100+ Cannabis Marketing Conversations

The lessons, mistakes, and growth strategies behind the industry’s most recognizable brands.

Browse Episodes

You Can't SEO Your Way Out of a Bad Reputation: The New Rules of Cannabis Search

Cannabis operators have spent years building SEO strategies to win visibility on Google, but the rise of large language models like ChatGPT and Claude has changed what "visibility" actually means. In this session, Cana Content co-founder Stella Morrison joins MediaJel CEO Jake Litke to unpack why brand reputation, not keyword tactics, is now the strongest signal shaping how cannabis dispensaries and brands show up in both traditional search and generative AI responses. Morrison explains how Google's long-standing emphasis on quality and E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) has carried directly into how large language models pull, summarize, and recommend cannabis businesses.

The conversation covers what percentage of traffic is realistically coming from AI search today, why Google Business Profile and website content should be an operator's top priorities, and how customer sentiment across reviews, Reddit, and social platforms feeds directly into AI-generated answers. Morrison and Litke also break down content chunking, FAQ strategy, and why a strong, specific brand identity outperforms trying to appeal to everyone. This session is built for dispensary owners, cannabis brand marketers, and MSO marketing teams who want a clear, practical framework for managing reputation across every place their business appears online.

Full documentation in Finsweet's Attributes docs.
Watch Now

Key Insights

  • Search still drives the overwhelming majority of cannabis business traffic. Google holds roughly 88 to 90 percent of traffic even as AI search grows, so reputation work on Google Business Profile and your website should remain the top priority.
  • AI search visitors convert differently than traditional search visitors. One brand saw 90 percent of AI-referred traffic come from new customers, arriving with a specific question already in mind rather than searching for a business they already know.
  • Large language models summarize sentiment across multiple sources at once. A recurring complaint across reviews, Reddit, or social platforms is more likely to surface in an AI-generated answer than it would be to appear on page one of Google.
  • Responding to reviews is both a reputation signal and a ranking signal. Consistent, timely responses on Google Business Profile can influence how a business ranks against competitors who are less active.
  • Brands that clearly define their audience and lean into a specific identity perform better in both search engines and AI tools than brands trying to appeal to everyone.

Expert Answers

[{Why can't good SEO fix a bad brand reputation?}

SEO tactics can influence rankings, but they cannot change what customers are actually saying about a business. If there is a recurring issue, such as inconsistent product quality or a negative service experience, that sentiment will keep surfacing across reviews, social platforms, and community forums no matter how strong the technical SEO is. Search engines and AI tools are specifically built to prioritize authentic, consistent information, which means the fix has to start with the actual customer experience, not with content tactics layered on top of it.

{How much traffic are cannabis brands really getting from AI search tools like ChatGPT?}

Right now, AI search tools typically drive less than 1 percent of total traffic for most cannabis businesses, with the majority of that coming through ChatGPT specifically. Google still accounts for roughly 88 to 90 percent of traffic. That said, this share is growing, and brands that already have strong, quality-driven content are positioned to capture more of that traffic as it increases, without needing to treat it as an entirely separate strategy.

{Where should a cannabis brand focus first when managing online reputation?}

Google Business Profile should be the top priority, since the majority of traffic and branded searches still route through Google, including simple searches for dispensary names or the word "dispensary" itself. A brand's website should be treated with equal priority, since it is the one place a business fully controls and where large language models look for authoritative, first-party information. From there, monitoring other platforms like Reddit, Instagram, and review sites on an ongoing basis helps catch issues before they compound.

{Why do large language models sometimes give incorrect information about a cannabis business, like wrong store hours?}

Large language models generate answers based on patterns and predictions from the information available to them, and they are designed to produce a confident, direct answer even when the underlying data is imperfect. This can result in outdated or simply incorrect details being presented as fact, even when accurate information exists everywhere else online. This kind of inaccuracy is largely outside a brand's direct control, which is why the more productive focus is ensuring accurate, consistent information exists everywhere it can, rather than trying to control every AI-generated response.

{Should cannabis dispensaries invest heavily in FAQ content for their website?}

Yes, FAQ content remains a strong investment because it is inexpensive to produce, aligns with how users have been conditioned to look for information, and directly supports how large language models break down longer, more specific questions into smaller pieces to build an answer. FAQ content that addresses real customer questions, such as product use cases or store policies, gives both search engines and AI tools clear, extractable information to pull from when responding to a related query.

{How should a cannabis brand think about targeting a specific audience versus appealing to everyone?}

Trying to be relevant to every possible customer tends to confuse search engines and AI tools about who a business actually serves, which weakens visibility overall. Brands that clearly define their ideal customer and build content and messaging specifically for that audience tend to perform better in both traditional search and AI-generated answers, because the signals sent to these systems are more consistent. This same clarity also makes paid advertising more effective, since organic content and paid media should reinforce the same brand identity.]

Put these Insights into Action

Whether you're optimizing product mix, improving customer retention, or measuring market performance. Mediajel helps cannabis operators turn data into profitable growth.

Marketing Attribution

See exactly which campaigns generate dispensary revenue - not just clicks.

Track Campaign Revenue

Programmatic Advertising

Reach the right customers with data-driven targeting that maximizes ROI.

Launch Programmatic Ads

Cannabis SEO

Capture more high-intent shoppers and increase organic visibility.

Boost Cannabis SEO

Webinar Highlights

00:00 – Introduction and Background

Stella Morrison shares how she and her Cana Content co-founders moved from journalism into cannabis marketing nearly a decade ago, building an agency focused on the East Coast ahead of New York and New Jersey's adult-use markets.

09:37 – Quality and the E-E-A-T Algorithm

Morrison breaks down how Google's E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) has consistently rewarded quality content over keyword-stuffing tactics, and why that same principle now applies to how large language models evaluate brands.

18:06 – How Much Traffic Actually Comes From AI Search

Morrison shares real numbers on AI search traffic share, explaining why the impact is currently small but growing, and why abandoning proven Google strategy in favor of AI-only tactics would be premature.

24:11 – Why AI Search Visitors Convert Differently

Litke shares an internal analysis showing that AI-referred visitors were far more likely to be new customers, arriving with a specific question already in mind rather than searching for a business they already know.

34:33 – Ranking Reputation Signals: Google Business Profile, Website, and Reviews

Morrison lays out a clear priority order for cannabis brands managing their online reputation, starting with Google Business Profile, followed by website content, and ongoing monitoring of reviews and community platforms.

52:26 – The Cloudflare Crawler Setting Cannabis Brands Should Check

Litke flags a commonly overlooked DNS setting that can block AI crawlers from reading a website entirely, and Morrison explains the copyright concerns driving some businesses to use it.

Frequently Asked Questions

[{What is E-E-A-T and why does it matter for cannabis marketing?}

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness, and it is the framework Google uses to evaluate the quality and credibility of content and websites. For cannabis brands, this means demonstrating real authority in the space, such as being recognized as a top dispensary in a given market, matters more for visibility than keyword volume alone.

{Does Google Maps traffic matter for cannabis dispensaries?}

Yes, Google Maps and Google Business Profile traffic has grown significantly for cannabis retailers, with simple searches like "dispensary" increasingly surfacing local map results directly. This traffic is also increasingly pulled into Google's AI Overviews, making Google Business Profile accuracy and review responsiveness even more important.

{Is Yelp important for cannabis brand reputation?}

Yelp gets mentions and should not be ignored, but it should not be prioritized the same way as Google Business Profile, since Yelp makes it more difficult to manage or contest negative reviews. Businesses should still monitor and respond to Yelp reviews rather than leaving them unaddressed.

{What is content chunking and why does it matter for AI search?}

Content chunking refers to how large language models break a longer, more specific question into smaller sub-questions in order to build a complete answer. This is one reason FAQ content performs well, since it naturally answers these smaller, specific questions in a format AI tools can easily extract.

{Should cannabis brands try to rank in every AI search tool equally?}

Not necessarily. Different AI tools such as ChatGPT, Claude, Perplexity, and Grok can return different answers for the same query, and it is not realistic or necessary to optimize equally for all of them. A stronger approach is prioritizing the platforms and search engines that drive the majority of traffic while maintaining accurate, consistent brand information everywhere.

{Can competitors manipulate reviews to hurt a cannabis brand's reputation?}

Attempts to manipulate ratings, such as coordinated negative review campaigns, are generally flagged as spam and addressed by review platforms fairly quickly. The more common and controllable reputation risk comes from genuine, recurring customer complaints, which require addressing the underlying issue rather than the review itself.

{How does personalization affect the answers AI tools give about cannabis brands?}

Large language models build a profile based on a user's chat history and preferences, which means two different people can receive different brand recommendations for a similar question. This makes it less useful to fixate on appearing in every possible AI response, and more useful to focus on being clearly and accurately represented for the audience a brand is actually trying to reach.

{What should a cannabis brand do first if it takes over a dispensary location with a bad reputation?}

Start by claiming and cleaning up the Google Business Profile, since it is the most visible reputation signal and carries ranking weight. From there, ensure the website reflects accurate, current brand information, and set up ongoing monitoring, such as Google Alerts, to track mentions on Reddit, review sites, and social platforms.]

Cannabis Podcast Full Transcript

{}Chapter: Introduction

Jake: Okay, stand by for one second, let's connect everything. We have to do the LinkedIn connection. Takes a couple seconds.

Stella Morrison: Go for it.

Jake: Redirect… Okay, I think we are good now. All right. Welcome, everyone. This is another episode of Cannabis Marketing Live. Today we have Stella Morrison. We are going to be talking about cannabis content, AIO, GEO, whatever you want to call it. I don't know if the jury is fully out on what we're calling generative AI and search and all of that.

Stella Morrison: opinions.

Jake: Okay, well, we'll get into your opinions then. That'll be perfect. Well, Stella, starting off in the beginning, I'd like to, get into how you found yourself in the cannabis industry. We talked to a lot of people around a lot of categories, and, you know, no one started here, generally. Well, a few people did, but mostly not. What was your background, and what was the transition? How did you find yourself here?

Chapter: How Stella Got Into Cannabis Marketing

Stella Morrison: Yeah, so, I started out in the newsroom. I'm a former recovering journalist, depending on who you ask, and I met my business partner, Adam Uzhalco, in the newsroom. We were switching… I was transitioning out of the industry, he was taking on my beat, and we were… we collaborated over the years, just, you know, we kept in touch, and at one point, he had, also transitioned out of journalism and into. a different publication, where he was covering the cannabis business beat, and this was back in… this was almost a decade ago, so very… definitely very new, especially on the East Coast, where we're based, where it was… a surprise that this was an industry in the first place. And he, just by virtue of his byline being on articles about the cannabis industry, he was getting hit up by people who needed marketing help. So he called me. At that point, I had already started my first agency, and he called me, and he's like, just chatting about what it would look like to, you know, work as a freelancer, like, just get some advice. By the end of the call, it was clear that this was a business. So he brought on our third business partner, Bree Brewer, our chief creative officer, to come in, and the three of us really put everything together that summer. This is July of 2017. So, just 9 years ago, I flew out to the Pacific Northwest, and I won the flight, whole story. And I took the flight, Southwest over to, that whole area, Seattle, Portland, just so I can kind of see how adult-use cannabis worked, and I remember distinctly going to Seattle Hemp Fest, and that… it was really cool to me to see, like, the culture and business, sort of, like, the same tents in the same, like, area. Had that dawning moment of, oh wow, this is a whole industry, and… it made a bunch… it just made perfect sense for us to continue on the path that we were on, and by September, we had our first client. So, we moved pretty fast with the intention of growing with the East Coast. So we opened four years before New Jersey and New York came online, and that was intentional, because those are our communities from New Jersey, our founder… co-founders are in New York, And we wanted to make sure that we were supporting the people who were there, so it was a very strategic decision just to grow with them as well. So, the journalism to cannabis pipeline is definitely, definitely not the most uncommon one, but, that's how we ended up, where we are.

Jake: Well, it makes sense. I mean, if you're following the news and the zeitgeist as a journalist, you would see, kind of, what's going on in these emergent markets happening. You know, we started in 2018, you were in 2017, that's pretty early on. It's definitely not the beginning, but it is… You know, early in the stages where, you know, we've seen a lot of maturity happen over the years in terms of how cannabis operators are running their businesses, what they're doing. And you started Can of Content. That's what it was called, right? The… that's where you're at now, kind of con… yeah. Okay. So, that's pretty ahead of the game. I know when we started trying to roll out digital advertising in 2018, there was a lot of confusion about what we were even talking about, right? So, you know, we're talking about, you know, digital ads, and most people were… The cannabis companies had been stuck in the penalty box of, like, billboards and magazines, and that's really all they had access to, so there was a big educational cycle in terms of talking to cannabis operators about what options were actually available to them. And from the content side. How did you approach that, in terms of, here's what I can do for you, here's how content can help drive your business?

Chapter: Why Search Is the Safest Place for Cannabis Brands to Grow

Stella Morrison: Yeah, so… I've always held the position, and still hold the position, that search and content… well, content to a degree, but search, is really the safest place for a cannabis business to grow their business. Instagram isn't taking away their account, right? Like, there's nothing stopping anybody from Googling you, so it became a very, obvious first choice for us, when we were trying to position these businesses. And to your point about being early, on the East Coast, you know, New Jersey and New York had severely limited medical markets. There wasn't really exactly marketing opportunity. There was no need when there were 6 dispensaries in the state. Nobody needed it, right? So we actually… we were working a lot with accessories, a lot with, a lot with ancillary business, really in that camp, and then once the licenses started rolling out, we've really, taken off on the retail side and the brand side, and I think that's just virtue of the market, the market that we really have a lot of ownership in. So, for us, for the content role is… the… I've always been a stippler for writing quality. I have been in editorial in some capacity for 20 years, and bad writing, even if you don't know it, turns people off. So… To use content as a vehicle to still balance the regulations while hitting the concerns around stigma that people still have and definitely had 10 years ago, how do I use this vape thing? Like, what do I do with it, right? A lot of… I mean, we still have to do a lot of educational basics, but content gives you the place to explore and to connect with people who learn in different ways, right? Whether that's… blogs driving SEO, whether that's people who like to read, people who like to swipe a graphic, people who like to look at a video, content really becomes a driver, you know, almost to the point where you don't really realize you're kind of in cannabis school, right?

Jake: Yeah. Yeah, absolutely. And, you know, the… the rules of SEO have… they're always changing, right? And even faster now, it seems like algorithm changes are happening. you know, there's sort of this ongoing joke about SEO content, where it's… people say. like, the average consumer that's just looking for something doesn't understand all of the work you need to do to play ball in Google's world. We'll just talk about Google, because that's still the majority of search content. Agreed. And, you know, people would say, like, well, why do I need to read an article that's 2,000 words long about someone's grandmother's apple pie in, like, the baseball game they used to go to in order to get my apple pie recipe, right? I understand why that was happening. I feel like that's calmed down a little bit, I do think that… you know, Google's effort to try to deliver results to people, at least on the organic side, is, you know, directionally correct. That's what they're trying to do. They're trying to find content. But it's an arms race, right? You've got an army of people out there trying to… Optimize or manipulate, depending on which adjective you want to use the algorithm to get their content at the top, because being on page 1, or in the top three, makes all the difference in the world for a brand. And I do agree with your statement about search being the safest place, and if you look at, sort of, the… old-school, traditional marketing categories, you kind of three circles, right? You have paid media, you have owned and operated, and you have earned media. And they kind of can overlap to a certain extent. But your owned media is the place where you have the most control, and it's the place where you build an annuity from a marketing perspective, no matter how efficient… and we work in paid media, so… but no matter how efficient your paid media is, you're always spending some amount of dollars or pennies. to get the traffic. Whereas, if you put the foundational work in to get your content in place well, in ranking, that pays in the long term, until they change the algorithm.

Stella Morrison: You touched on… sorry, finish your thought.

Jake: I was gonna say, it's a very long preamble to… Over the last 8 or 9 years, how have you sort of navigated the algorithm changes and helped your clients adapt their content to stay relevant, and sometimes you have to do things that are a little counterintuitive, right? And that can be difficult to explain to your client about why we need to change your content in a specific way to… To… to deal with the world that we live in?

Chapter: Google's Quality Focus and E-E-A-T

Stella Morrison: Yeah. So, you touched on a couple things in your introduction that are actually played right into this, right? So, one constant that Google, and I agree with you to stick with Google. One place that Google has consistently and constantly prioritized is quality. That's where the journalism background came in for us, that for… this is why we are human-driven. AI… Gen AI has its time and place, not with, generating creative, in my opinion. And… one thing that we are consistently and constantly seeing, as I said, is… is quality always being preferred, right? So, this is because whenever Google rolls out an update, they roll out multiple types of updates a couple of times a year. Every so often, these updates are gonna be pretty drastic. Some of them are way more drastic than others. Either way, a lot of the times, you do see, like, demoting spam as part of that strategy, right? Because Google's goal is to make sure that you're getting the information that you want. And if you have to read that 500-word preamble about the baseball game with Grandma to get your recipe, that's actually in one of their, updates, I think, I don't precisely remember when, but one of their updates basically said, hey, recipe people, cut that out and get straight to the recipe. So they… they responded, right? So, and that's similar to any kind of content you're putting out there. When, you know, 10, 11 years ago. The strategy in a lot of places was flood things with content, right? Just keep writing about this thing, mention this thing. There's this precise science where you should mention this thing 12 precise times, and you're gonna somehow be… number one with that. That is just a very… long gone are those days, right? Like, when I first started out. you would see, you know, for example, like, you know, I worked at the lender who had a separate blog just about lending topics, hoping to kind of flood information about them on the internet. That doesn't work anymore. That doesn't work anymore, and there's a reason for that, because Google tries to keep up with what's useful. An important algorithm for people to know, it's called EEAT. This stands for Expertise, Experience. authoritativeness, authority, and trustworthiness, right? So, when Google's trying to evaluate what is… A worthy piece of content or a worthy website to rank higher on search results. Is this a place of authority? Does this entity have a place in this world, right? So, for dispensaries, for example, that becomes an education tool that the internet needs to understand that you are a top dispensary in New York City, or, you know, the best dispensary in California, or the most respected dispensary in Nevada, whatever that is. like, search engines need to understand that and help build those bridges that help elevate that authority. So, it has severely limited, the possibilities that someone can basically you know, frankly, talk at their ass and hope to rank. That doesn't… that's not something that can, happen anymore. People are trying it with LLMs, they're getting slapped down pretty quickly, but this is, definitely something… so we're… there's a little bit of element of everything old is new again on the large language models, like ChatGPT side, but, overall, the biggest con… constant with all the changes I've seen with Google over the last 8 or 9 years, quality has always been at the center of it. We can argue all day about what quality is and what that looks like, but quality is their intent at the end of the day.

Jake: Sure. Well, quality is qualitative by nature, right? So, it is subjective. Okay, so… now let's transition a little bit over… just, I think that was a helpful conversation about SEO, which is still… where the majority of organic traffic is coming today, right? We, we have an SEO practice here at MediaJel. We help some of our clients with their SEO, generating quality content, using, you know. good editor. We try to, again, focus on quality, and we have for quite a while. And that has turned out to make the transition into Well, maybe we can have this debate now, G, GEO, AIO, whatever, you know, generative AI chat responses where people are asking questions and getting directed in a way that's different before, which is. And Google does have their own AI responses now, but let's just… put that in the generative AI category and talk about the fact that what people used to do is type in some keywords, maybe a question, get an array of responses, and then decide from that array of responses with a headline and a couple of pieces of text, maybe an image, where they should go next. Whereas, what people are doing now, more and more, is just asking a question. They may be using voice. A lot of people have transitioned to voice, because it's easier, and people have learned you can… you can input a lot more content into ChatGPT or Cloud. You can ramble for, like, 2 minutes about your whole, whatever it is. And that actually can be helpful. And then instead of getting a list of things to look at, you get some actual instructions from the agent. With some analysis, and then some citations of places you can go, and that's usually gonna be two. maybe 3 places, right? So you don't have 10 selections, you generally don't have ads in there. I know ChatGP does have an ad-supported version as well.

Chapter: How Much Traffic Is Actually Coming From AI Search

Jake: And for us, what we've seen with companies we work with over the last year, the percentage of traffic coming from generative AI is approaching double digits, but still kind of single digits. Now, those numbers are being reported by the analytics on their website, like Google Analytics or whatever they're using. I think a lot of people, are organically sidestepping the whole thing in the first place, which is they're getting the response and then, like, typing something in. However. I think that people that, like yourself, that were focused on quality, when that transition started to happen, I think you had a leg up. Because if you were just using mechanical keyword stuffing before, LLMs don't fall for that at all, right? Because they can take 200,000 words and look at it all at once, and I do think that They are inherently somewhat better. At understanding Quality, what are your thoughts on that?

Stella Morrison: Okay, yeah, so you said a couple of important things. The first thing I want to contextualize is I want to explain to the audience what… how much traffic is actually coming from large language models like ChatGPT, because this is something where… There are folks who like to position this as end of the world, and if you don't show up today, your business will go out of business tomorrow, which is not the case. The average, it is extremely hard to estimate across industries, across, you know, across different sectors, but the average that we see, cannabis, not in cannabis, less than 1% of your traffic is coming from large language models. The vast majority of that is ChatGPT. Because it is… it's kind of winning the brand more in terms of visibility, whether or not the public eye is a different story. But, so you're seeing less than 1% of traffic coming from search, you know, half a percent, 0.25%. We… anecdotally, I am seeing that, there seems to be higher click-through. I think people are just highly suggestive. I think that's a definitely different conversation. But Google still holds 90% percent of your traffic, 9-0. So, that has dropped over the years. It used to be closer to 95, 96, now it's more like 88-90% as search diversifies. That is enough of a dip to definitely make Google panic, but it also means that we don't ignore Google in exchange for a fad that's going to work on necessarily 0.25% of all your traffic, right? And… Keeping that in mind, I hope, calms down some folks who might be nervous about, I'm not in ChatGPT yet. Yes, are you missing opportunity? I believe you are. Are you… are you gonna miss 100% of all your revenue? You're not going to yet. It's going to grow. I firmly believe the .0… that 0.05%, that 0.25% will grow. I don't think it'll become 50%, 75%, but I think it'll grow. So we want to encourage people to think about that now. In terms of the terminology itself from, like, SEO and GEO and AEO, we get asked, I want to say twice a week, like, do you do GEO and AEO2? Yes, we do. And the reason we do that is because, GEO stands for Generative Engine Optimization, and it refers to, like, tactics that you would use specifically to try to appear in LLMs. Answer engine optimization is similar, some people apply it just to AI overviews, but this is evolving terminology, and I don't want people to use this terminology and to think that they need to do a completely separate service specially for this, because the VIN diagram between search engine optimization and GEO or AEO, it's almost a circle. It's almost a complete circle. There's a lot of fundamentals that overlap, and I'm sure we can get into that later. So I think some just basic education on these rapidly evolving pathways is incredibly important for people to know before you start getting into, like, the nuts and bolts of the… Getting into the nuts and bolts of the discussion.

Jake: Yeah. Yeah, that makes sense. I think we've seen, and it varies wildly, I know for us at MediaJel, we… the transition was relatively smooth for us because our whole strategy as a company is around content, right? So we don't have, like, a sales team. All of our business comes from really just content marketing. And so, we have, like, 500 pieces of content that we've built over the last 8 years. A lot of it's in the format of an academy, and we've spent a lot of time building a good amount of quality content. And so, when traffic started, when people started moving over to ChatGPT and started typing in cannabis marketing, we kind of just, like, out of the gate, were right there, because we had so much quality content. Now, I will say, for some of the companies that we work with, we've seen… the last… statistic that I remember, which is a little bit dated, was seen as high as 7%. Which is still single-digit, and I still think Google's gonna continue to be 90% plus, for quite a while, but we did see, we have one of our dispensary chains that we work with where they were actually getting 7% of their inbound traffic with ChatGPT as the UTM code. It is mostly ChatGPT. I use Cloud a lot myself. But I'm doing development work and other type of things that are not, like, necessarily consumer activities. And I'll, you know, we can follow up, and I can see if those numbers have moved at all. For myself, I have found that I generally now, when I want answers to things, I'll go to either Claude or ChatGPT or something like that. I still heavily use, if I want to find… like a business or a restaurant or something like that, like a location-type thing, Google Maps is probably still the… definitely the winner on that.

Chapter: Google Maps and AI Overviews

Jake: And they're new, I don't know if you've tried it. I just noticed it, maybe a month ago, is the… in Google Maps, they have the, like, ask… ask Google Maps feature now? It's on the mobile app. That, actually, that thing works great. And it works better than… I was traveling somewhere, and I saw that, and I was just looking for what, you know, things to do this weekend in that city, and the response I got back was better than any other… search I'd tried in the past, right? Yep. Alright, now I've got myself off on a tangent, I don't remember what we were talking about.

Stella Morrison: Just maps, and I mean, listen, Google Maps is a whole other discipline, and one that actually, it's separate from LLM, so I'm not going to get too far into it, but we have seen a drastic swing over the past 3 or 4 months, in terms of the uptick of traffic across the board, across all retailers that we work with, from, from Google Maps. There's definitely an algorithmic preference, because it is a Google product. definitely enforces the importance of, Google Business Profile for these dispensaries, for retailers, anybody who has that kind of, like, client-facing, that kind of customer-facing work. We're also seeing, I mean, the traffic is just exponentially spiking. We're also seeing Google Maps, results getting pulled into AI overviews, which is Google's summary at the top now, so we're starting to see that pulled in, which is… and that's the… that's a… excuse me, that's a pretty recent change as well, so… As you kind of alluded to earlier, this is a discipline that's changing At the blink of an eye, and is changing faster at a faster pace than it ever has. The terms of traffic, you know, we are seeing, you know, you mentioned the 7% metric. I have seen that before. I have seen that before. I think it's a bit of an anomaly, though, just from what I'm seeing, what I'm experiencing, but there seems to be no doubt across clients, that the click-through rate is extremely high. So, like, for example, there's a, like, a law firm that we work with, and This is all anecdotal, by the way, because there's really no good way to measure precisely how many times you're appearing or being mentioned. I'm kind of a skeptic when it comes to the tools that propped it up, but I use it as, like, I use them as guideposts, right? So anecdotally, the client would tell us, I think it was 10 calls, they said, yeah, we got 10 calls this month that said that they saw us on ChatGPT, and then we looked at their metrics, and they had 7 clicks that month. That's 70% click-through rate, which is, to me, speaks to the power of suggestion of these tools, and how powerful these tools are going to be in order to get people to actually, like. take the action you want them to, to show up, to… to shop, to sign up for an event, whatever that is. And I think that's only… that power of suggestion might even get stronger. So, there's something… there's something there to explore. I think we're a little bit too immature yet, but this is where, you know, this is anecdotally what I'm seeing, and I think this is… I think there's something really valuable to pursuing that.

Chapter: Why AI Search Visitors Convert Differently

Jake: Yeah, we ran an analysis recently, specifically trying to dig into what the cohort differences were from traffic from an LLM. Versus search, and this is just one particular brand, so it's not a longitudinal study by any means. But 90% of the people that came from the Agentic search were new customers. Right? And… I believe it.

Stella Morrison: 100%, yep.

Jake: And I think that's because, many, many people use search as a shortcut to just find someone they already know about, right? Whether it's brand or non-brand. search, they're looking, like, no one's saving their dispensary's phone number in their phone. Generally, they're just gonna go to Google and type in, you know, dispensary near me, or whatever the name of their dispensary is, and they're going there. So there's a huge amount of volume of people that use Google as their Rolodex. And people that are coming from a AI search context are people who actually had a question. about what they should be doing and who they should be talking to. And that is a… that's a pretty powerful number, 90% new customer take rate and high click-through rate, and I think that the… There, whether it's warranted or not. There's a natural inclination for people to have a much higher degree of trust. when they're coming from an AI agent, because they got a bunch of explanation about what it was. They didn't get 3 listings surrounded by a bunch of ads, and whatever else is going on, and they've had… they've already… they're already kind of mid-funnel at that point. Usually you talk about, you know, top, mid, bottom funnel. Display being the very top of funnel, search being sort of mid-funnel, but you don't have enough space inside of a SERP result to do any kind of real brand selling. And when someone has already spent a few minutes researching the topic, gotten a bunch of information, they're coming in as a much warmer lead, to your business from a AI conversation than they would off of a search result.

Stella Morrison: Folks who are using large language models are very much in a discovery phase. They are… they have a very specific problem that they're looking to solve. And this is because AI mode, ChatGPT Claw, it gives you the space to write out your thought, as opposed to trying to distill it into a keyword, no matter how short or long word that keyword is. So it's a different kind of conditioning when you're working with this, right? So, yes, you're looking for… you might be looking for dispensary open late on Google, and then the maps will come up if you're the one that's open at 1 in the morning, and that's what you're looking for. Or dispensary open now, right? But, you know, you could put more context into that, into an LLM, right? So you can look for, you know, I need a dispensary that closes… that doesn't close until after midnight, I'm going to be in this place, what's the closest? Because, you know, they could kind of use it to spit out that itinerary for them, right? So, you're dealing with things that… you're dealing with the Really deep, granular, specificity that… is not easily traceable in something like Google. We have an absolute plethora of keywords, of the way people are searching. what they're searching for. We have all that information, but each person is going to communicate differently. They're going to type things into a large language model differently, right? So it's about, you know, I call it the oversharing era, is what I've been calling it.

Jake: I feel like we've been in the oversharing era for a while.

Chapter: The Oversharing Era and Feeding Detail to LLMs

Stella Morrison: I absolutely have. Absolutely. So, when I describe the oversharing error, what I mean by that is Any detail that you think might be inconsequential to your business is consequential to somebody else, right? For example, if you're a dispensary, you don't think it's important to include, you know, if you're woman-owned or Black-owned or AIPI-owned, but there are people who are using the LLMs to give me a list of all the Black-owned brands in New York that I can shop with, right? Make sure you include those details wherever you appear, so it helps increase your chances of showing up. So, think of… I want to think about ways to… I want to encourage business owners to think about ways that they could be sharing more detail about the kind of customer who they're trying to reach, what they offer, and finding those connections, because that'll help feed the likelihood of starting to come up in LLMs, but also helps for Google as well. detail, as long as that detail is actually important to your searcher and answers a question for your searcher, it's good for search as well, and that's one of the many places that, like, I'm going to call more traditional search engine optimization and working within these large language models really tend to overlap.

Jake: Yeah. Alright, let's, it's such a big topic, I think this… this long, extended, half-hour time preamble before we even get to the topic we're supposed to be talking about, but I think we're ready to do that now. We have a headline for this, or a byline, or whatever. I'm not a journalist, so I don't always use the right words, but, we've got a phrase here from you, which is, you can't SEO your way out of a bad reputation. Alright, so we've talked, I think, a fair amount about some of the fundamentals of what's happening with large language models, sort of the history of SEO, where it's come from, where we're at today. what does that statement mean? I mean, I could… like, how… how do you explain the underlying premise?

Chapter: You Can't SEO Your Way Out of a Bad Reputation

Stella Morrison: There was, in the earlier days, definitely an element of fake it till you make it. I alluded to it before when I was talking about, you know, people who would just flood the internet with content, or fake material, fake sites, or just, excessive… let's just try to drown out the noise by putting up our own noise. Those days are absolutely long since gone and have been for a long time. So, if you are dealing with a brand reputation management problem, you know, there are… I mean, there's definitely, like, more, like, more underground… what they call black hat tactics that you… people use for this, but… It is… if… it's gonna be really hard… it's much harder now to keep something that you don't want people to know about your business under wraps, right? So, over time, Google, as I mentioned, has caught up and really doubled down on, you know, what's authentic, what's real, what's authoritative. Where can I get the best source of truth of information? And, you know, part of that is from the brand itself, that's reputable reporting. This is all going to play into, you know, how LLMs pull their information. it's about, like, where does that source of truth come from? So… and now that we are… A, people are looking in more than one place, they're not just Googling, they're going to Reddit, they're gonna go look on Instagram for your profile, they're gonna… maybe find you on LinkedIn, they're maybe gonna read some reviews, they are gonna read some reviews, nearly everybody literally reads a review. And it's gonna… it gets a little harder to fake that, right? Anytime someone tries to, like, brigade a rating, right? They try to tank a 5-star rating out of Google profile down to, like, a 3 or a 2, that gets flagged as spam and taken care of pretty quickly, right? So these tools have really caught up with any ways to manipulate that reputation. So if somebody is, you know, if there is, like, a… a consistent bad customer experience that, that comes up. Maybe there's, like, a bud tender issue, maybe there's, you know, some funky flavor in your product, and, you know, people are just talking about it in different places online. It's not gonna be… it's… having, like, great SEO is not going to solve that problem. What's going to solve that problem is actually correcting the issue. So, it does make it a little bit harder for, you know, anything that might be negative to get, you know, sort of buried. It does make it harder. But I've argued this for a long time, where, this is… your brand and what people think about your brand and your reputation, that's so intricately tied to search, and I think it always has been. It's just gotten renewed life in large language models, because, you know, if you're going to Google, looking up the brands. Maybe you got lucky and that, you know, that bad thing you don't want coming up is on page 2, and it's very unlikely to be seen. Even at the bottom of page 1, it might be unlikely to be seen. But if you are using something like a chat GPT or a Claude, and part of that job's tool is to summarize what it's seeing, and 4 or 5 different places, a theme comes up of that bad customer service or that weird taste, aftertaste comes up. that's not… that's more likely to be shown. So I think this… this concept has really just gotten renewed life, but I… I will always argue that your brand reputation always has been important, first and foremost.

Jake: Yeah, let's talk about brand reputation, especially… With large language models, which have… just a huge breadth of information that they can draw from, right? And that's… that's a changing landscape, because there's licensing issues of whether or not your… the LLM is allowed to read Reddit, or not allowed to read Reddit, or allowed to read Google, and how much of it. However. there are… I would… I would assume there's some signal sources that are more important than others. And let's… let's look at this in the context of, a cannabis brand, or a cannabis dispensary. that they all have limited resources. They're just very difficult businesses to run between excise taxes and 280E, which maybe is going away. And just everything else. It's very difficult. Everyone is operating with constrained resources. most… Most of these companies can't afford to have a huge agency budget, or a huge marketing department, and so they need to kind of pick and choose their battles. And if you were to give advice to a dispensary owner in terms of. There's… there's Yelp, there's Google reviews, there's Reddit, there's your own organic content, companies go through life cycles. There's a lot of dispensaries that have changed hands. Someone… and you can't just sell a business in cannabis, because the license is attached to the address, so there's all these things that make it difficult, so… there's companies that have come in, they took over a new location, maybe that location has a bad reputation. Those signal sources are kind of scattered around. How would you rank those signal sources for someone to go and tackle, right? Like, let's say, okay, you should go as the brand owner and respond to some of these negative comments on Yelp or Reddit. Like, where do you see the more powerful signals being today?

Chapter: Ranking Reputation Signals: Google Business Profile, Website, and Reviews

Stella Morrison: This is a separate answer for Google and for large language models. Okay. But I would say Google Business Profile should be the first and foremost priority, because, as we mentioned before, 90% of your traffic is still coming through Google, and we have seen the pendulum swing in the last couple of months of way more traffic coming through these local searches. Even just the word dispensary is bringing up a dispensary now through Google Business Profile. So, negative sentiment that goes on… addressed there becomes a brand reputation management issue. It is also a signal to Google that you're an active business. Not responding to reviews can have a negative effect on your ranking. If your competitor is responding to reviews. they are more likely to show up ahead of you. So, keeping on top of that and responding, you know, not 100 at a time, like, don't go and catch up on 100 tomorrow, that could flag some stuff, but tackle a few at a time. As they're coming in, make sure you're responding as they're coming in, you know, don't give it, you know, it doesn't have to be 5 minutes, but don't wait 3 weeks. So I would say Google Business Profile is definitely your most important of all of these. The second would be making sure the information on your website is what you need it to be, right? So this is not just your technical information down to your hours and how people pay. It also involves, you know, what are the brand values that you want people to know about you? Making sure that that is reflected on the site, because A, it's… your website is gonna be… One of the first things that comes up, assuming things are done right, for when people look up your brand, and I do agree with you that I think more branded search is done on Google, because people just look up dispensary X, and they… that's how they find the phone number, instead of typing in dispensaryx.biz, right? So, I agree with you on that. That's actually… what some of the evidence says, you know, for our clients as well. But I also think that, what we are seeing is that for a time, and, you know, it's fading in and out. What I hear versus evidence is always a little bit of a conflict in this space, but… Large language models, in theory, should be looking at your website as a primary source of information. Because you are the authority on your brand. You have full control over your website. You have full control over what goes on there, what you say, how you say it, and large language models should be pulling from that. Ai overviews should be pulling from that. So, making sure your website actually aligns with the company you want to be out, out there for… out there as, is incredibly important. You mentioned Yelp specifically. Yelp is a weird place. Definitely… it definitely gets, mentions, but it's always… it's always a little bit harder for them, so I would say… I would not prioritize it the same way as GBP, but you don't want to leave, like, a few reviews, like, lingering, unresponded to, right? Or at least take care of it. Yelp does make it a little bit harder, too. kind of clean up those more aggressive reviews, they definitely make it harder. And then on top of that, it… no one's gonna like my answer, but try to keep an eye as many places as possible, right? There are tools out there, some free, some paid for pretty cheap. you could use to just kind of, you know, keep an eye on, well, even a Google Alert, which is free. Like, hey, what came up on this Reddit thread about Dispensary X, right? What was somebody talking about? Just, you know, looking for those kind of tools can really go a long way. And you can kind of create a knock-down list from there, and then sort of see what the sentiment is there, and if that involves, like, a larger… does it become… does it become larger than a marketing decision at that point, right? Is this something more institutional that you have to tackle? So, bottom line, to go quick. Google Business Profile and your website, I would actually put those in equal priority, but… We're talking about brand sentiment, things out of your control, GBP first, website, and keeping as many eyes as possible on other places where you appear.

Chapter: FAQ Content and Content Chunking

Jake: Yeah. In my suggestion. Along that vein, There are… And, you know, it changes rapidly, the LLMs change rapidly, so… you know, if you're looking on YouTube for AI marketing advice, and there are… plenty of people on YouTube trying to give you marketing advice and sell you a course, potentially. But, I've seen some, you know. And I would call most of this anecdotal, because people don't really know what's going on under the hood, but I've seen kind of a top piece of advice for people on their website, is make sure you have a really thorough, like, FAQ or multiple FAQs on your site, because those are questions, you provide answers, and LLMs prefer question and answer content because it is easier for them to digest. I'm just curious about your opinion on that particular talking point, and whether people should be focusing on that for their website.

Stella Morrison: Yeah, so FAQ is a type of content that search engines understand pretty easily, and there's… there's technical reasons for that I'm not going to necessarily get into here. Faq was definitely… is definitely important. As a standard, it has been deprioritized because, we're starting to see FAQ as, like, it's still important, but it is when you're dealing with some of the more technical requirements, it is really intended for the Cleveland Clinics of the world, right? Like, for the… for the… for medical, for things like in the YMYL, Your Money, Your Life camp, like, things like med… you know, medical, more authoritative areas. It's not meant for, like, every single business ever to have an FAQ. That being said, we still highly encourage them for two reasons. First, well, three reasons. First, I think people have been conditioned to expect them, and there's an element of user experience and, for websites that involves meeting user expectations. Second, it does no harm to have the FAQ. It doesn't harm you. It might not be the… absolute rocket to the moon that people expected a year ago, but it's definitely important. Third, does come back to this LLM situation. So, to be able to connect this, I think it's best for me to very, very high-level briefly explain how LLMs are actually pulling information. So, large language models are pulling from the body of information that it has learned, and from, like, basically, in summary, current Google… well, search results, Bing is some… Google, whatever it is, and they're… taking that information and basically summarizing an answer for you based on predictability. So, it's making predictions based on what it's learned to answer the questions that you already have. So, one way that it does this is it breaks down the longer… remember, we talked earlier about shorter queries and longer queries? It's breaking down that longer question that you have into multiple pieces, right? So instead of looking up best edible for sleep, you might look up, where's the best place to buy edibles for sleep in New Jersey? So now what this query is doing is it's looking up dispensaries in New Jersey, like, places to buy cannabis in New Jersey. It's looking up edibles, it's looking up edibles for sleep. It's breaking these down into, like, these micro areas, quote-unquote. They… it's called content chunking. I'm not a fan of the term, for a couple of reasons, but, they break it down… basically, they break it down to smaller, Questions in order to build up an answer that you're then going to read. So, an FAQ is just a natural extension of that, because you can answer things like, do you have, to use the edibles example, you know, do you have edibles for sleep? Do you carry edibles? You know, how many edibles can I buy in New Jersey? You can structure an FAQ in order to feed that. So, it's a bit of a… you're always gonna get a bit of a longer and a bit more complicated answer with me, but that's, essentially, you know, that's my take on the FAQ. I do personally recommend them. I do… I do believe that they should… they have a presence on, basically any dispensary or brand website.

Jake: And it's inexpensive content to generate. Absolutely. Like, it really is. Absolutely.

Stella Morrison: You have to be a poet to write FAQs.

Jake: Yeah, and I think something you said, earlier about people are somewhat programmed to look for those, because they've become very common. But I also think that there's a, a psychological benefit, because, especially when someone's asked… like, if they're coming into a brand or topic that they're not experts on, and, like, no one's an expert in cannabis, even experts, it's such a big topic, I think people naturally look to the questions that other people have asked, right? They may have a question like, oh, I'm interested in sleep, right? But there's a whole bunch of other questions that they would have not thought to have asked, and so they look to the FAQ as a place to be like, oh, what should I even be… asking about. Not even…

Stella Morrison: that, but it also… this creates a whole, universe of content that helps extend that, right? Are you losing sleep because you're in pain? Are you anxious? Or do you have trouble falling asleep and staying asleep?

Jake: Those are two different things, yeah.

Stella Morrison: 100%, and, like, oh, do you have trouble? Like, is it… is it waking up groggy? That's the issue? Like, whatever that is, like, there's so many, like. smaller in pieces, it's, you know, it's called intent, right? There's so many other intents within these questions that, really helps shape what predict… what answer you're going to get, right? So this is where the predict… in my opinion, this is where the predictability of… well, the prediction part of the LLM comes in, right? If you're… if it hears, you know, and this is where, especially if you have an account and you're using it all the time. if you have 20 chats that are about, you know, I have restless… you know, I have nerve pain, right? It's gonna figure out contextually that you're actually looking for products that might calm down nerve pain and, as a result, help you sleep, right? It could… it eventually could get to that point. another reason why I think tracking is just… is almost a lost cause, to some degree, in large language models, because it's just so hyper-specific, and you're not going to tell your entire website to that one query, right? But it goes to show you just, like, the sophistication of what the kind of information that people are looking for, right? They heard that this can help. How do, you know, it's our job as search engine marketers to help figure out, well, what's the intent behind that question, right? There's so many, you know, is the person looking for, things about sleep because they have pain, because they're anxious, because they can't fall asleep, whatever that is, or maybe, you know, whatever, whatever that looks like. You know, it's our job as strategic marketers to figure out… to kind of reverse engineer that pathway, and then make sure that we feed the content necessary in order to help people get to the conclusion that they're… that they're hoping for.

Jake: Yeah.

Chapter: AI Personalization and Hallucinations

Jake: On… on that thread. One of the things that happens at this point with anyone who's… even if it's a free account. is the… the AI chat services now start to build a profile on you as a person, right, based on the things that you've asked. This… this signal, I think, is much stronger than anything that people have had to deal with from an SEO perspective. In the sense that… let's use a topic of sleep versus nerve pain versus whatever it is. It's very likely that the individual who is searching or, or, chatting. with an LLM is had extended conversations over a period of time, and they… and there's now a profile that the LLM has built about that individual. And they are going to use that information when they decide what to reference and direct you to, right? And so, you mentioned… and we could go with a different category, we could go with, like, political leanings, or, or, things that people are passionate about. And so, you see these responses now when you talk to the… with LM. You could say… actually, I don't know what would even happen right now if I went to chat GPT and say which dispensary would be the best one. to go to, but I guarantee the preamble would say something along those lines of, well, based on what I know about you, that you have cats, and you have nerve pains, or whatever, it's now going to give you A completely different set of results than it's gonna give someone else. I have… yeah.

Stella Morrison: This is something that I really want business owners and marketers to understand, because, you know, it's a lot to hear the pressure of, you have to be in these systems, you have to be representing these systems, you're falling behind, you're not doing XYZ obscure thing that I don't understand when I'm telling you it, and you're not doing it, but take a breather, take a second, and think about what this is, right? So, it's… it's… it's that kind of… it's so deeply tailored to the person, and as an aside, just because I have… I understand how it works doesn't mean I don't have moral qualms about the whole thing, but, you know, this is a system that is… It is designed to tell you what you want to hear, and that is a conversation we cannot get into today, because we will be here for 4 hours, and you have things to do. But, this is… this is where hallucinations come in, right? This is where the information, like, I continually see, you know, for example, like, we started working with a dispensary recently that, you know, when part of our onboarding is, I… I go, I have, like, a blank account. I'm not signed in, chat GPT, I just want to see what it knows about the brands, right? It had its hours wrong. It had its hours wrong, even though the hours were correct everywhere else. But it wanted to know… it wanted to tell me what I wanted to hear, which was, you know, what is it… was it… I had this example on the brain, late at night, right? What's open late at night? So, the hours are just… they were… they were made up. They were not changed by the client recently, nothing like that. They're correct in every place where it appears. The hours they gave me were just never the hours, right? So, I think that is a separate conversation than the mechanics of how the system works, because Reality, in my humble opinion, there's really nothing that we can do to control that, right? There's, you know, we can put out all the information as best we can, and make sure that we are doing everything right from a technical perspective, from a content perspective. If the LLM hallucinates, that unfortunately becomes out of our control, and I think that's just sort of a weird path to go on, but going back to the customization of it all, I think there's something really important to remember here, which is People are going to get very different answers depending on What they're searching for. how they're searching for it, and the platform they're using. Every study that's been done, and every piece of… I'm not gonna call, but I do a scientific study, but every, you know, every… all the work that we do behind the scenes. You see different answers in ChatGPT than Perplexity, then Claude, then Grok, then Clarity, then all of them. You're going to get different answers on each one. Do you want to spin your wheels knowing that you're coming up only one time in one and two times in the other? is that really worth it when you have 10,000 clicks coming through a month from ChatGP… from Google, right? So, this is where it's like, it's okay to not… Controversial, sorry everyone, but it's okay not to know everything, right? And it's okay, out of 10,000 clicks if you have two data points that are interesting, but it's not going to collapse if we can't follow them the way we do, and it's because there is this level of, like. personalization of what's being experienced on the other side, all these, you know, very scary stories we hear about, you know, people using ChatGPT to do all sorts of things, right? I think just goes into, like, a more of a moral dilemma than a technical, like, consideration, right? It's a different… I think it's a different animal.

Chapter: Building Brand Identity and Audience Targeting

Jake: Yeah, I think from a marketing perspective, you know, you definitely can't please everyone all the time. But I would… I would think that a good approach would be to… try to lean into your strengths as a brand and the experience that you want to give, right? If you are a high-volume, value type of operation, you should lean into that. If you are a bespoke, high-touch, custom thing, you should lean into that. And you can build your content strategy around that, right? So, who's your ICP? or three ICPs, and say, okay, well, let's try to build content on our site that when the LLMs look at it. they're going to recommend it to the people that we want in the… anyway, right? You can't get everyone, but try to get the people that are authentic to your brand, and build your content strategy around that, so that when the LLM suggests your brand to that people, you're more likely to build a long-term relationship with that consumer.

Stella Morrison: You are 100% correct, and leads me to two, like, pieces of homework that I think people who are listening might actually benefit from to go for, which is, you know, when you are poking around in ChatGPT or any other large language model, you want to kind of evaluate, look at two things, right? Is what's coming back the consensus of what you want people to understand about your company or brand? And also, take note of who is talking about your brands, right? Are those the publications that make sense, right? If you're, like, a very fashion-forward, you know, design modern kind of dispensary. Do you want that Architectural Digest article to come up, right? Is the ownership correct? Were you correctly characterized? Or that could indicate that there's a larger issue of people not understanding who your ownership is, right? So, that kind of… it gives you… these are all breadcrumbs that leave you clues of what you need to be doing next, and what you need to be focusing on, you know, whatever that strategy looks like for you.

Jake: Yeah. One other thing that is a weird… it's such a small checkbox, but, you know, the… that's… for your website, if you're using something like Cloudflare or something else like that to manage your DNS, there's a checkbox there that people may not be aware of that just disallows LLMs from going to your website at all.

Chapter: The Cloudflare Crawler Setting and Copyright Concerns

Jake: And that's something that people should be aware of, because you could be doing all this work, but not even know that you've… you've banned. Now, whether the LLMs respect that or not is a different question. But, it will certainly get rid of some amount of signal that you could be giving to LLMs. I don't know if you've run into anyone that has had that issue yet, where it's like, oh, we're getting no traffic at all. Well, let's look at your DNS settings and make sure you're even letting them in the door.

Stella Morrison: We have not run into that yet, because all of our clients want to show up on large language models, so we have not run into the issue yet. The root of the issue comes from copyright infringement, right? Is folks who… and let's be real, the content libraries that we have built for clients get ripped off all the time already, and we didn't need large language models to do that, right? Yeah. So that's a different, that's… that, I think, is a different… sort of a different conversation of the why. I would encourage people to let the crawlers, what's called the crawlers, so basically, just for non-tech-oriented people, the algorithms… for lack of a better term, come to… they come to your website, and that's how they read it. That's how they read the content, right? We want to make sure that large language models can read your content, that Google can read your content. And I have… I've had… I have had clients before who weren't crawled, and it's because they had that checkbox, either on WordPress or Squarespace, that says, deny the crawlers, and that solved all the issues, was the one checkbox, right? I have yet to see this from the LLM and the Cloudflare, specifically from that angle, I have yet to see that, because I have yet to run into people who are concerned from, like, a copyright perspective about being put in, put into the, you know, getting, getting read that way. But 100%, in alignment with you about what you're saying about reaching your target audience, right? There's something, we cannot be the same thing to everybody. If your target audience is 21 plus with a wallet, you have already lost out the door. We are so… far away from, if you build it, they will come. We are at a point of maturity, right? I mean, I… we're dinosaurs at 8 and 9 years in this industry, and their people have been doing this way before us, right? This is a point of maturity. You don't just buy it because people want to buy weed, man. That's just not how… that's not how it works. There's sophistication. to… The retail experience, the same… the same applications and the same thought that's going into store design, web design, retail design. is going into dispensaries, and those are the people who are coming out the other side. The clients we work with, who know who they are, who have a strong online presence. Who take their online reputation seriously and know who their clientele is, are the ones who perform best in search engines and the ones who perform best in large language models. And it's because they know who they're talking to, they're not trying to appeal to everybody, because if we're putting out 12 messages for 12 different people, then we start confusing search engines about, well, where are you located? Where are you? And then we… we build on… you can build on that and add more audiences over time, but avalanching this has always been what's worked for us. It has always worked for me, you know, as well, where… we know who we're hyper-targeting, and then you build up from there once you've built that, right? And that takes time. I'm sorry to tell you, there's no shortcutting to that. But this is also where we work hand-in-hand with ads, right? From an advertising perspective, you know, we're, you know, paid ads is your world, it's not my world. But… the… we're complementary to what you're doing, right? You want to make sure that you're at the top, and that's one way to help, you know, reinforce that brand message at that very important real estate at the top of these programs, right? ChatGPT is first rolling out its ads program. I'm actually not even sure if cannabis is allowed to advertise through OpenAI. I'm actually not even sure if there's… if they put out regs on that yet. Has anyone… have you tried? Is that something you're trying or looking into, or… What's going on there?

Jake: I'm not sure, maybe? Yeah.

Stella Morrison: so new. No, it's so new, right? It's one of those things where I think I saw, like, one or two ads, and it was, like, it was, like, PetSmart or something. It was, like, some very, very big company. It was… it was not a cannabis ad, but, like, there's also, like, also other features for shopping, and it's just like, well, is ChatGPT going to allow access to that, too? It's… these are things that we are… Learning about as we go, and… it's gonna take some experimentation, right? But going back to just, you know, the… making sure we know who we're talking to and reinforcing, like, that ownership at the top, right? It's… you know, the ads are bringing people in the door, and then the content's what's converting them, right? It's… they're interested, oh, you're talking to me, right? You're talking to the tired mom, you're talking to the marathon runner, you're talking to… the stoner who wants Cheetos, right? You're talking… you are talking to me. These… these two elements go hand in hand, and, you know, I always recommend, pay a mix of paid organic and PR for anyone who's, like, first starting to break in. That's the mix I have, that's the mix I've always seen work, and I recommend it for anyone who's trying to, you know, break… break in somewhere.

Chapter: Closing and How to Reach Stella

Jake: Great. Definitely touch.

Stella Morrison: Like, four things at once in that time.

Jake: Yeah, well, I know it's… we're trying to cover a lot of ground in a short amount of time. It's multiple big topics that are interleaved, so we are running up on time now, though. I wanted to take a minute for you to explain… I think we've… probably have a general idea of what you do, but kind of content, for people that are interested in hearing more, learning more, potentially getting, some advice from you, how can people get ahold of you on the internets, and who is your ICP?

Stella Morrison: Yeah, so, we love talking to, folks in emerging states, in newer markets, and in restrictive states, so, any… if the marketing is extremely challenging in your state, come on, Ohio, either in Cleveland. I would love to talk to you. That's definitely a target of ours. And we are… we like to work with people who understand the importance of visibility, and how difficult that visibility is in the world, and how to break them in. That works at anybody in those restricted states, restricted markets, we, we definitely want to be, talking to you. I am pretty active on LinkedIn. If you find me, I think I'm tagged in this event, so just look for Stella Morrison, I'll probably, pop up. As of now, that will change soon. My hair is straight in the photo, promise it's me. And we also, on Instagram, Cana underscore content, you can see a good library of just our creative work. there, our… our, Chief Creative Officer's incredible, like, just… her brain is magical, and just how she works there, you can see… you can see a lot of that as well. And on the internet, RakanaContent.co, the, the contact button box will go to our team, or just find me on LinkedIn, and we can I'd take it to the conversation, but if you're not in for 3 hours, you might regret that.

Jake: Fair enough. Well, Stella, thank you for all of your wisdom and experience today. It was, like I said, I could… we could go all day on some of these topics.

Stella Morrison: I really could.

Jake: I'll have to do a follow-up conversation and pick a more narrow topic, which might work. I go on tangents, so we'll probably end up on another topic anyway. But again, thank you for your time. I appreciate you, spending it with us today.

Stella Morrison: Absolutely, Tang, just make for a great conversation. Thank you for having me, Jake, I appreciate you.

Jake: Yeah, have a great day.

Stella Morrison: Thank you, bye-bye.

Jake: Right.

Read More

Featured Speakers

Stella Morrison
Stella Morrison

Stella Morrison is a cannabis brand and content strategist who has spent years watching brands try — and fail — to outrank a bad reputation. Her lens is brand fundamentals and generative-search visibility: she helps cannabis companies fix the messaging, positioning, and reputation foundation that AI search now depends on before a dollar goes toward optimization. Having built content libraries for established multi-state operators and worked across the cannabis marketing ecosystem, she brings a rare combination of brand discipline and technical search fluency. That combination is exactly what operators need in an era when ChatGPT, not just Google, decides whether your brand shows up at all.

Related Cannabis Podcasts

Webinar Highlights

Related Articles

No items found.

Other Ways to Enjoy the Cannabis Podcast

Listen to the Audio
Download the Transcript
Follow us on LinkedIn

You Can't SEO Your Way Out of a Bad Reputation: The New Rules of Cannabis Search

08/06 | 11am PST | 2pm EST

When someone researches your dispensary, they're not just Googling you anymore. They're asking ChatGPT, reading the AI Overview, scanning Reddit, and skimming reviews before they ever land on your website. And here's the part that should keep you up at night: 97% of consumers read online reviews before choosing a local business, and nearly 60% of Google searches now end without a single click. The machines deciding whether you show up at all are summarizing everything the internet says about you. You can rank #1 and still lose, because you can't SEO your way out of a bad reputation.

For years, search was a game of showing up first. That era is ending. Large language models don't just index your website — they synthesize Reddit threads, review platforms, social mentions, and press into a single answer about who you are and whether you're worth the trip. If that picture is inconsistent, thin, or negative, you become invisible before a buyer ever compares you to the shop up the block. In cannabis, where competition is brutal and brand fundamentals are often an afterthought, this is where deals are won and lost. Stella Morrison joins Cannabis Marketing Live to break down the new playbook of branded search — and how operators build a reputation the algorithms actually reward.

What You'll Walk Away With

  • The "reputation everywhere" map — the specific off-site surfaces (Reddit, review platforms, social, editorial, GBP) that shape how LLMs describe your brand, and which ones cannabis operators ignore at their peril.
  • Message consistency as a ranking factor — why conflicting or wishy-washy descriptions of your business confuse ChatGPT and make you "appear weaker," plus how to lock down one strong, repeated brand description across every surface.
  • The "build on concrete, not quicksand" audit — the brand fundamentals that must exist before SEO or GEO spend does anything, and how to spot when you're optimizing on a broken foundation.
  • How LLMs actually read your brand — a plain-English explanation of how generative search pulls from archived and real-time sources, and how to reverse-engineer what the models already believe about you.
  • The budtender-to-search connection — how in-store experience and review sentiment flow directly into your AI-search visibility, and why reputation management is now an SEO discipline.
  • A 30-day starting move — the first concrete step to take this month, whether you're a single-store operator or an MSO with dozens of locations.

Why This, Why Now

AI Overviews now appear on roughly half of tracked Google queries, and zero-click search jumps to over 70% when they show up. Buyers are forming opinions inside AI answers you don't control — built from sources you may not even be watching. Cannabis brands that get their fundamentals and reputation aligned now will own the answer box; the ones that keep chasing rankings alone will quietly disappear from it.

Speakers

Stella Morrison
Stella Morrison
Founder & Content Strategist | Canna Content