Subscribe
Share
Share
Embed
Always Be Testing explores the experiments, insights, and growth stories shaping the future of affiliate and partner marketing in B2B SaaS. Hosted by industry veterans, the show dives deep into real-world lessons from the people driving measurable impact at companies like Google, HubSpot, Ramp, Webflow, G2, and beyond.
Each episode uncovers what happens when today's most innovative marketers challenge assumptions, run smarter experiments, and build programs that scale revenue through meaningful partnerships. If you’re in affiliate, partnerships, or SaaS growth — this is your front-row seat to how the best do it (and what they’ve learned along the way
Tye DeGrange (00:01.144)
Hello, hello, hello. Welcome to another episode of the Always Be Testing Podcast. I'm your host, Ty Degrange, and I'm really excited to talk to Mike Malazzo today. Mike, what's up, man?
Mike (00:12.817)
how are you, sir? Thank you so much for having me.
Tye DeGrange (00:16.152)
Dude, you're you're looking good. You're looking strong. You got leather bound books behind you, you know, beautiful photos. It's it's game time.
Mike (00:23.323)
Yeah, well, yeah, they're my wife's law school books, you know, and I put them up so that I look a little bit more intellectual and erudite than I actually am when I do these interviews. So, you know, hopefully we fool a couple of your listeners here.
Tye DeGrange (00:32.462)
you're you're quite the intellectual in this space and give an assist to your wife for the for the beautiful books. I'm a fan. I have too many. I'm reading too many at once, I'm reading too many in general. My wife has to get them out of my house, so I don't read don't like clutter the house with too many books. So there you go.
Mike (00:58.041)
Nice. Well as long as your book smells of rich mahogany, then you're doing all right by me.
Tye DeGrange (01:03.554)
That's it, man. That's what we're trying to go for here. So let's do it. Mike is an amazing guest. You guys are in for a treat. We're gonna talk fact fiction, GEO, AEO, try to talk about what's the hype and what is actually real. Mike's had an amazing career in growth, in writing and observing of like all the trends that's happening in growth, in partner marketing, in AI, GEO in particular.
head of growth at forum brands in the past, writer of the Zero Clicks newsletter, GM and lead writer at Alephic currently. So he's got it all. Let's let's do this. Mike, what else do you want to share with the team that the group that I'm not I'm not getting right?
Mike (01:46.46)
No, no, I think you nailed it. obviously a lot of your listeners know about my time at at PayPal as well, where I was sort of the eleventh hour GM of Honey and working on our agentic commerce product. but yeah, here at Alephic I am the general manager of our media division, which comprises three different publications. one called Ride AI, which is the most premium event and newsletter in the autonomous vehicle space, a podcast called Forward Deployed, which covers AI and software engineering.
Yeah, and then a newsletter and event called Brand, which lives at the intersection of AI and marketing. and has our flagship event at the Time Center on November fifth for four hundred enterprise marketers who are thinking about what what the job really looks like in a world where L L Ms mediate brand preference. So yeah, excited to talk about all of that today.
Tye DeGrange (02:34.251)
Amen. Love it. Excited to hopefully I can make it to that event in early November. It sounds great, November fifth. kicking off what's something people might not know about you?
Mike (02:45.341)
Well, something people might not know about me. good question.
Tye DeGrange (02:48.493)
Besides leatherbound besides leather bound books, maybe something outside of the writing and and professional realm.
Mike (02:52.921)
Yeah, a couple different things. I've been involved in brachiocephalic dog rescue my entire life. I've been involved with bulldog rescues and just adopted a boxer here two and a half weeks ago who's sleeping at my feet. what else do I have that's kinda random? I believe I'm still my high school's record holder in the four hundred meter hurdles and eight hundred meters. because we'd have a, you know, not so prestigious track and field history.
Tye DeGrange (03:07.605)
That's amazing.
That's really cool.
Tye DeGrange (03:16.202)
Wow.
Mike (03:21.905)
I also think one of the greatest Well, you know, the times are if you know track the times are fine. I think my best sort of esoteric skill in the world is I think I can slice garlic thinner than anybody except maybe Polly and Goodfellas. so you know, it like liquefies in the pan on contact. So yeah, yeah, a couple little fun idiosyncrasies here.
Tye DeGrange (03:22.381)
Hey, that's that's pretty prestigious.
Tye DeGrange (03:38.761)
yeah.
Tye DeGrange (03:45.806)
A true Renaissance man and and you're thropping pop culture references, so you're really you're firing on all cylinders from from where I'm concerned. Well done. Well done. Try to weave in some more gr goodfellows references, or or maybe just Scorsese references in general.
Mike (03:53.581)
Yeah.
Mike (04:03.507)
it can be arranged. If you ask me questions about meta, I'll certainly have some good fellows' references for you. So, you know, we'll get
Tye DeGrange (04:07.917)
I I for what it's worth, I dropped a Scorsese reference earlier today, from in a professional, semi professional setting. So do that with what you will. departed. Departed. Yeah. Some some good stuff in there. That it's one of the more quotables quotable of his movies, if not the most quotable, in my in my humble opinion.
Mike (04:18.523)
Which one?
Mike (04:31.111)
You're being a little coy. Was it a was it a PG thirteen departed line or was it a departed line that we can't that we can't repeat for the more sensitive ears among
Tye DeGrange (04:38.571)
It was more inappropriate, which is tends to be my my style. So sorry in advance to the audience that you will not enjoy that one.
Mike (04:49.757)
He
Tye DeGrange (04:52.309)
well you you that was one of the better answers and with very little prep. so you know you you're let's get into the world of AI, right? You you're in this intersection of AI, media, commerce. What is something that's kind of changed your mental model around you kind of referenced at the beginning, a customer acquisition. in particular over the last, you know, twelve months, give or take.
Mike (05:02.929)
Yeah, correct.
Mike (05:20.517)
Yeah, so I think the most interesting question in our profession right now, that really nobody has an answer to, and I'm trying to kind of figure out my own worldview day by day, is really around what marketing to agents will look like. and what does our profession look like as agents start to sit between a marketer and the end customer. you know, I would say twelve months ago I was very skeptical that the idea of marketing to agents would even exist.
At Mike McNearney's MARC event last summer, I interviewed the CEO of Profound, James Cottawalder. and he was kind of espousing this world where the job of a marketer is really to persuade an algorithm. and I didn't quite call bullshit on it on stage, but I I certainly had a very palpably skeptical view to the audience. and while I don't agree with James on everything, I think he was directionally right about this idea that a real part of the job now.
is is is persuading an LLM as much as it is persuading an end person. what I would just kind of say on that though is, you know, I think there's this sort of fallacy in the marketing ecosystem that marketing to agents or marketing to LLMs is going to be this sort of very rational endeavor where you just create really clean markdown files and really clean catalogs and you pursue
whatever the latest AEO arbitrage hack is on Reddit, and you'll be able to kind of convince the models. I think that's wrong for a number of reasons. I think you know with each model release, and and this obviously was in the news this week with Dorakesh's post on AI developing civilizations and all that, but with each model release, the sort of LLMs are becoming a little bit more anthropomorphic, a little bit more human-like. they're
learning to kind of look for heuristic shortcuts that don't use as much compute. It's kind of what that means is that AI and LLMs are going to have brand preference the same way that people do. I'm actually writing a blog post right now that that's called How to Sell Lucky Strike to an LLM, kind of drawing on the madman illusion. But I think that's the really interesting question is how do you make an LLM or an agent develop brand preference? I think that's the highest order thing that everybody's trying to accomplish with
Tye DeGrange (07:28.125)
Great episode.
Mike (07:35.823)
AEO and GEO and a lot of the sort of affiliate tactics, but I haven't seen a lot of marketers kind of zoom out and and think about
Tye DeGrange (07:42.807)
Yeah. I I had a lot of things going through my head, but like maybe for people to understand like how much of that is happening now, it just to get people up to speed in your opinion and based on what you're seeing currently.
Mike (07:55.804)
Well, I think it's two parts, right? I think obviously consumer usage of LLMs goes up every day. you know, Google obviously is effectively becoming its own zero click interface, right? So the idea of the entire web being kind of contained within an answer engine is is very, very real for for a lot of people today. What is obviously more nascent is any like real sense of agentic commerce, right? Where an agent is autonomously purchasing something.
Tye DeGrange (08:14.241)
Yep. Yep. Yep.
Mike (08:24.911)
end to end. but an agent already mediating a lot of what you see between a prompt and between when you actually reach a a sort of human set of eyeballs, that is already, you know, sort of very prominent. right. If you think about, you know, the percent of Google searches that are that are strictly zero quick now, plus the kind of aggregate usage of Chat GPT cloud perplexity, you know, most I would say brand touch points at this
Tye DeGrange (08:42.753)
Yeah.
Mike (08:54.237)
point are touching some sort of an AI system before they are getting to publisher or brand content. So that part of it is, you know, obviously anyone working in the space feels it already. That part is very real. The obviously agenda commerce piece is is more forward looking, but this idea that how either ChatGPT, Google, to a lesser extent other Frontier LLMs kind of perceive your brand as dictating your fate, that is that is a real trend.
Tye DeGrange (08:56.011)
Yep. Yep.
Tye DeGrange (09:20.235)
I I think it's interesting too because in I mean this might be mildly spicy, but the we've kind of been through this in a way where for quite a while Google was that gatekeeper, right? Google was that search gatekeeper where it was like, okay, you're you're kind of having to bow do battle with or you're gonna have to contend with this intermediary, obviously very different. it as it evolved Meta, it it took it to a pretty
pretty high level in terms of the algorithm determining quite a bit of like maybe not if you were showing up at all, but maybe how effective that what that spend and that placement was, especially from an ad and from an SEO perspective. So I'm not saying they're the same, but is it that is it that natural extension of, you know, you're you're basically going over the next technology hill to get in front of customers.
Mike (10:18.159)
Yeah, I think what is underappreciated though is just how kind of weird sometimes the L LMs can be, right? There was not really like I think one search that's that's really fun to for people to kind of sort of appreciate the gravitas of this is if you search for a really expensive item in ChatGPT, car, new set of golf clubs, high end furniture.
It's very likely that there are going to be touch points in which ChatGPT is like recommending you buy a used car or recommending that like you look on resale marketplaces for for deals on furniture. Right? Think about that from Toyota or Honda or West Elm's point of view. Like, what the hell? obviously that could exist on Google as well, but it was a different experience where
You know, you'd be linking out to an external blue link that clearly was like a blog that suggested you buy refurbished, whereas in a AI experience that's all kind of within the the chat window. so there's like this slightly less in a sense commercial aspect to especially Chat GPT, but even some of the the Google Zero experiences that I think is sort of a new wrinkle for for marketers to navigate.
Tye DeGrange (11:33.57)
Yeah. I think that's super interesting. Especially because, you know, new waves are usually are often where you're seeing alpha, where you're seeing really good performance for brands. You know, I was lucky to be one of the first advertising on the Facebook news feed and getting really high quality, low cost downloads at a time, like a lot of us were in those early twenty ten eras where you're just like, This is like shooting fish in a barrel.
Mike (11:58.622)
Well so that's what's weird about this moment, right? Yeah, if you think about early Facebook ads, right, there was arguably six or seven years of arbitrage opportunity that a lot of great D to C brands were built on. Now, a lot of them raise too much venture capital and whatever else, but you know, you were able to basically pretty predictably put a dollar into the Facebook machine and get three or four dollars out. Very similar dynamic on Amazon, right? When Amazon sponsored product ads started to really scale in the mid twenty tens.
All kinds of early Amazon sellers were basically able to build brands well into the tens of millions of dollars with names that nobody had ever heard of on just paid advertising arbitrage. That now I'm not hands-on keys spending ad dollars at scale anymore, but you know, I still obviously follow the space. That doesn't exist yet with ChatGPT, right? ChatGPT's cheapest ads right now, I think, are $3 cost per click in their self-service platform. might be a little bit different on the managed service side.
They just have rolled out kind of the rudimentary conversion tracking that has the the product dialed in, but it's still a fairly opaque experience when it comes to which prompts you're targeting. So even adjusting for inflation and all that, this is not a ad platform that has a clear arbitrage opportunity yet. However, I think all of us that have touched growth and performance marketing, especially given the caliber of the team OpenAI has hired.
think that there's gonna be a moment here in the next, I don't know, three months, six months where there is gonna be real performance arbitrage on Chat GPT and who can be the first brands to hit that in a in a big way, I think is gonna be sort of generationally interesting. Like if I think about the biggest short term opportunity for, you know, performance oriented people, you know, I can talk to I'm blue in the face about how LLMs mediate your brand and I do think it's fascinating, especially for Fortune five hundreds, but if you're a performance oriented entity,
being one of the first folks that can really find the sort of arbitrage pockets in Chat GPT, I think it's going to be the biggest opportunity. what I'd say though, the last thing I'd say is that like the arbitrage window is probably going to be a lot shorter than it was on Google, Meta or Amazon in the mid twenty tens, right? These systems are just becoming these ad tech systems are just becoming perfectly competitive markets faster and faster.
Tye DeGrange (14:00.587)
Yeah.
Mike (14:17.851)
And if you think about the unstate of what OpenAI or Meta wants, right, they want to create a system where you put a dollar in and you get a dollar and a thousandth of a cent out. Right? Basically the the least possible kind of performance system for marketers to continue spending dollars. It took Amazon and Meta years and years to get there. Chat GPT might get there in a matter of quarters. So it it really is gonna be. I think there'll be this period where a few consumer brands and higher end B2B and services businesses find like
Tye DeGrange (14:31.949)
Just enough to keep you coming back.
Mike (14:47.665)
real alpha and chat GPT ads. but I'm not sure that it'll last as long as you know some of the of the other platform waves.
Tye DeGrange (14:55.693)
Yeah, yeah, it's it's interesting. I differ ways to think about it. It seems like I don't know, I I think they still have their time to to figure out. It's not massively been forever in terms of them announcing. I I do think the anticipation has been quite high. And I think some of us either deep in ad tech and closer to performance marketing are probably like, this has been a little underwhelming yet has it? And
Haven't given them enough time. That's an interesting thought. for those curious about how you're defining GEO, AEO, maybe just break it down for us a little bit about those definitions and make sure we're clear on what we're talking about.
Mike (15:26.267)
Yep. Yeah.
Mike (15:39.324)
Yeah, so I use so it's funny. I've been looking into this to try to kind of like decide what vernacular I'm gonna use. I prefer AEO, answer engine optimization. GEO is the more popular industry term right now per some chat GPT and Google Trends data generative engine optimization. same thing for me, interchangeable. I actually think the better umbrella term though, if you're talking to like an executive about this, is AI search.
and again, this is a little bit of a semantic argument, but GEO and AEO sound like tactical things. they sound like SEO used to sound, and SEO used to get boxed into a small kind of silo and never reach the CMO's desk. I think AI search is is is the better sort of umbrella term to kind of, you know, plan strategy around because it's not really around just optimization for a given answer engine per se. It's really about you know, navigating the sort of
new way that consumers find information online. So long-winded way of saying when I when left to my druthers I define it as AI search if I have to choose between the acronyms AEO is my
Tye DeGrange (16:48.641)
Yeah, that makes sense. I I like the framing up of hey, how do we talk about this intelligently to the executive team? And it also I think also shines the light on partner marketing. Would you agree when it comes to thinking about AI search?
Mike (17:02.469)
Yeah. Yeah, exactly. I actually think the good thing about partner marketing is that it's one of the easier AI search strategies I think for executives to to grok, right? This idea that especially doubling down on publisher and creator partnerships where models have or specifically open AI has a licensing deal in place. I think there's a very nice straight line you can draw between, you know, a piece of coverage you get with a People Inc. publication and
you know, that informing an LLM's results. So I do think partner marketing's in a sort of fortuitous spot that among the different sort of inputs that can influence how an LLM perceives your brand, it's a very clear, I think, straight line to a lot of executives how, you know, pursuing marquee content partnerships will help. I think the other thing that's important that I I stress when we have AEO discussions at Alephic and I've written about in brand is
Like the more that you can anchor your AI search strategy in things that are just good practice anyway, I think the easier it becomes for for for mass buy in, right? Very few people are gonna argue against pursuing top tier content creator partnerships, especially if you can do it on a CPA model. Very few people are gonna argue that it's a bad idea to write richer catalogs that have more semantic kind of attributes about your product.
Very few people are going to argue that it's ever a bad idea to make sure that your brand and products are represented consistently across all of your content. But it's a much bigger deal now with with LLMs who are almost averse to competing information within brand taxonomies. so I think with all of this, it's it it comes down to like even if AI search wasn't a thing, what is just good practice for me to prioritize anyway as a marketer? I think if you're a affiliate manager or an AEO manager trying to get more budget for AI search initiatives.
That's probably a pretty smart place to start.
Tye DeGrange (18:59.095)
Yeah. Yeah, I love that. In in the zero click world, are there units of measurement that you are seeing come up more regularly and you're seeing like, hey, this is where we should be honing and thinking about citation, recommendation, and product mention, the conversion.
Mike (19:12.881)
Yeah, it's really hard because, you know, so much of the industry is focused on citation count, or, you know, various sort of prompt tracking things that I think are kind of pseudoscience. like for a very nascent brand, I suppose raw citation count matters a little bit. but you know, an LLM could just as easily be saying not to buy this brand and that would show up as a citation the same way as a best wire cutter article would.
So you know, I think it's, you know, more sophisticated brands I think are coming up with interesting ways to track sentiment. Profound has re released an interesting feature in the last few weeks called fact check that is a way of helping brands understand if LLMs are actually perceiving their brand correctly and not making factional error factual errors from either bad training data or just like retrieval from weird ass blogs and things like that. you know, I come back to
You know, the the number one metric that I really liked when I was on the gross side selling brands that were were sold through marketplace was search query volume on Amazon. both branded search and like unbranded search that I thought was kind of indicative of our brand kind of standing out. I think you know the more that you can proxy
Tye DeGrange (20:32.971)
High intent query.
Mike (20:36.101)
It's a really hard concept because right, LLMs don't search in the same don't don't hold search query data in the same way Amazon and Google do. But the more you can start to understand the equivalent of tracking branded search on Google and Amazon within LLMs, I think the more you can start to understand how you're actually doing. And that doesn't always correlate to just raw number of citations, especially for for larger enterprise brands.
Tye DeGrange (20:59.745)
Yeah. Yeah. I I I think it's interesting. I I'm hopeful. I mean part of the conversations we've been having around measurement and curious to get your take is I think that the in our industry in partner marketing, I think is intending to, you know, add a lot of levers in terms of control for brands and ability to look at better signals around creator, better signals around the impact in
AI L LM queries and citations and other things and giving brands and partnership, you know, practitioners a better way to like pay out, not just on click, not just looking at flat fee, CPC and conversion, but also looking at things like those nuanced actions and things. I I don't know what your take is on that, but I know that that industry our industry is moving in that direction, and I'm certainly encouraging it.
Mike (21:49.105)
Yeah.
Two thoughts on two two thoughts on this. I think the first, not to just kind of like talk your book or talk up the affiliate leader's book. but everything you just said is why I feel so strongly that someone with formative experience in partner marketing should own AI search strategy more than any of their colleagues who might come up through more conventional search or social backgrounds.
Because at the end of the day, like so much of a brand's success in AI search is based on how the external third party content marketplace is portraying a brand. And nobody is going to know the nuance of that better than somebody who has background affiliate partner creator marketing. Right? Someone who actually understands the difference between a publication that actually reviews products and a publication that writes AI slop that kind of looks like a review. The difference between those things might actually form how an
kind of creates an opinion about a product or brand. And most of your classic performance growth marketers are just not going to have formative experience there. And most of your affiliate partner, creator, influencer marketing folks will. So to me that's the first piece. I think this should be a career defining moment for a lot of people with affiliate roots to to just sort of own more of the marketing stack. And I think, you know, for for especially agencies that bridge the gap between affiliate and growth, I think it's a really exciting
So maybe the first thing. go ahead. Anything should
Tye DeGrange (23:16.961)
Yeah.
I I just I mean, obviously that's compelling and exciting from my perspective, but it also echoes what I've shared a lot, which is around the nature of affiliate is in it there's so much there's so many payment types and flavors and channels housed within it. It's almost behaves like an opportunistic growth engine in itself if it's managed correctly. So I like I appreciate and obviously resonate with where you're going with that.
Mike (23:45.0)
Perfect. second part of what I was kind of getting at is I think, you know, this is gonna be hard for a lot of growth marketing people, but I think there has to be a little bit of an embrace of how little we actually know in terms of being able to directly attribute conversions, right? I think there are some incredible nascent products like Vantage Point, which Partnerize is working on, and and some really interesting things that Cloudflare and Stripe and others are doing. But at the end of the day, like
You know, web analytics, even GA, right? It was always like kinda BSy. Right. I used to always joke, I remember I was in a discussion with a board member at a previous company who was asking me about some of our GA data. and I like we were just kind of talking around each other and I was struggling to kind of convey what I was trying to say until I came up with an analogy that was like, you know, I just kinda came up with this off the cuff and thank God it landed. But it was like you gotta think about GA sort of like
you think about the unemployment rate. Right? Like the unemployment rate is three percent or whatever. Like nobody believes that only three in one hundred Americans are out of work and don't have interest in finding a job. Like everyone kinda gets that's bullshit. But we track the unemployment rate as we do because economists have realized it's the best like microcosmic way to understand directional changes in the economy. It's directional, right? And it's it is in that function. It is it is kind of useful. that's how like web analytics used to be.
Tye DeGrange (24:55.373)
Underemployed.
Tye DeGrange (25:01.847)
Directional.
Mike (25:10.897)
You don't even have your directional data now a lot of the times though. So it it kind of comes back to like, yes, I think some better zero click attribution solutions will be built. yes, maybe there'll be a little bit more science around understanding which sources shape LLMs, but like right now, people that have a good gut read, people that have just like a good killer instinct around which publishers and which creators will actually drive performance.
Are super valuable because I think you're not going to see it obviously in the data for a long time. so it it it's it almost makes the job a little bit more fun again, because it means that people that actually have an eye, I think this is even more true on the creator side than on the publisher side. Like I've worked with some people that are just incredible at having this killer instinct around looking at what look like two identical pieces of content to me.
And being able to say like option B is gonna drive a fifty percent higher conversion rate because this person talks about the product in such a way or this person has the product frame. And I'm just like, what in the world? But I think people like that are gonna be just incredibly valuable and and again, a lot of people like that I think will have roots in in you know, partner or affiliate or or influence or market.
Tye DeGrange (26:11.629)
Yeah.
Tye DeGrange (26:23.809)
Yeah. It it's a fun dovetail because it's it's an area where affiliate creator partner done the right way I I think is is can be a storytelling mechanism, can be a brand builder, can be an awareness driver. And I think the amount of people that are aware of that is far too low. And the amount of people that actually utilize it and do it as such is is also shockingly low. And I think it's a really good call out.
Mike (26:54.117)
Yeah. Yeah, exactly. And again, I think, you know, there's gonna be, I think, a lot of CFOs in large marketing organizations that are gonna start that are quit are gonna question some of these investments, right? That are gonna question whether it's, you know, exorbitant sponsored content deals or creator partnerships that might not be myopically performing like they were three, four years ago.
And but, you know, I think a strong sort of partner strategy is gonna be the backbone of of pretty much any strong AI search strategy. so I think one of the most important things for marketing leaders is gonna be how do you tell that story, you know, sort of internally in a way that is discernible to a non marketing leader who's just used to kind of looking at more sort of direct performance oriented metrics.
Tye DeGrange (27:48.632)
Yeah, yeah, I love that. Love that. I trying to work with some individuals in our industry in particular, it's real I feel it's related. I'm curious to get your perspective, but the notion of MMM becoming more important over the last few years, re revisiting those econom econometric models, actually determining incrementality to where it gets, you know, mildly measurement nerdy. Yes, you can use MMs for
And plug in affiliate data. It's obviously very different than how other channels operate. It's almost inverted. Therefore, you you have to think of it differently. You have to tee up the models differently. And so this comes up a lot in our measurement conversations with leaders of networks, top publishers, and curious to get your take. Cause I in com in conversing with folks around it, I've been trying to encourage look the click and
Persian data is going to be usable in MMMs, get it. But the exposure and impression data is going to make it even more compatible with MMMs. And having that affiliate and MMM compatibility long term, from my view, feels like something we need to go down, especially with the citation work that you're referencing, and especially with the GEO impact.
Mike (29:04.571)
Yeah, I have a I have a different view on this, probably than most people in the industry. I actually think MMMs and MTA is gonna get less useful over the next couple of years and not more broadly. The reason that like my basic thought and this might be incorrect for like very, very sophisticated performance marketers who are running campaigns that touch search social, AI, connected TV, whatever, but I think for the vast majority of brands that you work with that
even I will will work with. Like I think MMM and MTA is going to tell more of a faux precision story than a a real one in the next kind of three to five years. unfortunately I don't think that's going to stop more brands from trying to adopt more complex MMM models and probably poorly allocate some media spend as a result. But I would almost be until there's a little bit more signal and just more
Tye DeGrange (29:55.169)
Ha ha ha.
Mike (30:03.349)
effectively better at a better web analytics for zero click interfaces and AI search. I'd be thinking about attribution as back to like largely last click with incrementality holdout testing. Totally. Like truly. and just understand the limitations of that if you do it. But
Tye DeGrange (30:20.161)
I I I yeah, totally. And I and I not to stop you there 'cause I want to keep hearing that, but I think you're you're right in a lot of cases. I mean the amount of brands if we take a step back and look at them and say, you should actually be
ingesting an MMM, building an MM, working with a third party MM is you're probably talking about a not a massive a set of brands. I mean it is a I would say they probably have to be across at least four pretty sizable channels that are all pretty, you know, you know, different. I think there's a a minimum spend amount that's quite high. so I I would I I some of what you're saying I agree with.
Mike (31:06.245)
Yeah, there are pockets where I think there's it like it rather than sort of general purpose MMMs, I think a tool like Prussian AI is pretty interesting if you're spending a lot of money on Facebook for D to C and you want to understand how that's impacting your Amazon or Walmart sales. but to me that's like it's different than a lot of MMM thinking because that's a very scope that's a scoped problem, right? It's a scopes and that this is why I like you know, incrementality holdout testing as well. It's like a very scoped problem.
Tye DeGrange (31:16.375)
Yeah.
Tye DeGrange (31:26.764)
yeah.
Mike (31:35.282)
that you can kind of get yes or no answers to rather than a lot of MMM inputs which are gonna give you, you know, sort of abstract numbers that suggest influence of different channels in a lot of ways that I think is is is really difficult. again, I do think this will get better, right? I I there it's not gonna be entirely opaque and indecipherable forever how LLM or how different sources impact kind of hidden LLM.
Tye DeGrange (31:52.077)
Totally get it.
Mike (32:04.827)
traffic like that will become more transparent over time. But we're in this moment right now where there's a good chance that a lot of your especially your partner marketing investments, but a lot of your other advertising investments are show are impacting how you kind of show up in LOMs and you're just not gonna be able to measure that. Like really all that well. so you know what can you control? You can do holdout tests for those channels. You can
Tye DeGrange (32:23.745)
Yeah. Yeah.
Mike (32:30.459)
Just try to take some educated guesses based on again which publishers and creators seem to broadly be cited by the models. but like the idea that a even the most sophisticated kind of twenty twenty four Mm model is gonna really account for AI search correctly, I think is a a little bit of a a little bit of a fallacy.
Tye DeGrange (32:50.731)
Yeah, you alluded to it earlier. I think we you we have to be honest and comfortable about where we can't measure and what we don't know. And I think that's an important part of the conversation that everyone needs to be candid about. That that back to the GEO kind of realm a little bit. And obviously, theme of the pod, the the levels of testing and experimentation.
Has there been something where that you can speak to? And if you can't, that's totally fine, that you've seen as an as a test of some sort within that was impacting GEO, AEO, something that was interesting.
Mike (33:30.629)
Yeah, I think this is where yeah, so this is an interesting one, right? Because when I was on the brand side in twenty twenty four, we did some sort of growth hacky test that would be aggressively counterproductive to try in twenty twenty six. involving like AI generated programmatic SEO content where we w wrote like one really good brand owned blog post and then had a combination of
2024 AI and a team in India just like spam a hundred versions of that. But in 2024, that kind of stuff was getting picked up by Google AI overviews and driving like thousands of visitors in some cases. I think what the really important thing about, I think like more performance-oriented GEO things is just to kind of like be honest about the testing window. Like I think the first step, if you if you're thinking about doing any of these things, is like, am I pursuing a long-term
AI search strategy here, or am I trying to find an arbitrage window? I think too many times those things get conflated. for example, right, if if you were gonna green light a Reddit astro turfing campaign last year, if you were a performance-oriented direct response brand, I would say, yeah, probably do it. But probably understand you're trying to capture traffic for six to twelve months before this kind of stuff gets cleaned up. so I think more than other channels like
I think have to be honest about like what is the arbitrage window I'm going after if you're trying something a little bit more sort of hacky. you know, I at this point from where I sit writing a newsletter and working for a company that basically has Fortune 500 clients, I'm less interested in in obviously this sort of growth hacky test. There are three things that I think are gonna be very enduring on the AI search side. The first, which we feel like we've kind of beat to death, is you know, high quality affiliate.
partnerships, right? I think it's very it is not likely that any of the models are going to steer away from citing consumer reports or wire cutter in the near future. now this I mean wire cutter is a little bit of complexity given the some of the lawsuits with the times, but the the broader point still stands, right? Like high quality expert reviews. The second I think is going to be sort of creator content.
Mike (35:42.684)
I just only see creator content being more impactful in impacting the models, right? Google is gonna lean on YouTube more and more because they own it to to kind of serve content. I think OpenAI is gonna have to do a creator platform deal in the next 12 months to try to bring more UGC and creator advertising into their platform, whether that's with Instagram or TikTok or LTK or ShopMai or all the above. I don't know, but it feels pretty likely to me that
especially given, you know, the fact that ChatGPT wants to be more of a consumer product, that there's gonna have to be more creator content brought in both to power organic results and ads. and then the third for me is is high quality brand owned content, right? the models are gonna continue to especially for more technically oriented products, the models are gonna continue to draw on and favor brands that just do a really good job exposing their canonical product information, and even doing things like
if you're at the furniture brand putting assembly guides online and all these other things that are valuable to users. So you know, I think there's a type of marketer that should continue doing sort of some of the hacky tests and this, that, and the other thing. But if you kind of came to me like you have ten minutes with a CMO and they you need to help them kind of just prioritize zero sum what to focus on for AI search, it would be green lighting, more budget and team to pursue high quality affiliate partnerships.
probably spending spending even more on creator than than some brands are today. and then really, really increasing the investment in brand owned content. these aren't like these aren't super sexy silver bullet kind of things, but these are the things that I think as the models start to favor, you know, sort of different arbitrage hack de jour, these kind of three areas will stay pretty kind of evergreen in terms of shaping how you know, especially Google and ChatGPT view the world.
Tye DeGrange (37:38.03)
Like amazing. I I think that's a phenomenal, like, you know, punchline that you just landed on of like, hey, this is what, you know, the the core basic healthy meal is gonna make you much more healthy in your training program. It didn't change drastically, but there it's a prioritization, it's a smart way of thinking about it. It's obviously exciting for you and I because it it it kind of jumps into our worlds.
And I I'm impressed to to hear and I always enjoy our conversations around it.
Mike (38:09.307)
Yeah, and the other thing too is like if yeah one thing I've been thinking a lot about is like if you're I don't know, Procter and Gantt, like some massive CPG brand, do you wanna even be pursuing the latest AI surge hack and going up against some two hundred million dollar a year supplements brand that is always gonna be better at it than you? Or do you wanna just stick to the things that you know will compound over time?
which again I think is is right now is is brand own content creator and and and high quality affiliate partnerships. So I mean again, there's always gonna be a certain type of marketer or entity that is gonna have to look for pockets of growth alpha. and even large brands should do that in other parts of the business. Do it on Meta, do it with your Amazon ads, etc. But I think for AI search and AEO, you know, it's really gonna be about kind of the like it like you said, the healthy meal.
and you know, I would focus more of the sort of arbitrage growth hacky calories around you know, trying to to find alpha and paid channels.
Tye DeGrange (39:11.041)
Yeah, I love it. just coming down the home stretch, Mike, it's been amazing to talk to you. Is there a resource that you suggest? Is there like a it could be a book that's about the topic, it could be a non f something fun that you just highly recommend for people to read? Is there a book that you think must have that you suggest? Since we have all these leather bound books behind you, I figured, you know, we we could finish strong. Bring it all back together. Just pull one off the shelf and
Mike (39:32.252)
Yeah, what do I have back here? this is a little bit out of left field. I'm just making sure I get the title of this book right. give me one second. I just finished just I can't remember the freaking title of it. This is embarrassing. This is how I know I'm getting old. hold on. What? This is like an embarrassing blank to be drawing live. Destiny of the Republic!
Tye DeGrange (39:44.455)
Left field is a good is a good part of the pod.
Tye DeGrange (39:53.901)
No, it's okay.
Mike (40:02.705)
I could not remember that for the life of me. so Destiny of the Republic is a biography about James Garfield, who was assassinated like sixty days into his presidential term and kind of like forgotten. Kind of forgotten by history, but was actually one of the most interesting political thinkers of sort of the post Lincoln era in the United States. and I think the like it's no hyperbole to kind of
Tye DeGrange (40:03.084)
Ooh.
Tye DeGrange (40:10.888)
dude.
Right up my alley.
Mike (40:29.617)
suggests that the trajectory of the US would have been radically different had Garfield not been assassinated for the better. most likely. He was just a a really kind of phenomenal thinker, like I said, in the post Lincoln model. So it's both like a historical thriller and just like an interesting kind of book for anybody that's into politics or post Civil War kind of America. so that would that would be one
Tye DeGrange (40:51.627)
This is amazing.
Mike (40:54.065)
I do think everybody listening to this, even if you're not like deeply AI pilled or following closely, should read Dwarkesh's latest substack, which published over the weekend, called The Rise and Fall of AI Civilizations or Agent Civilizations. it's about the open AI hugging face incident. it's been heavily criticized for basically anthropomorphizing AI too much and giving it lifelike character.
characteristics, but I think it's the best kind of explainer for an event that was covered by the press but like not really appreciated. and just sort of understanding how you know how open AI or how AI might sort of transpire in the security realm. So obviously not a book, but I think certainly a piece that it's worth
Tye DeGrange (41:40.365)
I saw his thread come out. What was the name of the title of the article on Substack?
Mike (41:44.294)
It's called The Rise and Fall of Agent Civilizations. It was published over the weekend, August 29th. so yeah, that would be a another sort of kind of recommendation, I think, for more kind of current events. on more of the core marketing side, I'm gonna pump one very tactical piece. Tom Critchlow, who is our AI search specialist at Alephic and myself, wrote a piece called the AI Search Maturity Matrix.
Tye DeGrange (41:58.093)
I love
Mike (42:12.579)
in the BRXND newsletter that is a framework for marketers to walk through with executives on sort of the five levels of maturity in AI search. it's a little bit of a pithier summary of a lot of what we've spent forty five minutes talking about. So if you would if you could read those three pieces, you would get a a case study in American history, an understanding of where AI is taking us, and then something to present to your boss to hopefully get additional investment or partner marketing.
Tye DeGrange (42:38.861)
Little something for everybody. A little something for everybody. Yeah, that the presidential history stuff is a hoot. I absolutely love when I can get a chance with my spare time to do that. But the other two I saw those pop up this weekend. I saw the high level summation, but I did not get into the details. So I definitely want to check that out. And Mike, e exciting to have you. Appreciate you. Where else you I wanted
give you a chance to share details on the event. We'll make sure to post those when we post the pod, but hit us with the event and where else to go to check you guys out.
Mike (43:16.347)
Yeah, so event which is live at BRXMD dot AI is on November fifth, the New York Times Center. We have incredible speakers, Chief Creative Officer of Amazon, SVP and Microsoft, Chief AI officers of Rocket Money and Bark, as well as a number of incredible sponsors that we're gonna announce next week. the event is open to sort of any enterprise marketer or agency or really anybody that listens to your pod will get approved by me.
but tickets are seven hundred fifty bucks for another couple weeks. so still on a nice discount. and yeah, I think if you're even mildly curious about what the kind of world will look like as as agents mediate more of brand preference, it's gonna be sort of an incredible show and hope some of your team and some of your clients can make it.
Tye DeGrange (44:04.449)
you. We'll be definitely getting folks sharing that out and sharing with pod listeners, clients, teammates, hoping to be there and always appreciate talking to you, Mike. Thanks for coming on. This has been an amazing episode and always is a good conversation with you. Appreciate it.
Mike (44:21.231)
All right, man, this was a ton of fun. sorry, I think in my somewhat delirious, somewhat sick state I was a little too verbose on some answers, but hopefully the audience bears with us and I promise I'll be a little pithier next time.
Tye DeGrange (44:34.646)
You were pithy as can be and maybe not your your perfection view of it, but you did an amazing job and I'm I'm grateful. It was a lot of fun and we should do it more often. Thank you.
Mike (44:39.408)
Yeah.
Mike (44:44.882)
Hell yeah, man. Thank you so much.