 00:21:00.000 --> 00:21:18.000 Because now you can pull the same information that our B2B friends can with recordings, but you don't have that. So that's one of the things that I would do there. And if you don't have a dropdown on your website that says, like, how did you hear about us? Be careful. I checked our own site. We had, like, 10 of them, and I think that if you put too many in trying to get too granular, you're gonna ruin your data 00:21:18.000 --> 00:21:37.000 But if you only have a few, because you really are just trying to figure out, are they using AI or not? So just be like, did you use Google AI or other, right? If you can do that, then you start to get people that's identified that way, and you can also ask them to send over prompts and kind of walk you through how their decision process was. In early days 00:21:37.000 --> 00:21:56.000 When we put it on our website, it sent me an email, and I would go and I'd email those people and be like, yo, I can see that you said ChatGPT, I'll give you 10 or 15 minutes of my time. Will you also be willing to walk me through all your prompts and whatnot? And it gave me a plethora of information that, a keyword analysis kind of tool or some made-up synthetic prompts never would have given me. 00:21:56.000 --> 00:21:59.000 Really good stuff. I love that. 00:21:59.000 --> 00:22:17.000 So the other thing to keep an eye out for is Andrea Haley from our team has been conducting these AI moderated user interviews all year. I think she's done about a dozen at this point. I want to say she has over a thousand prompts that she has collected from real customers 00:22:17.000 --> 00:22:34.000 across, again, maybe a dozen different types of businesses. We are finding that AI is very good at taking that raw list of prompts and turning it into a template, kind of a Mad Lib style template, right? And everybody searches differently. Everybody prompts differently. 00:22:34.000 --> 00:22:51.000 But we're trying to create prompts that are directionally accurate, and everything will just describe the sales calls. First of all, you can get some of those Mad Lib templates that way, but you can then use the paid search data, use the sales call data, use the go-to-marketing messaging that you've created 00:22:51.000 --> 00:23:02.000 and populate those Mad Libs prompts for pretty directionally accurate view of what your visibility looks like, depending on how people are prompting. 00:23:02.000 --> 00:23:03.000 Oh my God. 00:23:03.000 --> 00:23:07.000 Alisa, I gotta share with people what's going on on my screen. 00:23:07.000 --> 00:23:14.000 Real quick. So something I think that we should all be challenging ourselves to do is what's running for me automatically while I'm here. 00:23:14.000 --> 00:23:34.000 Right? Like, what is running for me automatically and making me better at marketing while I'm here? So just so you guys know, there's two things running on my screen behind the chat we're having. So I'm chatting with you guys. Two things are happening. One, I just got a notification that one of my bots has gotten 128 gigs of files off of Slack, of videos that I've left the team internally 00:23:34.000 --> 00:23:46.000 It is currently pulling down the transcripts from every one of those internal Slacks, looking for the 10 that they can automatically cut up into video. That is happening, like, while I'm here with you guys, right? So I want you guys to think about 00:23:46.000 --> 00:24:02.000 When everybody talks about agents, you know, I tend to think, how can I be doing one thing completely locked in, and other things that I see as valuable are also running in the background. So, I'll chime back in with the second one that's running and share that with you guys in a bit, but I wanted to just stop and be like, that is, like, this new world we're in 00:24:02.000 --> 00:24:13.000 It is the best way to think about it is how can I have my full attention in one place where I'm most needed, but then also have something doing something else for me that is also really helpful that I don't need to manually be checking in on and doing myself 00:24:13.000 --> 00:24:33.000 Yes. Yep. Yeah, totally agree. I mean, I think that's what this world is becoming. And when I say this world, I mean digital marketing. There's, yes, we're all in on agents and there's a lot agents can do, but I think the value is in these systems that your agents can help create, your agents can then run autonomously 00:24:33.000 --> 00:24:51.000 Because in the past we had such concise dashboards and good sets of information we could use to say, okay, here's where the opportunity is, here's where the movement is. Now it is so all over the place, right? There's so many surfaces to track. So creating these systems, again, created by humans 00:24:51.000 --> 00:25:14.000 run autonomously by agents, you know, we're it's something we're still chipping away at at Seer and Will is, you know, driving the forefront of that. But I think as we get into 2027, this is going to be more of a, you got to have this set up to compete in this space as opposed to, okay, this is something that sets you apart 00:25:14.000 --> 00:25:35.000 All right, next question I have, how are you tracking your recommendations from AI tools? How do we measure the impact of improved AI search visibility? And so I'll answer the latter part first. I always try to take a step back with this question and think about what is the way in which we are trying to use this channel to win 00:25:35.000 --> 00:25:54.000 Because I think it's very easy to fall into the trap of, okay, how do I measure the traffic? How do I measure the revenue that's coming from this channel? Which you can do, right? Google analytics now makes it easier than ever to do this. In the past, you needed custom groupings. Now, for the most part, what is trackable is coming into Google Analytics 00:25:54.000 --> 00:26:09.000 The asterisk I would put there is how much is trackable. I don't know, because the data story of the last couple of months is a little bit all over the place. And it's leading a lot of us to think that a lot of this data is obfuscated, especially at the traffic level data. 00:26:09.000 --> 00:26:26.000 So, when I think about how do I measure the impact of improved AI search visibility, I wonder, am I trying to be seen? And if so, I'm looking at share of mentions for the prompts that I'm tracking. I'm looking at how often my brand is appearing in query fan out 00:26:26.000 --> 00:26:44.000 If I'm trying to be believed, I might be looking at all branded prompts, right? And that's something that's a little uncharted territory for us as SEOs. You don't necessarily have a strategy around branded keywords because you just create a page on your site and you're going to rank for your brand, right? 00:26:44.000 --> 00:26:47.000 You know, 9.5 times out of 10 00:26:47.000 --> 00:27:02.000 That's not the case with AI search, though. And we know these models are asking or these users are asking models, should I trust this company? Is this a legit company? By the way, I'll plug that ChatGPT released what I think is their third version 00:27:02.000 --> 00:27:09.000 of data that how consumers are prompting ChatGPT 00:27:09.000 --> 00:27:35.000 1 in 5 conversations on ChatGPT now is an information seeking question, and that's the highest we have seen. I would expect it's only going to get higher as long as the experience continues to be better than Google. And then lastly, how do we measure the impact of improved search visibility on being chosen? I think you have to be able to take a step back from your pipeline and look at overall, am I driving more leads? Am I driving more sales 00:27:35.000 --> 00:27:51.000 Why is that, right? And I think we try to take credit based on, if not the last click, some kind of attribution window, which really obfuscates some of the nuance of how conversions are actually created 00:27:51.000 --> 00:28:06.000 But I think we can look at, hey, we had this increase to this specific category in terms of share of mentions, and then we started to see more pipeline for that product category. So we got to get used to the fact that these we're not going to have a straight line 00:28:06.000 --> 00:28:22.000 You know, we still do today, but very soon I think we're going to be in a world where clickstream is so unreliable, just like we did one day with, you know, 10 years ago when we had keyword level data in Google Analytics that told us exactly 00:28:22.000 --> 00:28:31.000 how much revenue we got per click of the keyword, we're going to have to take a step back and be like, okay, let's look at this as search as a whole 00:28:31.000 --> 00:28:32.000 Go ahead, Will 00:28:32.000 --> 00:28:34.000 You know, it's funny, I 00:28:34.000 --> 00:28:44.000 So I just want to be honest with you guys about how I run my business, which insulates me from some of the ways that you guys have to run yours and the organizations that you're in. So with that caveat 00:28:44.000 --> 00:28:46.000 I don't believe in attribution at all. 00:28:46.000 --> 00:29:03.000 Like, I'm so close to the data. Like I'm looking at Muse coming in and I'm trying to identify it with Cloudflare and be like, is that a muse or not? And I'm having team members go out and run specific tests for me to understand these agentic things. And I think the closer you get to trying to figure out what data is there, I think 00:29:03.000 --> 00:29:05.000 The direct line attribution 00:29:05.000 --> 00:29:16.000 We've gotten so used to that direct line, pipe, pipe, pipe, drop, that now you're going to have to start to have some educated guesses that you can defend 00:29:16.000 --> 00:29:22.000 to a CMO or a CFO, but you're like, this is the reality of the data that I have 00:29:22.000 --> 00:29:25.000 And we're gonna have to get really good at communicating 00:29:25.000 --> 00:29:37.000 The fact that, yes, you have seen search for this long, and it has always given you these things, but now you don't have that. So, for instance, for me, my number one metric I look at is organic social 00:29:37.000 --> 00:29:42.000 It's been the number one metric I've looked at for the last 3 years. In a B2B business like mine 00:29:42.000 --> 00:29:43.000 If 00:29:43.000 --> 00:30:01.000 If I can't win over actual humans, and they don't see what we're writing, and be like, that's good, I want to show up to that webinar because I've been watching what you guys have been doing. Like, to me, all the spillover that is going to grow our business is gonna come from us having the discipline to say, if we write things and put them out on social and they don't perform 00:30:01.000 --> 00:30:18.000 It is a tremendously great signal that like people kind of are like not feeling this great, let's pivot and move on. Let's pivot and then move on. And if somebody were to challenge me on that, I'd be like, well, Will, what's the ROI of like this one post that got 4,000 views in a week? I'd be like, I don't know 00:30:18.000 --> 00:30:37.000 But it's thousands of people that loved it. It got us covered by a PR, it got us covered in this place, and it got us some people to sign up for the newsletter. Now, how many people then came in direct that were actually this and that? I think that's a fool's errand. I think you gotta be really careful of believing that these things are super trackable, because then you're gonna get all the way to the bottom 00:30:37.000 --> 00:30:49.000 And spend hours or weeks or months, and then somebody's gonna be like, but what about this? And then you're like, so what about the fact that cookies get deleted after seven days on Apple? Oh, , now I gotta go back and think through my whole 00:30:49.000 --> 00:30:51.000 You know, so it's just something I would caution. 00:30:51.000 --> 00:30:53.000 I love it. 00:30:53.000 --> 00:31:08.000 All right, I'm gonna, I've got another question. And this is a David versus Goliath competitor situation. So I'm going to read the question. I'm going to give an answer and then I'm going to hand it over to you to weigh in on as well. 00:31:08.000 --> 00:31:30.000 So the question is, our brand has dedicated a lot of time and effort into creating content that matters to our audience, and yet we are still second fiddle to the leader in our field. They've been around since the early 2000s while we launched in the early 2010s. What are some ways we can get the needle moving and close the gap, especially when it comes to AI recommending our brand over theirs 00:31:30.000 --> 00:31:47.000 And I immediately related to this question because at Seer we compete with other agencies that are private equity funded or venture capital backed, and it is really annoying. It is unfair, right? It is an unfair advantage in many ways 00:31:47.000 --> 00:32:06.000 And this is not Will's ethos, right? Like nobody's nothing's fair. The world isn't fair. But from where I am sitting, it feels unfair to go up against people who might have five times or 10 times the staff we do. How do you compete with them? This is the fun part of marketing, though, and this is the fun part of art too. It's the constrictions and the constraints, right 00:32:06.000 --> 00:32:08.000 Yes 00:32:08.000 --> 00:32:09.000 Tell them. 00:32:09.000 --> 00:32:23.000 So what we have done at Seer and again, we have a very small but mighty team, especially when it comes to how we invest in go-to-market and marketing. What we have done the past few years is 00:32:23.000 --> 00:32:39.000 share unique data, and on our team, Tracy McDonald is the godmother of this for us. We have long kind of hoovered up all of this all of this search data right 00:32:39.000 --> 00:32:56.000 about our where our clients are ranking, where their competitors are ranking. But most interestingly, how search features are rising or falling, and we're getting to a point now within this industry that this data is hugely valuable 00:32:56.000 --> 00:33:15.000 The big providers are offering less and less data to make decisions by, and it's part of our ethos to share everything we're learning. So I know other agencies may not have this data and they're strategically saying, we are not going to share this data because we don't want anybody else to know this. We want this to be a competitive advantage. 00:33:15.000 --> 00:33:32.000 That's not our ethos, right? We're like, hey, let's share this out, and whether you work with us or not, we can all learn together. Over the past few years, there's been a lot of good stuff happening here, but this, I would say, is one of the things I'm most proud of. The amount of coverage, the amount of 00:33:32.000 --> 00:33:49.000 visibility we have gained by sharing this data, by updating this data. It's really changed our engineering roadmap, and we've said, hey, we've got to be able to enable this more easily, because people find this very valuable in the hypothesis is, if we keep creating value 00:33:49.000 --> 00:34:01.000 The leads will come, the businesses we want to work with are pipeline is as strong as it's ever been. And some of the brands that hit it, I'm like, this is a marketer's dream. I'd love to work with this company or that company 00:34:01.000 --> 00:34:18.000 So given that background, Will, and you can't say unique research, what do you say, what would you say to this person asking this question? 00:34:18.000 --> 00:34:21.000 Major brands 00:34:21.000 --> 00:34:25.000 If you've got a major branded competitor 00:34:25.000 --> 00:34:29.000 And they've been out there for a lot longer than you have. 00:34:29.000 --> 00:34:32.000 I don't have a good answer for you other than patience. 00:34:32.000 --> 00:34:38.000 Like, well, one, find the niches where 00:34:38.000 --> 00:34:39.000 Slipping 00:34:39.000 --> 00:34:46.000 Sorry, let me let my brain go on this a bit, because I don't like that answer for you. So I would… if I was in a place where I could get reviews 00:34:46.000 --> 00:34:50.000 I would start to see what is the negative reviews typically saying? 00:34:50.000 --> 00:35:03.000 Because the AI is probably going to pick up on those negative reviews. Great. So now that the AI is… and then I would be running my prompt tracker and whatnot to see, are they mentioning the things that are negative in those reviews? And I bet you, as Elisa's already shown you 00:35:03.000 --> 00:35:18.000 If you see enough negative reviews on something, it starts being like, I don't know about them for that, right? It was one negative review, and I've been the victim of one negative review affecting 40% of my prompts. So it doesn't take much. So I would use that as intel to be like, this is where they're probably weak 00:35:18.000 --> 00:35:20.000 And then the harder part is 00:35:20.000 --> 00:35:25.000 Do you have the ability to influence your organization in a way 00:35:25.000 --> 00:35:28.000 Where you're like, all right, product team 00:35:28.000 --> 00:35:42.000 We gotta make this part of our product better, because I'm seeing that for a year and a half, people have been saying this part of their product sucks, and does that fit our ethos? Does it fit our roadmap? Does it fit where we want to go? But if my boss is asking me, yo 00:35:42.000 --> 00:36:01.000 How are we ever going to eventually beat these guys? It's like you got to be on your fucking shit, man. You got to be like, I'm watching them like a hawk. Like for me, I would be everywhere that they are. I would look at where they have gaps in their go to market, right? So I would have an agent going out and crawling their meta ads library and then I would compare it to their landing pages and compare it to this and compare it to that 00:36:01.000 --> 00:36:16.000 And I'd be like, where are they slipping? It might be worth bidding on their brand name so you can get the data and the search terms of how are people exactly searching for words around their brand, right? It's like those are the kind of things that I would be looking for 00:36:16.000 --> 00:36:17.000 And then 00:36:17.000 --> 00:36:34.000 you know, I would hope that I can find a crack or two to try to win, against them. The other thing I will say is this, a lot of the research shows Kevin Indyk has just tremendous research on these things, so you don't want… you know how people have paid newsletters on Substack, and most of them ain't worth 00:36:34.000 --> 00:36:41.000 Kevin's is worth actually spending money on. So if you've ever been like, I don't know, it is worth it. 00:36:41.000 --> 00:36:52.000 But what you find is that, like, people will reorder the mentally, they will reorder the answer that comes back from AI based on the brands they already know 00:36:52.000 --> 00:36:59.000 So it's like, you might be mentioned third or fourth in a list, but if they're like, but I know them, so then what is that, what is, but I know them come from 00:36:59.000 --> 00:37:21.000 social, right? But I know them, I've been seeing them around, they're showing up in my LinkedIn groups, they're showing up in my private WhatsApp groups, I heard of them, right? And what do you do in a world where all the answers doesn't include a brand you know? Like, if you were like, I'm looking for a luxury hotel in Philadelphia, right? And if none of them were the Ritz-Carlton or the Four Seasons, you might be like, but what about the Four Seasons? Well, why did you put that in? Because of things that weren't content marketing 00:37:21.000 --> 00:37:26.000 It was building the brand over time. So that was what I would… that's what I would recommend. 00:37:26.000 --> 00:37:30.000 Hopefully that's helpful. If not, just ping me on LinkedIn and DM me and I'll help you out further. 00:37:30.000 --> 00:37:51.000 Yeah, the only thing I'll add to that, Will, is I'll tie that to what you were mentioning earlier because I think some people hear that you don't believe in attribution and it's like, all right, get rid of all of it. It's like, no, we're still data-driven. We're still data driven, but we no longer have faith in the macro view, which means you've got to have a micro view 00:37:51.000 --> 00:38:07.000 And what Will is describing here, these are dashboards, these are reports that do not exist, but that is the power, that is the beauty of AI to be able to build whatever you need. And you don't need to set up a whole data pipeline. You know, it does need to be this arduous thing 00:38:07.000 --> 00:38:22.000 You are trying to cheat at Jenga. That is the mentality. And you're hitting every brick, and it doesn't work, and it doesn't work, and it doesn't work. And then all of a sudden you get an easy one. It's like, this is the one I'm going to go after. 00:38:22.000 --> 00:38:36.000 You've got to be able to look at these little pockets of data to understand what is the thing I'm going to go after. What do I believe will happen if I'm successful? How do I measure that? How do I keep investing in that thread if it seems like it's working? 00:38:36.000 --> 00:38:53.000 Yeah, well, Lisa, just one follow-up here. Cassandra Santa Maria asked, I'm in the opposite situation. We've been around longer, but they are owned by private equity and can't compete with that type of money. So real talk, we just had a client get acquired by a very aggressive. So some of these private equity firms look for 00:38:53.000 --> 00:38:56.000 older school companies 00:38:56.000 --> 00:39:11.000 that did not adopt technology fast enough. They had this big brand, but they've slowly fallen behind. And then these people have come in and the messages that I've seen coming back to our team would make you want to vomit the way that these people treat people and the way that sometimes they act 00:39:11.000 --> 00:39:23.000 So that is an asset for you, right? Like, when I was small, and I didn't have the money to compete with the big players, I was like, it's like exactly what Elisa said. I can out-care my competitors. A private equity firm 00:39:23.000 --> 00:39:26.000 Investing in a company is built for one fucking purpose 00:39:26.000 --> 00:39:28.000 Get Daddy a third yacht 00:39:28.000 --> 00:39:44.000 Okay? That's all it's built for. How can I get more so I can brag to more people that also do my job, and how can I make your yacht the size of my dinghy that goes to my yacht, right? So, real , real conversation that I heard. So. 00:39:44.000 --> 00:39:49.000 Because that's how they think, they will not invest in caring… out caring 00:39:49.000 --> 00:40:00.000 For the customer. The problem is, is older school companies get in their own way because they're like, I don't want to do the AI thing. You know, and they get all like that, and it opens a gap. Here's what I believe. 00:40:00.000 --> 00:40:08.000 If private equity firms are beating you out in an industry you've been in forever, you let them do it 00:40:08.000 --> 00:40:21.000 Because you're slow to adopt things, they saw the market and they went, we're going to buy this firm who's slow to adopt, pour a bunch of money in, speed them up, and just completely destroy everybody else who's also not as fast 00:40:21.000 --> 00:40:24.000 Because the one thing they don't have is patient capital 00:40:24.000 --> 00:40:31.000 They want their return in 3 years. They want their return in 4 years. So they'll just keep pushing on that company 00:40:31.000 --> 00:40:44.000 Because they gotta get that money back in 3 years or 4 years, which means there are certain things that you know are right for your customer that they will not do, because it doesn't scale. And that's the way that you ultimately win, in my opinion. 00:40:44.000 --> 00:40:47.000 Yeah, I love that. 00:40:47.000 --> 00:40:51.000 I like that line. Daddy needs a third yacht. 00:40:51.000 --> 00:41:00.000 Tell you a little bit 00:41:00.000 --> 00:41:01.000 Right. 00:41:01.000 --> 00:41:03.000 That's what they fucking are. I mean, they hit me up to try to buy our company all the time. And I'm like, I can't make my why you're going to beat me and my team up to grow so we can get some crumbs so you can freaking, you know, the money we got would be there like, yeah. 00:41:03.000 --> 00:41:04.000 Yes. 00:41:04.000 --> 00:41:05.000 Yes. 00:41:05.000 --> 00:41:08.000 Daddy got a gold anchor on his yacht. Thanks, guys. Good job, sir. I'm so glad you guys worked so hard. 00:41:08.000 --> 00:41:11.000 Yes, yes. 00:41:11.000 --> 00:41:12.000 Fuck them. 00:41:12.000 --> 00:41:29.000 All right, we got we got another good question. Let's take a right turn here. So the question is, I'm really wondering how the landscape is going to change for businesses when every consumer and client has a proactive, always-on assistant like Dot 00:41:29.000 --> 00:41:40.000 Muse, instinct. I don't even know what instinct is, so drop that one in the in the chat. Duet is now Claude's Gemini had spark. First of all, what I would say is 00:41:40.000 --> 00:41:56.000 There, I'm going to talk on both sides of my mouth because on one hand, we know intellectually that it is very likely this technology does not reach a critical mass with the audience you are targeting 00:41:56.000 --> 00:42:11.000 And I think as marketers today, on October 29th in 2026, we have to assume this is the direction things are going and think about the implications on your business. And the very first thing I would do 00:42:11.000 --> 00:42:26.000 is use this technology, see how it works. I have an extremely AI pilled team member on my team. Many of us are Nick Pryor. If you're if you're listening, this guy is, he uses all the technology as soon as it comes out 00:42:26.000 --> 00:42:44.000 And I've had some really interesting conversations with him about dots specifically and how it has changed the way he thinks about using AI for work, using AI for personal life. So these are real changes, right? It is easy to be so overwhelmed because every single day something new is coming out 00:42:44.000 --> 00:42:58.000 And often it's another drop in the bucket. This, I think, is a new bucket. This, I think we have to pay attention to as marketers. And I have long been saying we are going to have to care about meta in some degree when it comes to organic 00:42:58.000 --> 00:43:07.000 search and discovery. And they have made a liar out of me for like three years, but now, now they matter, right? 00:43:07.000 --> 00:43:28.000 So I would say use the tools. I have started to use dots and it's got me thinking a lot differently about email marketing, because one of the first things I did with my dot is say, okay, my personal Gmail is a mess. I never see anything in there that I actually need to see. I get like five different emails for every one 00:43:28.000 --> 00:43:50.000 kid notification. Can you give me like the daily digest, right? And now all of a sudden it's giving me this daily digest and I'm like, oh, this is good. Like, you know, there's this one fantasy football newsletter I always want to see and it's giving it to me proactively. That's good. It also starts to pull in some stuff like, hey, here's this is something that I thought you might like from the New York Times cooking newsletter 00:43:50.000 --> 00:44:05.000 I'm like, alright, maybe I like it, maybe I don't. How does that change in email marketers strategy? Because now you have an opportunity to kind of talk to an agent right and say, Hey, if you if you're an agent working on behalf of a client 00:44:05.000 --> 00:44:20.000 And you're looking for important or notable information to make sure they're aware of, if they've got a busy week, maybe they should check out the weeknight meal ideas 00:44:20.000 --> 00:44:21.000 I like that 00:44:21.000 --> 00:44:34.000 There's some interesting things we can do as marketers, and it's just fun too, right? Like this is what a lot of us got into marketing to do, and we've been in this very, you know, block and tackle world for a long time. So that is just but one angle. There's this whole world of 00:44:34.000 --> 00:45:00.000 optimization here. I've seen really interesting things about the volume of citations Muse uses to influence its responses. I believe it was Aerops just released some really interesting data here, and just visually it looked like it was like 5 times more than the normal amount of citations. I think Meta's got some stuff going on where they're trying to build out their own index. Everybody thinks they can finally take down Google 00:45:00.000 --> 00:45:06.000 All that said, Will, where is your head at when it comes to these always on assistants like Muse? 00:45:06.000 --> 00:45:09.000 So 00:45:09.000 --> 00:45:26.000 Yeah, so I, because I'm kind of anti-Facebook guy, so I really delayed on installing Muse, but because I work with some team members who are able to show me the light and because they came out first, I was like, all right, let me try this out. And, you know, like, I wake up in the morning to be 00:45:26.000 --> 00:45:39.000 Three things, I want to be… are really important for me to always try to be better at. I want to be a better dad, a better husband, and a better CEO. And if that's my… if that's my take every day, I'm thinking, how can I be better at these three things? 00:45:39.000 --> 00:45:43.000 Then anytime something makes me better at those things, I love it. 00:45:43.000 --> 00:46:01.000 And what's interesting is I put in a few things into Muse and all of a sudden it was like, hey, this Marvel thing is coming this weekend to this thing on the campus at UPenn, and your kids love Marvel, right? So it learned that, and then I'm like, that makes me a better dad, right? And then it's like, and then I'm taking my kids indoor skydiving. I have, like, date nights with them every once in a while 00:46:01.000 --> 00:46:17.000 And Muse did this, and this shocked me. I'm like, , when is Duet gonna get this good? So I have Muse, I have Claude Duet, I got into the… I got off the waitlist, and I have, Dots, and I got all them running, so I can try to understand how they work differently. 00:46:17.000 --> 00:46:20.000 And 00:46:20.000 --> 00:46:42.000 Muse is super proactive, so it knew the time slot that I wanted to have with my kids. Muse went out, crawled the website of the indoor skydiving place, checked to see if that was still available, found out that the two blocks around it were not available, and was like, you need to book that today 00:46:42.000 --> 00:46:43.000 Yeah. 00:46:43.000 --> 00:46:48.000 Because the blocks around it aren't available, and then you won't be able to do this thing you want to do with your kids. That makes me a better dad. You know, privacy, here you go, right? Now, for me, of course, because I don't like 00:46:48.000 --> 00:46:57.000 using their platform, I have now pulled mostly everything off of that. But what it inspired me to do was to build my version of it in a platform that I trust. 00:46:57.000 --> 00:47:16.000 That makes a lot of sense. That's really interesting. It would be really interesting, too, to see how aggressively all these agents violate robots.txt in terms of service, because you can imagine Meta's just like, I don't care what you're telling me I can and can't do. I need to be so impressive. I need to show something different 00:47:16.000 --> 00:47:23.000 And so maybe the other agents are like, oh, I can't access this. I can't. And it's like, no, I'll just boof myself into something else and get through. 00:47:23.000 --> 00:47:28.000 Yeah, I mean, what we found was, so we had our team members, if anybody wants to test Muse 00:47:28.000 --> 00:47:43.000 One of the things you can do… also, Jason Thompson, I can't share exactly what he's testing because I don't think he's made it public yet, but I know he's a good dude. He's testing some really interesting stuff on Muse, so, you want to give him a follow. He tested stuff that I had not even thought of, and he dropped it on me yesterday 00:47:43.000 --> 00:47:59.000 related to email marketing. So I think he's worth a follow, because whenever he does drop this, you don't want to miss it. What I did, if you want to test Muse, especially for my friends on the call on Cloudflare, super good, because Cloudflare has logs, and Cloudflare does the work to identify the AI agents for you already. 00:47:59.000 --> 00:48:00.000 So 00:48:00.000 --> 00:48:13.000 What I did is I gave individual people at Seer their own unique URL to go to, so if you want to test Muse, give everyone… not everyone, give, like, 5 people in your company the same task 00:48:13.000 --> 00:48:15.000 to use and muse. But 00:48:15.000 --> 00:48:30.000 to start from a URL that's only for them. And then you can use Cloudflare or your log file analysis and GA to be like, did they all come through the same way? Oh, in this city, for some reason, they sent you through this thing 00:48:30.000 --> 00:48:47.000 And through a proxy, this person got it differently, and that's what makes me and Elisa start being like, these guys are gonna send out these agents through a bunch of different things to try to stay not detectable, especially when Amazon blocked them all, right? So Amazon blocked a bunch of these. So you've got to start to think, I think that's the other thing 00:48:47.000 --> 00:48:55.000 About, like, if you're knee-jerk reaction is to block AI bots, it's like, if you're Amazon, you can kind of get away with that. 00:48:55.000 --> 00:49:15.000 Because they have an install base on Amazon Prime that'll choke a freaking hippo. But the rest of us who are trying to freaking, like, get a client, one, two, three at a time, for you to be like, I'm gonna block, like, these, because I don't like them, it's like, then I just won't buy from you, because I don't want the cognitive load of having to remember your brand every time I want to buy a new pair of shoes 00:49:15.000 --> 00:49:16.000 Yeah 00:49:16.000 --> 00:49:18.000 I'll give you guys Jason's LinkedIn in a second. 00:49:18.000 --> 00:49:34.000 Nice. That's a good one. All right, if anyone has a question that was sparked by anything we've been talking about, get it in in the next 30 seconds. Otherwise, this is going to be the one we close on. Will, I'm going to tee this one up for you. 00:49:34.000 --> 00:49:50.000 What advice would you offer a B2B manufacturing company on monitoring GEO? Although we seem to receive more AI conversions, maybe because we have a tech savvy buyer, a large portion of the advice that this user has seen 00:49:50.000 --> 00:49:57.000 targets B2C brands. So do you have anything specific for B2B manufacturing that comes to mind? 00:49:57.000 --> 00:50:06.000 Oh, man, this is such a broad one because manufacturing B2B manufacturing. So you're building things for businesses. 00:50:06.000 --> 00:50:08.000 B2B plus manufacturing, okay 00:50:08.000 --> 00:50:12.000 Like specs like product specs, that kind of thing. 00:50:12.000 --> 00:50:32.000 Yeah, it's like B2B. Like, so, here you go. Let's just say you build… I see, I'm still going B2C. I'm thinking of like a builder. Like you could be technically in B2B manufacturing through like 00:50:32.000 --> 00:50:33.000 Medical device 00:50:33.000 --> 00:50:37.000 I don't know, it's like… I think it's a hard… it's a hard niche, like, I would need to know more about the actual business, because my brain goes to things like, alright, so someone in a business 00:50:37.000 --> 00:50:55.000 We'd have to find out how they're using ChatGPT to solve a problem. So I tend to think of someone in a business being like, okay, I build out skyscrapers. So I'm working as a business that builds things for other businesses. I would have to find out how much are people using AI in their design 00:50:55.000 --> 00:51:10.000 Part of their brain, right? So, when they're designing a solution, are these people using AI? Are they not using AI? How are they using AI? What kind of prompts are they putting in? And at what point in their design journey are they doing it so that then when they go to buy product 00:51:10.000 --> 00:51:15.000 To your point about the specs and all that stuff, they know that my stuff fits. So, for instance, like 00:51:15.000 --> 00:51:19.000 I would be afraid of, like, all my specs being in a PDF 00:51:19.000 --> 00:51:34.000 I would be afraid of my specs being too table-driven, where it could get confused around cells and rows and columns. I'd be a lot more like, yo, like, just be like this, next to that, next line, this, next to that, because if I get really specific about, like, the apparatus that 00:51:34.000 --> 00:51:52.000 product needs to fit into in my business, I'm much more likely to be like, well, if you're a manufacturer, it means you make a physical thing. And does that thing need to fit inside of a certain thing that I may already have? The more clear you make it as to what sizes you have and don't have, and the more easier you make it to index 00:51:52.000 --> 00:51:58.000 I think the easier you would make it for the consumer that when they're doing research 00:51:58.000 --> 00:52:07.000 and whatnot to find your specific product. I would love to hear whoever said this, like, DM me somewhere or something, and I can try to get more, get you a more specific answer when I get more specific information. 00:52:07.000 --> 00:52:23.000 Yeah, yeah. What I would add to that too is with B2B, there is likely a buying committee, right? And especially with B2B manufacturing and somebody dropped in the chat that this question might have met back to medical devices 00:52:23.000 --> 00:52:39.000 And so you have, I would imagine, a wide variety of people who need to sign off on this thing. And so audience research, I think, comes in. I think of the brand Winter with a Y, W-Y-N-T-E-R 00:52:39.000 --> 00:52:55.000 They do a lot… first of all, they have a lot of good content on things like this and the need for buying journey prompts to be tracked. But what I'm getting at is you might, depending on what role in the buyer journey, you might be talking to different personas 00:52:55.000 --> 00:53:16.000 The person who needs to see you appear and you need to be seen for, they might be looking for specific things, and then they're like, all right, cool, this is who I want. Then a different person, maybe procurement, CFO, a different buyer is going to say, okay, can I trust Seer? Is Seer good for this? Are they, do they have this kind of, you know, certification, whatever else 00:53:16.000 --> 00:53:31.000 Totally different persona, totally different prompts. And then you may have different final criteria. So to me, I would try to map out those prompts and track in a similar way that your buyer journey is already occurring 00:53:31.000 --> 00:53:49.000 And I think the key there is doing audience research. And if you're like, no, but we already invested in that, like we did this whole thing two years ago, it's wrong. It is wrong. Tell your board, like, that is not relevant anymore, especially if you're already seeing conversions come in through AI search 00:53:49.000 --> 00:53:50.000 Go ahead, Will 00:53:50.000 --> 00:54:09.000 All right, so, med device got my brain going a little bit. I didn't see it when I was chatting. So I'll give you an example of something that's happening for SEER right now. So I just ran a bunch of prompts trying to understand something about questions people have about showing up in AI. And I found out that ChatGPT61SOL is like all searching like Anthropic and OpenAI and they're not actually searching like marketing sites 00:54:09.000 --> 00:54:25.000 Okay, so when you look at the fan out, you start to be like, where does this model have leanings to trying to answer this question? And where are those leanings in places that I'm like, I'm not getting content, I'm not getting links from OpenAI or Anthropic. I'm not getting mentioned there 00:54:25.000 --> 00:54:40.000 But then it's like, well, no, this is what it looks like to be a modern marketer, and this is why just IT and SEO working together, creating wins is donezo, because I'm going to walk you through what happened. So I went to SEO week. I'm a human. I went to a conference 00:54:40.000 --> 00:54:51.000 At that conference, it just so happens that the person that heads up Bing's Webmaster Tools was like, I saw that you wrote up something about us. I really liked it. Now I've made a relationship with one of the sites 00:54:51.000 --> 00:55:07.000 That is being mentioned in all these prompts. So now I'm thinking, oh my god, alright, I know this person, they recently pinged me on LinkedIn on something, so there is a relationship there. And he was like, I think that this idea that you had back then, we're now solving. So I had a relationship with one of the three sites 00:55:07.000 --> 00:55:22.000 It was Bing, OpenAI and Anthropic were showing up a ton for these answers. I literally am now like, wait, I could co-promote a piece with I can go and at least pitch it to this product manager at Bing Webmaster Tools to be like, yo 00:55:22.000 --> 00:55:36.000 You know how you've been saying that my stuff, social post, how do I value this? This comes all back to KPIs. So, oh, Will's over there talking to this guy at SEO week. What's the value of that? Oh, now you see the value of making those relationships in person. 00:55:36.000 --> 00:55:40.000 So now I can actually ping the person that's the product manager over Bing Webmaster Tools 00:55:40.000 --> 00:55:55.000 He has already seen that my is stuff that they ultimately put in the product, so he's like, check this out, we finally did it. So now, because of that relationship, I might be able to actually do a co-branded piece. I can at least ask, right? And now I'm showing up on the place that ChatGPT is leaning towards 00:55:55.000 --> 00:56:00.000 That typically would be off-limits to some kind of low-quality listicle spammer. 00:56:00.000 --> 00:56:03.000 I love that. That's such good stuff. 00:56:03.000 --> 00:56:07.000 All right 00:56:07.000 --> 00:56:08.000 Haha! 00:56:08.000 --> 00:56:15.000 That's it. We did it. We made it through another episode. Keep an eye out for the recap. Keep an eye out for more information here. There's a lot of good stuff. Two things. 00:56:15.000 --> 00:56:29.000 Go out and vote if you're in the US, make sure you vote next week or vote early if that's an option to you. Number two, if you or someone you know has been impacted by layoffs, I've seen so much in the news, our heart goes out to everybody affected 00:56:29.000 --> 00:56:45.000 We have some open roles at Seer so Hannah's going to drop a link in the chat. We are, I'm personally looking for a data engineering lead. So if you're fit, if you know somebody, please send them our way. And be good to each other, and we can't wait to see you again next month. 00:56:45.000 --> 00:56:48.000 Thank you all for joining us. 00:56:48.000 --> 00:56:50.000 Have a great day, guys.