The B2B Podcast Index
Index
All categories
MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
MethodologySubmit
Best of:MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
An independent project byFame
SearchBest episodesGuestsInsightsMethodologySubmit a podcast
Index/Marketing/Content and Conversation
Content and Conversation artwork

Inside HubSpot's Category-Leading GEO Program w/ Aja Frost

Content and Conversation · 2026-06-24 · 56 min

0:00--:--

Key moments - from our scoring

Substance score

62 / 100

Five dimensions, 20 points each

Insight Density13 / 20
Originality12 / 20
Guest Caliber16 / 20
Specificity & Evidence11 / 20
Conversational Craft10 / 20

HubSpot experienced a significant decline in blog traffic as AI models began answering common search queries directly, prompting a strategic pivot toward what Frost calls "B2B2C" (Business to Bot to Consumer) optimization. Rather than abandoning top-of-funnel content, HubSpot refined its approach by explicitly connecting educational content to product capabilities - ensuring that when ChatGPT, Claude, or other LLMs synthesize information, they mention HubSpot as a solution. The company now tracks visibility at granular levels (4-10 prompts per buyer journey stage) and measures success primarily through citations and LLM references rather than direct website traffic. Beyond owned content, HubSpot expanded its creator program and affiliate strategy to account for AI citation influence. Interestingly, audience size and influence correlate poorly with citation likelihood - instead, original data, clear structure, and easy-to-parse formatting drive LLM mentions. The company recently restructured affiliate commissions to account for citation value separately from conversion-driven revenue, reflecting the outsized importance of being visible within LLM contexts.

Key takeaways

  • →HubSpot's biggest predictor of purchase intent across all customer segments is whether prospects used AI to evaluate solutions, making LLM visibility the primary KPI for their content strategy.
  • →Top-of-funnel educational content now requires explicit product connections (e.g., 'you can segment emails using HubSpot') rather than generic answers, because LLMs will cite and repeat these specific recommendations back to users.
  • →Creator and affiliate program success is no longer determined primarily by audience size or conversion volume - original data, clear formatting, and LLM-parseable structure are the strongest predictors of citation visibility.
  • →YouTube optimization for LLMs focuses entirely on transcripts and metadata, not video consumption, allowing HubSpot to leverage its owned channels more effectively than paying external creators.
  • →Citation value must be separated from conversion-driven ROI in affiliate commissioning models, as LLM references influence discovery far earlier in the buyer journey than traditional metrics capture.

Guests

Aja Frost

Topics in this episode

LLM visibility and citationsAEO (AI Exploration Optimization)B2B2C content strategyHubSpot blog traffic recoveryCreator program optimizationAffiliate commission restructuringYouTube transcript optimizationLinkedIn and Reddit creator partnershipsBuyer journey mapping for AIProduct feature-level visibility tracking

Questions this episode answers

What percentage of HubSpot's customer conversions are influenced by AI discovery?

LLM usage is the single biggest predictor of purchase intent across HubSpot's customer segments (both upmarket and downmarket), and self-attribution shows that after jumping from ~5% to ~40% of traffic self-attributing to LLMs, the influence is substantially higher than direct citation counts suggest.

How should companies structure top-of-funnel content for LLM visibility?

Educational content must include explicit connections to your product or service - for example, stating 'you can segment and tailor emails using HubSpot's email marketing software' within explanations of best practices - rather than relying on generic answers and CTAs, because LLMs will cite and repeat these explicit recommendations back to users.

Do audience size and follower count predict how often creators get cited by LLMs?

No - HubSpot found that influence and audience size are far less correlated with LLM citations than expected; instead, original data, clear structure, and easy-to-read formatting are the strongest predictors of citation performance.

How does HubSpot track its visibility in LLMs?

HubSpot maps 4-10 prompts for each stage of the buyer's journey (from problem awareness to head-to-head comparison) and drills down to visibility at the product feature level, tracking both citation share and what LLMs actually say about HubSpot, rather than relying on citation counts alone.

How often do LLMs crawl and index new content?

HubSpot observes LLM crawl activity within hours of publishing new content, indicating that maintaining fresh, updated content is now critical for LLM visibility.

What our scoring noted

Our reviewer’s read on each dimension, with quotes from the episode.

Insight Density

13 / 20

The episode delivers a meaningful cluster of non-obvious operational insights - citation visibility vs. citation share distinction, the product-based team restructuring that bypasses mid-funnel attribution, and the finding that audience size doesn't predict citation volume. However, there is meaningful filler, circular confirmation exchanges, and several points that simply rehearse well-known AEO discourse.

we are finding that not only is it um, most customers but it is the single biggest predictor of purchase intent across segment
if you are optimizing and writing for bots, then you really do need to spell out how your product can help someone do this thing

Originality

12 / 20

A handful of genuinely fresh framings emerge - B2B2C (business to bot to consumer), the counterintuitive creator citation finding, and the property-to-product team restructure as an attribution workaround - but much of the conversation recycles familiar AEO talking points about ChatGPT disruption, Reddit's Google deal, and query fan-out.

I call it, um, B2. B2C. Business to bot to consumer. Your primary audience becomes the bot.
audience size. Yeah, it doesn't predict how many citations someone is going to get

Guest Caliber

16 / 20

Aja Frost leads a 40-person SEO/AEO function at HubSpot, a genuine category leader in B2B content, and has been running structured AEO programs since early 2023 with real budget, real team decisions, and real data informing a commercial product - this is a senior practitioner at scale, not a thought-leader circuit guest.

The SEO AEO component of it is around, uh, 40
we go down to the feature level and we map out between four to 10 prompts for each stage of the buyer's journey from I have this problem, but I don't know how to name it to. I'm comparing, you know, multiple solutions head to head

Specificity & Evidence

11 / 20

There are useful concrete data points - team sizes, the three dominant citation platforms, 4-10 prompts per feature, the creator program's 4-5-year history - but the most commercially interesting data (citation ROI model, actual visibility percentages, revenue attribution) is explicitly withheld, and many answers default to 'it depends.'

There are three sites that we see show up in the prompt set that we are tracking citation wise over and over again. LinkedIn, YouTube and Reddit
my total team is around 65 folks. The SEO AEO component of it is around, uh, 40

Conversational Craft

10 / 20

The host has genuine curiosity and lands some good follow-up probes (on UX changes, creator citation characteristics, Reddit community mechanics), but he consistently fails to push back on unsubstantiated claims, answers his own questions at length, and opens with extended flattery that sets a soft tone throughout.

HubSpot has always been known as kind of category leading to me from a strategy standpoint in the search world. So a lot of people would look to your team for kind of like industry leading best practices
okay, so. Oh, that's an issue. I feel like it's a good time to ask you, how does HubSpot track its own prompts?

Conversation analysis

Computed from the transcript - who did the talking, and the words that came up most.

Share of words spoken

  • Speaker A72%
  • Speaker B28%

Most-used words

hubspot52content45reddit36sense21community21team20visibility20marketing19llms18google18product18folks17long16blog15citations15makes15

Episode notes

Ross Hudgens sits down with Aja Frost, Sr. Director of Global Growth & Paid at HubSpot, to open up the playbook behind one of the most advanced GEO/AEO programs in SaaS. Aja shares how HubSpot has been investing in answer engine optimization since early 2023 and how the company turned a very public blog traffic decline into a "Phoenix from the ashes" story, becoming the single most visible CRM in LLMs. They get tactical on writing for the bot (her "B2B2C" framing and "tofu with a twist" content strategy), why visibility beats citation share as a KPI, how HubSpot tracks prompts down to the feature level, and why audience size barely predicts who gets cited. Aja also breaks down HubSpot's creator and affiliate strategy, the durable (and not-so-durable) ways to earn off-site mentions, a deep dive on doing Reddit authentically, and how she reorganized her 65-person team around products instead of web properties.

Full transcript

56 min

Transcribed and scored by The B2B Podcast Index.

Speaker A: Foreign.

Speaker B: I recently saw your AEO sensor and thought uh, it was amazing.

Speaker A: Thank you.

Speaker B: Yeah, beyond that intro, the fact that you all are doing more and more came top of mind for me to bring you on. I've seen your, you've had some great talks come out recently and some content online. You did an air ops webinar that was really good as well. Well, HubSpot has always been known as kind of category leading to me from a strategy standpoint in the search world. So a lot of people would look to your team for kind of like industry leading best practices and you continue to stay ahead of the curve there. So the fact that you're obviously taking this GEO thing, AO thing quite seriously made me think that you're probably cutting edge on that side. So would love to kind of deep dive into your practices today and, and go further there. So m excited to have you here.

Speaker A: Yeah, I'm excited to open up the playbook and show you what's in it. Yeah, we have definitely been deeply investing in what we call AO since the really beginning of 2023. I remember when ChatGPT, um, you know, kind of broke through and I actually just went back and I found a memo that I wrote to my leadership team um, around that time that said, you know, like, here's what this means for HubSpot, here's what this is going to mean for our customers and here's what we should do about it. Now I didn't get everything in that memo right. But what I called that I think has turned out to be quite accurate is discovery, uh, has moved to LLMs. Um, that is the primary way that I think most folks are now discovering and evaluating products. And if your company isn't thinking about how to optimize not just your visibility but the specific things that LLMs are saying about you. Yeah, it's like you're not on the first page of Google, you know, like the, the paradigm has really shifted.

Speaker B: Yeah, on that exact comment, one of the interesting things, I've literally had a conversation last week on your point of like most, we have a lot of B2B clients. I know you likely do at HubSpot. Um, you're mostly B2B right? B2B. Small business, B2B. Yep, yep. Um, so the kind of a refrain I like to give is like we had 5% of our traffic LLMs and then like at a certain point 40% were self attributing from LLMs.

Speaker A: M MHM.

Speaker B: Are you finding a percentage is pretty high? I'm sure it's pretty high for you, your customer base. Like what's the, the pulse there?

Speaker A: Yeah, well it's interesting that you say self attribution because that's definitely unfortunately the most accurate way to understand the true influence that LLMs are having on your funnel. If you look at the direct numbers, it's going to vastly under report. Yeah. So I think that that's what most folks should be doing. They should have some sort of um, self attribution motion. But yeah, we are finding that not only is it um, most customers but it is the single biggest predictor of purchase intent across segment. So whether you're upmarket, downmarket, HubSpot serves, uh, both then if you used AI to evaluate us, you're likelier to buy. Um, and we really attribute that to our fantastic presence in LLMs which yeah, is something that we've been really cranking on.

Speaker B: So on that note, the HubSpot storyline you all crush for a long, long time on the blog. One of those category leading things I'm referencing where you had ton of traffic there m somewhat public story is yeah, some people publish some things that that traffic went down a decent amount to that blog and you're not independent and alone in that. Of course many SaaS companies have seen a very similar story um, to that, to that blog. But some kind of correct me if wrong but it feels like with this GEO thing there's been a little bit of Phoenix of the ashes from the ashes moment. But what is the current kind of status of that blog setup and maybe you could speak to it.

Speaker A: I love that analogy. Yes, it was a very public, well documented, serious decline in traffic. Um, painful at the time as you can imagine. But ultimately I think turned out to be a blessing in disguise because we lost all that traffic from pretty low intent, broad keywords like what is content marketing? Um, and I think that loss of traffic really predicted what was going to happen to most companies publishing educational content online, which is AI intercepts that traffic and if someone is getting that answer in an AI overview or AI mode or Claude or ChatGPT, I mean, you know, pick your poison. Um, they're never making it to your website. And so much earlier than most companies we had to think about okay, how are we going to replace that demand and how are we going to optimize for a world in which um, our content is having an indirect influence on the buyer, our content is reaching the buyer in a different modal, in a different pain than on our website. So we've become through all of our work the number one most visible CRM in LLMs. Um, which I'm really, really proud of.

Speaker B: Uh, hyper competitive space.

Speaker A: Hyper competitive space.

Speaker B: And a lot of I randomly Google that just for fun. That's how much of a nerd I am. So I'm sorry for diluting your search data. Yeah, I searched best CRM because I'm just like, this is the most competitive SaaS term there is probably CRM software and best CRM software. So I unfortunately diluting all that data.

Speaker A: Yeah, yeah, yeah, yeah. Hey, as long as you're clicking, I don't know how engagement data factors in

Speaker B: here, but I won't ignore HubSpot, I promise.

Speaker A: Uh, so, yeah, so I think it is a real, um, Phoenix from the Ashes story. And the other benefit of being much earlier to thinking about our blog in that way, I call it, um, B2. B2C. Business to bot to consumer. Your primary audience becomes the bot. I think post Google IO we should all be thinking about your primary audience becoming the agent. Um, but you're structuring content differently, you're thinking about the takeaways differently, you're evolving your UX differently. Um, um, and, and we had a head start on that work that we've then been able to pass on to our customers, both in terms of Playbooks, but also in terms of our product.

Speaker B: Okay, so on the, could you speak a little bit more to the UX side? Like what are some examples of that?

Speaker A: Yeah, I mean I think we used to uh, really think about what is a user going to feel when they land on this page, really curating the experience around that. And now I think we are much more optimizing towards how fast is the page, how easy is the page to parse.

Speaker B: Okay.

Speaker A: And, and that's a different mentality. Obviously if a user lands on the page, we don't want them to think, oh, I don't like this, I'm going to bounce. But, um, we're thinking about how the person is going to experience the content in the LLM of their choice versus browsing through our website.

Speaker B: Okay, so you've made the jump to like the row of the LLM, um is preferred here. Is that what I'm hearing?

Speaker A: I think it was, I think it was a, um, a matter of necessity. Okay, like, okay, if we're not going to get the human, how do we optimize for the LLM?

Speaker B: Okay, had you completely. Is tofu even a thing on the blog anymore? Like how do you think about your content strategy there?

Speaker A: I, I still think about it as Tofu, but I think of it as tofu with a twist. Okay, um, tofu with some chili oil maybe.

Speaker B: Yeah.

Speaker A: Because where I see folks go wrong and this is kind of nuanced, so I'm glad we're getting a chance to walk through it. But they think, okay, I'm gonna optimize for bots, gonna structure all of this for bots. I'm gonna lead with the answer. Clear, simple language, lots of data, et cetera, on traditionally high, um, high funnel topics like what is content marketing? And they do all that and they give the LM all this great information and they don't include any link to their company.

Speaker B: Okay.

Speaker A: And so Now Claude or ChatGPT is coming to that page, synthesizing all that wonderful information, giving it back to the user, and there's no connection to. Oh, and here's why you should use HubSpot if you want to do content marketing. And so it's critical that in any kind of top of funnel content for LLMs, there is a very clear, explicit connection to this topic and what your product or service can do. Literally, I'm talking like, like, um, you know, if one of the tent poles of great content marketing is, um, well, it's funny because it's all changing now. So this is getting very meta robots. Yeah, maybe I'll do email marketing because that's a, a little bit of a cleaner example. One of the tent poles of email marketing is introducing, um, segmentation and customization. Then you would say, and you can segment and tailor your emails accordingly using HubSpot's email marketing software. And folks, I think aren't used to being that explicit. I think they're used to like including some CTAs in tofu. Some CTA is maybe a link to an ebook where you get someone's email address. Now it's all connecting. Um, but I think if you're optimizing and writing for bots, then you really do need to spell out how your product can help someone do this thing is.

Speaker B: So is the thought process that. Have you actually seen HubSpot reference in the what is content marketing? Or is it more about influencing their understanding for the best content marketing software,

Speaker A: uh, we, we see. So we see citations for the what is content? But more importantly, those citations are influencing our visibility. And so the LLM is much likelier to say if you have that explicit link we have found, if you want, if you need to do X, Y, Z, then you can use HubSpot. Like the LLM will actually repeat.

Speaker B: Oh, interesting. So it's kind of like a supplemental element of a tofu, which I've seen for sure at times. Yeah, yeah, yeah. Okay.

Speaker A: I think another distinction here, and this is why I said it's nuanced. Um, so hopefully this is all coming through. But, um, another distinction is the primary KPI that HubSpot cares about is visibility. I mean, I think citations are helpful in that they can certainly influence what the LLM, um, says, like the words around, you know, your, your brand and they can drive some direct traffic. But most people are not clicking directly on citations. And so if you're just looking at how many citations you get in optimizing your for your citation share, you're going to miss the bigger thing, which is how often are LLMs referencing you and what are they saying?

Speaker B: Okay, so. Oh, that's an issue. I feel like it's a good time to ask you, how does HubSpot track its own prompts? Like, what are you looking at?

Speaker A: We drink our own champagne, we use House aeo. Yeah. And, uh, we track. So it depends on how important, um, how much of a priority we consider the future. But we go down to the feature level and we map out between four to 10 prompts for each stage of the buyer's journey from I have this problem, but I don't know how to name it to. I'm comparing, you know, multiple solutions head to head. And so we're able to then drill down to, um, visibility at the product feature level, all the way up to HubSpot as a whole.

Speaker B: So on those various funnel stages, I'm curious for the things like what is content marketing, what is email marketing, what kind of percentage is realistic for HubSpot to show up there? If that's how you're tracking it, it's

Speaker A: going to really vary by the product or feature. And, um, how much HubSpot is, how much HubSpot has a right to be associated with that concept. So for something like content marketing, we see a pretty good percentage for something that we are less traditionally associated with. Obviously that's where a lot of the work comes in, but it's going to be lower. So it's a real classic SEO now. Ao answer of it depends. Um, but what I can tell you is those, uh, that tofu visibility can actually be incredibly valuable because you're kind of shortcutting everyone else who's going to come later in the discovery process.

Speaker B: Oh. So they may obviously hear the HubSpot name first and then they're going to obviously be led down the funnel. Is that what you're saying.

Speaker A: Yeah, I don't, I don't think it's that fully linear. Like the, the classic funnel or buyer's journey has really exploded, of course, with LLMs. But what we see is that if you ask an LLM a question like what's the best for your CRM? It's going to give you a short list. If you get visibility in those educational what is XYZ queries, you might be the only solution mentioned. And so that's really, really valuable because before someone gets to the stage of what is a good xyz, you're in their head. They, they, you know, you become their front runner.

Speaker B: Okay. They may even like front. Yeah, they may even skip like a best X search.

Speaker A: Right.

Speaker B: And just.

Speaker A: Yeah, m. We see that a lot. You know, we do a lot of, um, user interviews because again, that's one of the best ways right now to understand how people are actually going from discovery in an LLM to becoming a user or a customer. And they have so much trust in these systems. You know, I can't tell you how many times people have said, claude is my best friend, ChatGPT is my, you know, yes, go people. Of course we know because we're doing it too. Yeah. Um, and so when your deeply trusted co worker, assistant, you know, however you think about your LLM, says, hey, you're exploring this thing. And did you know that there's a solution out there that, that's quite meaningful for people? And I think it explains, um, the, the really high conversion rates that most folks are observing from LLMs relative to traditional sources.

Speaker B: A lot of people are going heavily bottom funnel in their content strategy on the blog. Is there a ratio like. It sounds like you haven't abandoned tofu.

Speaker A: No, not at all.

Speaker B: So it's a pretty mix. Are you a kind of classical 1/3, 1/3, 1/3 or any kind of rough?

Speaker A: I think what it, what I would recommend folks do is look at where they're visible today and then reverse engineer that into content ratios. So if you don't have any coverage for those, um, evaluation decision questions, that is where I'd focus. I think that that becomes your top priority. But if you actually have pretty good coverage there, then that's when I would absolutely start introducing more of the, you know, tofu with a twist content.

Speaker B: Interesting. Makes sense. And what, what are the things I think I just thought about as a byproduct, a lot of the LM corpus of contacts is at least of today can be like a year dated. And I'M guessing it's crawling the web a little less often on those tofu prompts. So is that fair to say or. I, I don't know if you, what your experience is, but this sounds like for for work is getting my brain spinning. But it's possible you may have a lag on some of this or may maybe not. Maybe you're seeing it more commonly.

Speaker A: Uh, you know, for all of our content, regardless of where it sits in the buyer's journey, we see LLMs visiting it quickly after publication and getting a lot of crawl activity and, and I think of LMS as these kind of V readers. Um, maybe that will change. I can't help but think that that must be extremely expensive among all the other balances that LLMs are running right now.

Speaker B: Yes.

Speaker A: Uh, so maybe the dynamics of that will change. But we, we see um, LLMs visit new content within hours when we look at our server logs. Yeah.

Speaker B: Nice. Yeah, makes sense.

Speaker A: Uh, that has, you know, it's a double edged sword because there's always been this race to have the freshest content. I um, think we're in turbo drive now. Makes us look like what we were doing before. You know, we were the minivan driving down the street and now we're like the F1 racers around the track. Um, it, it becomes a true necessity to invest in AI powered content workflows in my opinion. And also, um, to be really thoughtful around what you most care about from a content perspective.

Speaker B: Makes sense.

Speaker A: Yeah.

Speaker B: Went down a rabbit hole because I couldn't help but ask those questions.

Speaker A: Yeah.

Speaker B: But one of my initial question list things I was super curious to ask you about was some of the maybe 201 aspects of Aogo.

Speaker A: Yeah.

Speaker B: And things that have been top of mind for us as we explore these areas too. I've seen personally for a HubSpot, some like paid creator campaigns on Hub, on uh, LinkedIn. Um, maybe you're doing some of that on YouTube. Do you think about that from an AOGO lens at all or, and how, if so how are you kind of directing that?

Speaker A: We absolutely are. Yeah. There are three sites that we see show up in the prompt set that we are tracking citation wise over and over again. LinkedIn, YouTube and Reddit. Not super surprising given the relationships that all three of those companies have, um, with the frontier models. But so we've had a creator program. Uh, oh gosh. And I'm afraid to quote the year because I might get it wrong, but I think it's been four or five years. So it's been A long, a long time. Part of our playbook. In the last two years we started thinking, okay, we've really just been uh, thinking about the ROI of this content from a reach perspective. And then a traditional returns, you know, they click on a link, they end up converting. How much do they bring in revenue? Now we are thinking about a third dimension which is how much is that content getting cited and how much is it improving our visibility. And we are increasingly targeting creators that perform really well from a citations visibility perspective, even if they do not, um, look like our typical creator in terms of reach or revenue.

Speaker B: Okay, so is that circling back to maybe or is that definition like they're cited within the prompts you care about or is it. You're also looking at just net citation. I know they're HRS measures net citations, which you said maybe it was a mistake in some, some ways. How, how are you judging that before you've engaged them?

Speaker A: We will do pilots.

Speaker B: Okay.

Speaker A: Uh, so we have a good sense of what gets content cited on each of those places and that becomes the template that. Well, first it informs the creators that we reach out to. So they may or may not have already been cited, but we can, um, you know, we have a, we have a sense of what a citable creator looks like and what citable content looks like. So we use that to target our outreach. And then, um, if it's a creator that we work with, we use that to frame our recommendations, instructions. We still give them a lot of creator freedom. I think that's very important with any creator. You know, they, they know their audience. Um, and they also, you don't want to ask someone to post something that doesn't feel natural to them. Um, but yeah, then, then you know, we see how content performs and if it performs well, then we can move into a larger engagement.

Speaker B: Okay, so what does. You said the characteristics of a citable person. Is that just off. I'm guessing influence is one of those. But what is there any other?

Speaker A: Interestingly, influence is far less correlated with citations than you would expect. That was one of the fascinating early findings from our research. Um, audience size. Yeah, it doesn't predict how many citations someone is going to get, which from a, um, business, you know, I guess advertiser perspective maybe a good thing. Fantastic. Oh yeah. It means you can get way more bang for your.

Speaker B: Yes.

Speaker A: Yeah, yeah, yeah. And, and to be clear, you know, it, it's not that, um, like there are a lot of creators that have fantastic, highly engaged audiences that might not get citations that we would very Much still work with. Yes. Um, but there's just this new category. Um, so yeah, it's much less about audience and m. Of course original data does well. Ah. A lot of the things that we see being successful for owned content, we also see being successful um, in social content. So original data, um, well structured, easy to read. Those things tend to predict citations.

Speaker B: Okay. Oh, interesting. How about final stage? Are you instructing or nudging them to do bofu mofu Tofu?

Speaker A: Yeah, great question. We've largely focused on mofu and bofu because I think that those citations are um, more impactful in terms of like on site it's really easy to do both. You can do tofu and you can do mofoubofu. But I think off site it's a lot harder to get those mofu bofu citations. And so we try to optimize for those.

Speaker B: Okay, that makes sense. How, how about YouTube? Is that uh, difference different in approach? How are you thinking about YouTube?

Speaker A: YouTube is unique in that the, the, the video itself is not being watched. Right. It's all about the transcript. And so we try to leverage that. We try to make sure that the, the transcript is optimized as much as possible. Um, that the meta and the title are optimized as well. I would say for YouTube creators, a lot more of what we're doing revolves around um, our owned channels and the channels in the media network because that gives us a lot to play with before we ever need to go pay someone else.

Speaker B: Yeah, yeah, you have that leverage that other people do not have.

Speaker A: Yes, exactly. Which again, I mean so many of the things we are doing, and this goes back to, you know, AO really equals marketing. The lines between the disciplines are blurry at uh, best. Blurry if non existent. Yeah, there are so many of these plays that we were running that as the um, tactics of AEO began to emerge, we saw the opportunity to expand, um, or tweak them for AEO versus you know, something that we had to start net new. Exactly.

Speaker B: Okay, yeah, that makes sense. So what I see a lot in YouTube is often not, not an own channel, um, which is a good transition to affiliate. So I see you of course have an active affiliate program, did a little research, seems to be uh, best in class and market and CRM especially. How do you think about your kind of structure of affiliate, how, how it influences AO and all that?

Speaker A: Yeah, another great example of something that we've been doing for a long time that we really saw the AO opportunity with. And even before YouTube videos, affiliate sites were getting so much play in AO citations, it just, by far and away, it was one of the biggest categories that we were seeing in our own data. The challenge, and I think there are a lot of folks trying to solve this problem, is if you commission an affiliate site based on the conversions that they drive the ASP of those conversions, you're really not counting the influence that that content has on LLMs. And maybe you can get away with that for a few months, for a few years even. Yes, I don't think that you can get away with that anymore. And we always wanted to be really fair and respectful to our affiliate partners. So we've spent some time experimenting with different models. Um, I'm really excited, I guess now. Two months ago, um, I brought on a new senior director of Paid media who oversees our paid and affiliate, including our AEO work. Um, and he came up with this methodology for assessing the value of a citation specific to HubSpot. Like, here's how much a citation means to us. Here's the ROI of a citation. And that is so helpful because it allows us to plug it into this existing affiliate model that we have and say, okay, here's how much we will pay for these. Um, so that really advanced, you know, how you bid.

Speaker B: Bid strategy.

Speaker A: Yeah, how, how we're able to think about allocation, how we're able to think about who to work with, et cetera.

Speaker B: Okay, nice. I'm guessing you don't want to go too deep on that model. I'm like, I want to know that model.

Speaker A: We can talk after the show.

Speaker B: Yeah, yeah, but makes sense. Is there anything else with the, uh, do you go, Is it like, um, you feel like incrementally, it's good to just keep trying to get new placements that may impact the corpus? Or is it like there's 20 publishers showing up for best CRM? Or maybe it's not even 2010 that are showing up for best CRM software. Therefore, all of our energy should go towards that 10. Or have you seen incrementality from, like, trying to get a long tail?

Speaker A: Again, I think that audience or traffic doesn't correlate with the value of a citation. And so, um, at least right now, the more brand mentions, the better. Yeah, I also see a lot of diversity in the, um, websites that show up. I mean, don't get me wrong, like LinkedIn, YouTube, Reddit, you're going to see those for most companies, prompt sets, if they're a software provider. Like, there are certainly some recognizable names that are going to show up again and again. But if you look like beyond that, you know, starting five, you're going to see a lot of diversity.

Speaker B: Okay, makes sense. It. Do you think that uh, my brain goes to. Is that like Chat GPT? I mean there's a lot of people saying they just use Google's results.

Speaker A: Yep.

Speaker B: Claude, do they just use Google results? Yeah, but that would sort of hint at the fact that they maybe don't as much or, or. Because where would that. If they're, if we're all kind of going off this trusted seed set, clearly someone's going off some non trusted seed set to influence the citation share, right?

Speaker A: Uh, I think so. I definitely think that uh, they're using Google and maybe to this day Bing. Um, although Google is very, very much in the mix, I think what is happening is from this starting query then, um, the LLMs are using. Are you familiar with Fan Out? Yes. Yeah, so I think they're fanning out into these niche queries where there is a lot less competition to rank, which is where you see the domains that you might not typically recognize getting pulled in as citations. Um, and for your audience, your audience might already know how Query Fan out works, but it's basically how an LLM synthesizes a response to a complex question, breaks it down into you know, five to 20 sub questions, permutations.

Speaker B: Yes.

Speaker A: Looks at the results, the Google Bing results, um, pulls together the answers, assembles one response. There is also, I mean OpenAI has been very public about their uh, goal of building their own search index. I think their original goal was to build one by 2027 and then they came back later and said yeah, that's not going to happen. Um, but one of the people um, on our blogging team, Amy Rigby, found some really compelling evidence of just how OpenAI was trying to do that. Which I can share the link with you if you want to include it in the link. Um, so this is all a long way of saying, uh, there is definitely a world where ChatGPT answers are pulling from Google sometimes ChatGPT index sometimes.

Speaker B: Yeah, makes sense and that's a good way. Not everyone's like me and just searching best CRM software. I wish they were just doing the long tail variations. Yeah, but yeah, I mean that's that. I think that's one of like major takeaway for people across these. It's just like that. It's clear. I don't know if that persists forever. I mean a lot of.

Speaker A: Yeah.

Speaker B: Have you found this in kind of early medium research? Some Affiliates have woken up to that and changed their con, their, their how they charge.

Speaker A: Yep.

Speaker B: And are charging more upfront. Is that sort of what you've seen some of? Yeah.

Speaker A: Yes. Well, I think that there was this scary moment for the affiliate industry when it really questioned its longevity because if folks aren't going directly to affiliate websites and then I think there was this pendulum swing of oh no, we're more important now than ever. And I think we're still in the very early stages of figuring out what that looks like. But there are a few companies doing cool things in this space to try to figure that out. Um, and you know I, I see a very interesting closed loop where you are able to um, reach out to publishers or peers from within your AEO tool of choice. You're able to log that. You, you got that mention via that outreach. You're able to see the impact on your visibility. You're able to attribute that to increases, increases in folks coming to you from LLMs and the incremental revenue like we're gonna get there. It's just kind of messy right now.

Speaker B: Yeah, makes sense. I'm sure you heard the news of another AEO vendor announced a partnership where they do some high scale outreach for mentions. A lot of our industry immediately was like, this is basically like paid link outreach in a way. Yeah. Google, uh, and their geo guidance said inauthentic off site mentions. They might have said that maybe you remember. Exactly.

Speaker A: But yeah, they basically said like if you're buying brand mentions artificially, it's not going to work.

Speaker B: How, how, yeah, how do you think about that in, in prac. Like what does that mean in practice? I'm trying to think that through.

Speaker A: You know, I think it's hard because there's a lot of Google guidance that has later turned out to be m. You um, know, maybe not the full truth or not how we see things operating in, in practice. So I think it's always worth taking Google's guidance with a grain of salt. On top of that, you know, Google can speak for Google, Google can't speak for anthropic and OpenAI, um, perplexity. But I think what is very much true is if you are reaching out to a random website that doesn't have a connection to your space and you're buying mentions. Yeah, it's, it's not, none of the LLMs are going to like that. You can't poison the well. They're going to figure out a way to stamp that behavior out the same way you know, Google has done so much of that over the years. I think that if you are finding the people that are in your space that are relevant, you're reaching out to them with personalized outreach, you're finding something of value, you're making their piece hopefully better, more correct, more up to date, more comprehensive. I think everyone wins. Um, it's the harder, slower way to do it. But I, but I do think that that's a uh, approach that has, that's durable.

Speaker B: Yeah.

Speaker A: No matter how the elements evolve. Yeah, yeah.

Speaker B: And I've thought about it. I mean I think HubSpot would be a great example of this. Uh, how I describe naysayers of affiliate influencing ll is like I think you need to be a very strong business to maintain those positions on these because you HubSpot I'm sure converts better than other people, has strong brand awareness is going to get higher click through rate those combinations. Even if you're technically paying the publisher for conversion is not something Joe Schmo small CRM can do or keep up. They could do it for a month or two maybe.

Speaker A: Yeah.

Speaker B: Uh, so like if people are paying for placements they, if HubSpot did show up, every publisher would swap them for HubSpot for most 99% of CRMs, I uh, would guess.

Speaker A: Yeah. And I think you know, if we're talking about Joe so the, the competitive of the, the competitiveness of the space really dictates the pace. And so if you are a typical business, maybe you're a brick and mortar business, maybe you're a small B2B UM company. You may not need to go to those lengths to drive your visibility. You might be able to get what you need from owned content on your site and um, a thoughtful brand off site strategy and you may not need to go into the um, pay to play world. And so that's what you know people sometimes say they're like, well there's so much to aeo. How do you expect someone to do it all? I'm like well you know your, your resourcing is usually proportional to the level of impact you need to drive. If you're a huge multi, you know, multinational company then you need to be playing at a very high level. If you're, if you're just trying to win in like one small area of Southern California or uh, Austin, you probably don't need to invest as much to win.

Speaker B: Yeah, it's true. What I've heard uh, is a variant of that is because you are local, your ability to manipulate those rule results with a uh, paid Motion. It's also possible, potentially we don't have a lot of those clients, so I can't speak to it. But yeah, um, that I've heard about more of that of like LM poisoning and whatever and maybe like Barbershop Miami beach, you know, not, not any clients we're helping right now.

Speaker A: But yeah, Siege Media does not do this. Oops.

Speaker B: How about, um, number three in that portfolio? Right?

Speaker A: Yes. Oh, Reddit. We could spend a whole, we could do three hours just on Reddit. It's so unique. Um, you know, marketers have been trying to infiltrate Reddit for a long, long time because Reddit, Reddit's user base has, has long been active. We started to see folks append Reddit to the end of their queries, which, you know, really signaled how much folks trusted the content on Reddit. And then of course, Google and Reddit signed a deal and now I can barely search. I don't think I can search a commercial query without Reddit showing up. Whether or not I've added not even

Speaker B: negative Reddit.com that even works anymore.

Speaker A: It'll sneak in. Yeah, yeah. Um, so, uh, HubSpot has approached Reddit, I would say, you know, on like the softest defeat, because it's just so easy to go wrong. Um, and where we have really concentrated has been our own HubSpot community. R. HubSpot, because that is a group of people that have said, yes, I want to be here, I want to talk about HubSpot. And so we've concentrated on making that a hyper engaging place to be the really cool thing that we found. So we track how many people are talking about HubSpot within that community as well as across Reddit. There is a very clear relationship between those two things. When people talk about us more in our owned community, they go out to Reddit at large and talk about us more. And so that's kind of our way of doing Reddit in a way that is authentic and abides by Reddit's rules of engagement. We just make our HubSpot awesome and then those folks go out and they talk about HubSpot more across Reddit.

Speaker B: Makes complete sense because I'm a Redditor and now I think about just the cycle of like you engage on a threat. It's just like social media algorithm 201 or whatever. They will feed you more content of that type.

Speaker A: Yep.

Speaker B: So it's almost certainly what's happening in your case. Yeah, I would guess. Right.

Speaker A: Like, yeah, there are probably a bunch of ways that it's Happening. Let's say I went into our HubSpot and I'm like, oh cool. You know, HubSpot did an AMA with Ross of Siege Media. And you know, they, they were talking about AO and some folks talked about how they're doing a O with HubSpot. And now I'm browsing my feed and a uh, post from digital marketing comes up. Maybe I'm not even in digital marketing, but it's talking about ao and I'm like, oh, I just learned some cool things about AEO from HubSpot's AMA with Ross, you know. So, um, it's a really, it's a really beautiful, um, cascading motion.

Speaker B: Nice. Uh, uh, as I thought, it makes complete sense. I mean basically the two recommendations I've made of companies is your own community, which is that one second potentially paid social amplification. Are, are you doing that as well?

Speaker A: Yeah, yeah, we, we have long thought about Reddit, I think as a place to drive conversions, um, because a lot of our paid spend is optimized towards driving conversions. But we really didn't see that work on Reddit. Um, interestingly. And so we have more and more shifted to an impressions based strategy where, yeah, we're just reminding people that we exist. Um, but I don't think we have necessarily cracked the code on how that plays with the community strategy yet.

Speaker B: Okay.

Speaker A: Yeah, if you, if you have any tips.

Speaker B: Yeah, well my thought process has been if you drive more conversions with Redditors, therefore down the line those Redditors will be on Reddit. And then when mentions come up and you have a great product.

Speaker A: Mhm.

Speaker B: It's a long, basically long tail play or long term play. But the, but yeah, the catch 22, hypothetically is what you said. If you can't convert those users there, then that may be challenged in some ways, I suppose.

Speaker A: Yeah. How have your conversion rates if you can say, compare.

Speaker B: We're early there. Transparently we're working on adding Reddit. Um, so this is me asking on that side, but on the community side to that same kind of exploratory. This is interesting to me because I'm thinking about it for our clients is uh, how. I mean HubSpot's a ubiquitous brand in a lot of ways. Do you have specific motions to get people to that community? Like how are you channeling visibility there?

Speaker A: Yeah, a lot of it. I mean as Reddit has grown and grown in large part thanks to Google, um, you know, we've just seen more users on Reddit and then the more we make HubSpot a really well moderated, engaging community that just amplifies. Um, there are some, there are some motions where you know, if you're a top, if you're interested in becoming a customer advocate, which is a really thoughtful program that our community team runs, um, you know, we'll, we'll recommend that you join both HubSpot's owned communities and Reddit. But I think most of it is honestly just people HubSpot, uh, users or potential HubSpot users being on Reddit and organically coming across the community in their feed.

Speaker B: Okay, nice.

Speaker A: Yeah.

Speaker B: From a content perspective on that, you mentioned AMAs. Are there any kind of like go to things that community loves that you find, like content types?

Speaker A: We should also talk about what to do if it doesn't make sense to create your own Reddit.

Speaker B: Yes, yeah, community.

Speaker A: Because I could hear the people in my head saying it doesn't make sense. Right. Joe is not creating his own, um, but definitely having recurring threads and that's probably something as a Redditor yourself, you know, and, and recognize in some of the most active, well run communities, give people a sense of what to expect. Um, um, so it could be what did you learn about HubSpot this week or what were your wins this week? Um, um, definitely. AMAs can either be crickets or very engaging. There's kind of no in between. It's very dependent on if you choose a topic that will resonate with the AUD and you know, you choose a participant who um, really knows their stuff. Um, but I think some of the work that's been most impactful is the more invisible stuff which is um, making sure that folks are leaving relevant content and that you're not seeing a spate of really repetitive, low value content. Um, and having a really good co moderator and, and some of the stuff that you won't see as much when you just browse a feed.

Speaker B: Do you actively. Sure. At least pay attention to it. But is there any like what's the thought process of HubSpot mentions off community in Reddit in terms of like HubSpot comes up in slash. I don't even know if it's engaged community but CRM. MHM R CRM. Um, someone says something about HubSpot. I don't. It could be either way sentiment wise.

Speaker A: Yeah, yeah, yeah.

Speaker B: How do you, do you respond?

Speaker A: Do you jump in? Oh, do you respond? Okay. Yes. Um, normally we wouldn't respond if it's not our own community because again and for anyone getting into Reddit, if You're not already on Reddit, spend some time observing before you jump into the fray. But if it's not your community and someone calls, calls you out and you're like, well, actually, you know, it just can come across really not great. Um, so we, we typically would not. The exception might be like if, if someone says something and we're like, oh, we can help you with this. Like, if they're like, I've tried to talk to my, you know, CSM three times and I haven't heard from them, that's when we might say, step in and say, okay, check your DMs, like, we'll see if we can sort this out for you. Um, but Reddit does skew more negative in general, and I do think that that's why it's interesting. That's both why people trust it, because it feels unbiased to them. You know, even though it might be biased in a negative direction, but it feels true. Maybe because the rest of the Internet can lean.

Speaker B: Yeah, it's not at all just so positive. Who should not have a community?

Speaker A: Oh, yes. Um, probably most companies.

Speaker B: Yeah. Okay.

Speaker A: Yeah, yeah. I think, you know, it's, it's not helpful to have a community that's not active. Um, I think if you see if there is, uh, interest, organic interest from your customer base and connecting with other customers, and that's something that you're just going to observe probably anecdotally, um, like, oh, you know, have you talked to anyone else who's experienced this or tried to do that? Has anyone else run into that type of problem? You know, if you see any sort of organic selling organization, that's a good sign that there is enough there, there for a Reddit community. But mostly, um, I think that won't be the case. I don't think that means your brand shouldn't be on Reddit. That was a double negative. Your brand should still be on Reddit. But think about those communities that are the ones where your name might come up. Um, CRM is definitely one of them. For HubSpot,/Sales,/Digital Marketing,/ marketing, there's a bunch of marketing ones. Um, and add value, add value a lot before you ever mention your name. So participate like a true, you are a true member of that community. Um, give, give, give, give. And then once you've really solidified yourself as a helpful presence, that's when you can say when someone has, you know, when someone is dealing with something where your product or service can genuinely help, help. And, and by the way, here's how my product genuinely help.

Speaker B: Yeah, right. Yeah, makes sense. That's one way I've been thinking about the, uh, the advertising side is I'm sure you were doing this, but if you're running ads against that community, you at least could, maybe that could help the top of mind point of view when those threads occur. Of course, it's ROI model thing.

Speaker A: Yep.

Speaker B: But maybe like affiliate, you don't need to run at a super, super high profit margin if you're in a market where a lot of your customers are buying.

Speaker A: Yeah. I think Reddit ad functionality is evolving quite a bit. It's something, you know, it looks different now than it did a year ago. And so I think over time, as it becomes more sophisticated, my gut is that the math, you know, hopefully will

Speaker B: math better make a lot more sense.

Speaker A: Yeah, exactly. The math will math better.

Speaker B: We'll see. We'll see. So, yeah, you've been m making this move for a while. What does the HubSpot team structure look like in the context of AoGeo?

Speaker A: We just reorganized and it was something that my team leads and I put a ton of thought into. I have a large team. My total team is around 65 folks. The SEO AEO component of it is around, uh, 40. So you know, you're not, you're not making any decisions about 40 people overnight. And it's this balancing act of SEO still very much matters. So we don't want to, you know, just rip and replace all the SEO structure. So in, in the, in the, like, we're 100% SEO world. Um, my team was organized into the various parts of HubSpot's web presence that drove demand. So we had a team. Actually, the team that I initially started on when I joined HubSpot's SEO team, um, when it was like, I don't know, five people. Uh, the blog, uh, you have product pages, slash the website. That's your more bottom of funnel transactional content. You have academy. And that really stopped working in an AEO world because it goes back to this idea of you're not optimizing for the human anymore, you're optimizing for the bot that's giving the human information. Like, we're not trying to drive visibility for the blog, we want to drive visibility for HubSpot products and we might use the blog to do that.

Speaker B: Okay. Yes.

Speaker A: Does that make sense?

Speaker B: Yes.

Speaker A: Yeah. So rather than properties being the organizational framework, we needed to move to products and so what we have done. So there are two teams within the official title is Global Growth because it's SEO, ao. And then, as I said, some other stuff. Um, there are two teams within Global Growth. You have your English team, you have your international team. And within. Let's just go into English as an example. Within English, you have folks who are devoted to specific products.

Speaker B: Okay.

Speaker A: They are measured on two things. They are measured on the AI visibility of those products, and then the actual demand that we generate from those products. And so the middle is missing. It's like visibility, reach up here, conversions, revenue at the bottom.

Speaker B: Okay. Yeah.

Speaker A: And that helps us kind of leapfrog the attribution issue.

Speaker B: Okay.

Speaker A: You know that, like, is so endemic to AEO right now. Because if you have high visibility up here and you're generating conversions and demand down here.

Speaker B: Yes.

Speaker A: Like, we know that the middle is working.

Speaker B: Okay. Yes. Okay. Okay. Interesting.

Speaker A: That makes sense. And it also allows me to be like, okay, everyone can use every. What was starting to happen is we'd be having these debates about where content should live. Should it live on the website? Should it live on a product page, should it live on the blog? And, like, no one felt ownership because, again, they're not being gold on product visibility. They're being gold on how much traffic and demand their property is generating. It's just really like, it stopped working. And so now. Okay, if you are trying to increase visibility for our marketing product, use the blog. If the blog is the best place for that content, use product pages, use the website, create in a course on HubSpot Academy, work with affiliates. Like, you have full, um, control of all the different levers that you can pull to reach this goal of product visibility. And then product revenue.

Speaker B: Who owns, um, like, is there still ownership around? Like, could the blog just start looking really sloppy or, like, inconsistent in that context?

Speaker A: Great. M. Question. So we have a separate team, the media team that, um, owns the media strategy for our blog, our newsletters, YouTube, they're really like our right hand. Yes. Brand voice and, and the look and feel. And so they help provide that kind of view or eye of. Okay, here's how all this content comes together.

Speaker B: So we talked about Reddit and, uh, paid social. So are they. Are. Are there, like, a dedicated person in all of those, or are they going to the affiliate manager and being like, here's a thousand. Go figure this out for me. Or how's those channels? Yeah, not for them, for the partners.

Speaker A: Um, so the AO work via, uh, affiliates and creators. There are a few different teams working on it, but I would. It's much more of a partnership. The folks who are working with creators and affiliates. They really understand the landscape. They understand what's working, what's not working. And so it would be much more of a conversation. Let's say you're on the affiliate team. I would say, hey, Ross, I'm really trying to increase visibility for maybe the specific feature, and I've done some analysis on what's showing up. Uh, you know, here's what I'm thinking. Yeah, what do you think? And then you might say, oh, well, actually, I'm seeing something really interesting with this other thing. Let's bring it in. Like, let's try this content type. Let's try this tier of affiliates.

Speaker B: Yes.

Speaker A: And then we would build a strategy together.

Speaker B: Okay. Makes sense. Nice. Well, uh, this has been great. Anything I missed?

Speaker A: Uh, I think we did a real deep dive. No, this was fun. You asked me some questions that I hadn't gotten before, which I was.

Speaker B: That's good. Yeah. Um. Uh, I don't think there have been a lot of 201 deep dives, so I'd love to check in in a year and see what 301 looks like.

Speaker A: Hey, maybe we might be on 5. 0. At the rate things are changing, you

Speaker B: know, you'll probably have a, uh, an AI model that can walk in here.

Speaker A: Yeah. You won't have to fly me out to Austin. It'll just be a voice on a computer screen. Exactly.

Speaker B: Oh, yeah. So people should check out HubSpot AO.

Speaker A: People should absolutely check out HubSpot AO. Uh, so many of the learnings from my team have been directly incorporated into the product. So something that we didn't even get to is, um, we have a very close relationship to the product team that's building HubSpot AO. And basically, as soon as we found something that scales, we'll share it with them so they can figure out how to bring that to life in the product. Um, and they can connect with me on LinkedIn or I have a, uh.

Speaker B: You have a podcast yourself, right?

Speaker A: I have a podcast, um, with Kyle Denhoff, who is a leader on the media team that I just mentioned. It's called Field Notes. Uh, so you can check that out too.

Speaker B: Yeah. Awesome. Thanks for coming through.

Speaker A: Thanks, Ross, for having me.

Speaker B: Thank you.

More from Content and Conversation

All episodes →
  • What Your CMO Needs to Know About AI Search w/ Tom Critchlow
  • Eli Schwartz on Google’s GEO Advice and SEO Careers Today
  • Lily Ray on AI Slop, GEO, and What Actually Works
  • Is Content Syndication Worth It? w/ Stacker's Noah Greenberg
  • Building a Media-First Content Marketing Strategy w/ Devin Bramhall
Explore the best B2B Marketing podcasts →
All Content and Conversation episodes →