Why clean data matters more than the AI model

Welcome back to another episode of Limited Supply. Today, I've got a very special guest, my friend Maxx Blank, who we've been friends for a very long time now, it feels. We actually became friends. Actually, before we talk about that, why don't you just give a quick blurb of yourself? Yeah. Maxx Blank, co-founder of Triple Whale. Born and bred in Columbus, Ohio, Midwest. You know, we'd come to New York all the time. We have an office here. And spend a lot of time in the trenches with DTC and what's going on more broadly across the whole space.

And, you know, I don't know how long we've been friends for. I'm trying to think. Well, so it was MentorPass is what brought us together. Do you remember that? You were showing me Madison Braids. Yeah. Which is a business that you've run for a long time. Yeah. I mean, that's kind of what the emphasis of Triple Whale grew out of that. Right. Right. And probably the third or fourth call, it would be me and then you and AJ. And the third or fourth call, you started showing me these widgets in your browser that were basically your own reporting widgets.

I think on Shopify's UI and Meta's ad UI. And that was kind of the beginning of Triple Whale. Yeah. I was like, wow, I totally forgot. MentorPass. Yeah. So, I mean, before Triple Whale, you know, I was running Madison Braids, a few other brands too. But Madison Braids is really, you know, what led to Triple Whale. And AJ was working with me on that. And I remember just trying to meet whoever I could. And you were like one of the guys, you know, that was good to meet.

Yeah. I booked you. I figured what we were doing. But I asked you certain questions. We were trying to think about just general. Yeah. Influencer, increasing conversion rate. Yeah. And then somewhere along the lines, you know, with my background and just entrepreneurship and running brands and AJ, who we go way back in entrepreneurship in general. And just, you know, when you're an entrepreneur, right, you need to meet other entrepreneurs. Totally. You know, you're battery charged. Right. He's one of those guys.

And at some point we started working together on Madison. Madison and that, you know, the need to have like a tool that I could just like see all my numbers in real time. That was really the birth of Triple Whale. Yeah. And then it kind of just went, kind of went viral with Jason Wong tweeting it and like Nick Shackelford and all these OGs. Yeah. And you definitely were helpful in that as well. Totally. I think it's always interesting, too, to like go back and think about what was it at the time that you were doing with Madison Braids that like got you to even think, OK, first of all, I should rethink about, you know, software.

And I think, too, that was do you remember what year that was? That must have been. Yeah. That was like probably 2019, 2020. Yeah. So I'm around COVID or before or after in that area. Yeah. Yeah. And I remember at that time there was like there was actually no real proper analytics tools. There was maybe one. I forget the name of it. Yeah. Yeah. But the problem with a lot of these analytics tools where they were there were actually more business intelligence tools versus like analytics for e-com brands.

Right. And that's a specific breed of person. Right. You're not like super data heavy. You're not super tech heavy. Yeah. You're like, OK, I know media buying. I know I know Facebook ads. Maybe I learned some from like a dropshipping YouTube video and a specific type of entrepreneur. Yeah. Some Jon Loomer videos. Yeah. Or Ezra Firestone. You know, whoever it was. And like I did not have like a data background at all. And Triple Whale was like what it is today was not what it was.

No. Yeah. You know, it's been kind of like boom. Yeah. But it was very much for like the everyday sort of Shopify entrepreneur that just want to see their easily see their data. Right. [unclear phrase] Yeah. What I was trying to do. Can I get my blended ROAS and know like, OK, was it Klaviyo that did this or was it like my ads? Right. Was it all the creators that launched today or was it something else? And that's what I was trying to get.

Yeah. That was the beginning. Yeah. I remember in 2018 when I was at Hint, we still never had proper reporting. You know, we had like a Tableau dashboard that would get pulled by an analyst. And then a screenshot of that is what would go to the exec team every day in an email. Right. Basically saying, here's how we're pacing. How many brands still do that? Yeah. Yeah. Like I talked to brands that are today in the hundreds of millions of dollars still have a GA screenshots and have not got that.

We from my perspective, I'm like, oh, we've gotten everybody. It's like, no, we really haven't. And the plate, the space is so huge. This is still common. Yeah. Yeah. And what's funny, too, is there were there were a few players that came in and tried to do the analytics for Econ, but they all kind of fizzled out because they also, I think, took the approach of this enterprise enterprise contract. 12 month lock in approach and a bunch of people just kind of like stopped working with them after a while.

And and then Triple Whale came on the market and everybody was jumping to Triple Whale. So, yeah, it was just easy to get started. Yeah. And we started off. You can just come in, you can leave whenever you wanted. Some point that software has to mature and you have to get contracts, you know? Yeah. But I'll say going in from like the perspective of like what I was doing and feeling it and knowing it and knowing it was out there. Yeah.

And just going quickly and making it so easy. That's what kind of got us off to the races, not having strict contract terms and things like that. But as you mature, you can't scale a business if you don't know what's coming next month. Totally. So you have to kind of mature that way. Yeah. But we still have a [unclear word]. You can come in, get started for free and get the majority of what you need for a startup. Yeah. And at one point, too, we launched that Founders Dash together, which was like the free login, you know.

It's like thousands of downloads a month. Yeah. I mean, it's such a good idea because that was literally I think we were talking about like let's just make an easy way for people to see their numbers. Because even thinking back to those hint days, it was just never easy, you know. We have huge brands just on the free version. That's amazing. Guys, there's more here. But, you know, and it's fascinating. But, yeah, it's all about that. Yeah. You know, first of all, it's real time.

Right. Basically, sometimes you will see the numbers in [unclear name] before you'll see them in Facebook. Yeah, which is pretty amazing. Right. Yeah. The way we store data. We have a crazy infrastructure that we built in-house. Right. I think before we get into the product more, I'm curious, like when you started building Triple Whale, what was the goal or like what were you trying to solve for? And, you know, because that was probably much different than today where the product has evolved to.

For sure. I mean, the goal at the beginning was just how do I get my numbers in one place? Because I was just the, you know, just AJ and I. Essentially, like, AJ started media buying. I showed him how to do it. Yeah. And I was more like on the inventory and ops side. There's not enough attention that goes into ops. Right. And, you know, logistics and things like that. So I started focusing on there. But the idea was like, how can I just get all of my numbers centralized?

And I wasn't dealing with that problem. Yeah. It became very, you know, evident. So like at the beginning, the goal was not, I did not think it would get to where it's at right now. Yeah. It was just like simply get it all together. Right. You know. And was it initially, let me just make this for us? Yes. Get something that works. I always wanted to go, you know, wanted to have a software solution. But I, I'm always like, you know, how much can I focus on two things at once?

Yeah. It's hard. But now we focus on like a million things at once and I realized I could do it. Yeah. But it wasn't like until AJ and I started really working hard on it together that he was able to go super hard on. He's obsessed with product, which is awesome. And he was able to really just go deep on that. And then I was on the business side. And we were able to just kind of do our thing and had this magic chemistry that just, we were able to make it work.

Yeah. Do you think there was any. And then we just dropped Madison very quickly. I mean, we still use it internally. Yeah. Yeah. But yeah. It's always in some demo. A hundred percent. Yeah. Was there anything about running Madison Braids that, you know, was really hard that you felt like turned into a good lesson as you started building Triple Whale? I mean, for sure. Just knowing the problem so well. Like I still stay close to Madison because I want to stay connected.

Yeah. I mean, obviously I meet with brands all the time, but the brands that I meet with now are like our biggest clients. It's nowhere near the level of where I was with Madison. But I think that, that experience I had building something for myself and coming from the industry like that, that's invaluable. Yeah. Yeah. And also the scrappy sort of entrepreneur that kind of is bred. Right. Shopify space. Always keeping close to that. Yeah. Yeah. The best e-commerce entrepreneurs, they know every number from every day for the last, you know,

six months. Yeah. Yeah. Yeah. That's amazing. When did you realize like as you were building this that other, there were other founders that wanted a solution like this as well? I think immediately we started, once it was like ready, we started sharing it with others. Yeah. Starting with like the agencies that were buying media. Yeah. That were Madison Braids. And we just saw that it was moving. We started giving to people. Right. And then just DMing people. Yeah. Try this, try this, try this.

Yeah. Kind of like the influencer, like recruitment style, like getting people to try your product and then get pictures in return. Right. We kind of just sprayed and prayed. Yeah. And, you know, maybe like one out of every 150 person responded. Yeah. One of them being Jason Wong. I'll never forget that. But yeah. Yeah. And I think that was also a great way you guys got a bunch of like early backers and investors advisors on board too. Yeah. I think we kind of etched that playbook.

Yeah. Getting really everyone involved. Yeah. Yeah. A hundred percent. That's how we got Rabah early on. Yeah. Yeah. Yeah. He's hugely instrumental in the Triple Whale. This is like go to market. This is now all nostalgia for me. Yeah. Totally. Rabah, Alexa, the whole crew. It's a great crew. That small office across Whole Foods in Austin. Oh, yeah. You still have that? No. No. Yeah. That was a good office. Good old days. Yeah. Yeah. I'm curious.

Like, so I want to dive into Triple Whale. So before we start talking about Moby, like walk me through kind of the product evolution. Yeah. You know, it started with the widget. Today, it's a enterprise level all in one solution. But like, how did it get from the widget to where it is today? Yeah. You know, we are just obsessed with the product. You know, sometimes we put too much stuff out there. You know, sometimes we take stuff back, you know.

But we're always testing and iterating. And nowadays, to test and to iterate has never been easier and more like affordable with agentic coding. Yeah. We can test things quickly, bring things back, you know, iterate super fast. It's like, you know, I'm sure you're feeling it yourself, you know. Totally. But we've always been obsessed. And as the brand, you know, in our space, like you can go from like 250K a month to like millions of dollars really quickly. Yeah. And we just held on to the relationships of those brands that were scaling.

And we scaled with them. And they needed more reporting. We bought custom reporting. They needed more connectors. We brought more integrations. And they needed like a data infrastructure, which these are performance marketers. They don't really know much about like data infrastructure until it become big. 50 million, 100 million, you start to build out like a data warehouse and things like that. So we just latched on to a lot of the darlings that grew up with us. You know, the darling brands.

Yeah. And we built them like ruthlessly, just gave them stuff for free just so we could have the knowledge and just the feedback loops. And we just kept up with that. We didn't stop at like, okay, this is good enough. We just kept building and scaling with those real operators and learning, you know? Yeah. That's really like how we got to a lot of the bigger brands we work with now was just doing that. Some brands started with us when they were a couple million bucks.

And now they're doing hundreds of millions of dollars of revenue. Yeah. And you see them, that sort of evolution of, okay, the founders and like a couple of friends, they get product market fit and they start to scale. They start to go into a few different channels. Maybe they're bringing agency on at like 10, 15. And then they start to bring on Amazon, maybe some other marketplaces. And they get to like 50, 60, and they start to dabble into retail. And maybe at some point they open up a physical store.

And we've just seen that iteration where they get into Costco or they get into some other retailers. Right. And you see that progression. As opposed to maybe like 10, 15, 20 years ago, you'd maybe start opposite. Right. We maybe start with a store, old school, like back in the day. Yep. Maybe you'd start with, you know, retail. But now it's clearly you're starting online and you're progressively scaling. And we see all that all the time. You know, and it's Shopify has made it so easy to do these things.

Yeah. A hundred percent. And then we come on top and organize all the data and the infrastructure for these brands that really can unleash growth for them by having these things organized.

Specifically in AI, right? Yeah. The data in is going to be the quality of the answers out, the actions, the recommendations out. And I feel like that was something you guys saw really early. I think our first AI summit that we hosted was two years ago. Yaakov came. Yaakov came and spoke. And I'll be honest, I think when he spoke, I was confused about what is Triple Whale doing. For sure. Right? Because out of nowhere, chat was basically just chat at that point.

But you guys had already gotten access to the Gemini APIs and you started making creative and people were like, well, how is this even happening? And there was limited to eight second clips and you guys were able to put that together. And then also on the data side, you guys were so early on data organization for the purpose of AI, which again, I think conceptually was understood that you need good data going in. But at that time, no one had any idea what AI was capable of or how it was used or, you know, the way people use, whether it's Moby or Codex or Claude Code today, completely like not even a thought.

Different. Back then of any of this possibility. But had you guys not thought about data organization and building a warehouse and again, supplementing what a performance marketer doesn't know, you wouldn't be able to use these tools like that today. And you see it actually when people use something like Codex or Claude Code, but they don't have a Triple Whale built in to handle the data organization. The results are actually not, the numbers aren't clean. They're quite messy. Right. It doesn't know your actual business.

Yeah. Yeah. This is something that on our pursuit to scale with these smaller brands that then, you know, scaled, they needed access to like custom reporting. Every brand or agency or media buyer has like their spreadsheet. Right. Every day at the pacing. Right. So as those brands get bigger, that spreadsheet gets bigger and more complicated. Yeah. So we started building the data infrastructure that would actually replace that. Lo and behold, like that is the same type of infrastructure that you need to have clean data that you can run AI on top of.

Right. And then on the warehouse, like we built our own, it's a forked version of ClickHouse, which is a specific type of data warehouse for real time analytics. And what that allowed each account is to basically have their own stored data over time. As soon as you connect any integration, we save all that data from when you opened up your ad account. Every on, off, budget, scale, anything saved. Mm-hmm. So you can now run Moby on that data and say, like, give me all the actions I took.

Give me some type of like insights from like what happened here. Yeah. Like that is massively powerful. Yeah. Something like that is, is Triple Whale as a whole also getting smarter from these inputs coming in from brands or is it just specific to that one brand? Specific to that one brand. However, we do have a data set that's our benchmark data set that is anonymized trends, peer to peer, like cohorts. Yeah. That can, it's like, you know, people go on X and say like, what's going on with CPMs today?

What's going on with, you know, this or that today? And everyone's kind of trying to figure it out. So there's a benchmark inside of [unclear name] that can kind of tell you, is it, is it, is it us or is it, is it, is it you or is it the market? Yeah. And so Moby will get access to that and you can, you know, query it and say what's going on. Is it me or is it, is it them? Yeah. You know?

And so that, that, that is in there. And we'll get more and more learnings from that that we can share. But, you know, privacy and, and data privacy and things like that, we take very seriously. So it has to be like anonymized, has to be a certain amount in that as well. Yeah. So that it's like, you can never figure out who is who. Right. But there is an element of like it getting smarter over time for sure. Yeah. I think the data and privacy is so interesting.

You know, Instinct, for example, came out recently and there was all this talk about, well, these guys are basically saving a copy of everything that you do in their servers and it's not deleted. And, you know, people on Twitter have shared screenshots of, oh, I got a response from somebody else's Instinct account in my, you know, in my messages. Those are like, those are just like hallucinations. For sure. Yeah. Or some of them could be, you know, a competitor or something.

But nonetheless, I think it still raises the concern for a lot of people about like, you know, what's private or what stays private to me?

How does this data get used? For sure. Yeah. I'm curious, you know, if we talk, start talking about Moby now, like what's, what are some, some things that today, you know, brands have almost off-boarded to Moby entirely? Like, even just thinking about how, how teams have evolved over the last like three years in e-commerce, right? You're now seeing a lot leaner teams. Even I think Steven from Cuts was on a video a couple days ago just talking about how his new team setup is so lean.

Like, it's just a few people now. It doesn't require, you know, 30 plus people to run an e-commerce business. And a lot of it is because, you know, players like Moby are able to take a lot of these jobs and do them. So what are some of the things that we could talk about? So, yeah, I think, I think it's an incredible conversation. I want to preface it also by saying that like people's adoption curve with AI is different. Yeah. It's different.

And it's, there's a lot of anxiety, I think, with AI, like, oh, keeping up with like what is possible. Yeah. The only way to keep up is you have to be unemployed. You've got to jump in and try things also. Totally. I would say like the rate of adoption, like if you had a chart and I could say functionality is like straight up into the right, you know? I'd say adoption is like so low still. Yeah. People are trying to figure out there's a lot of testing.

There's a lot of, is this working? Is that not working? It's confusing. It's hard. You know, one of the things that we set out to do, which is kind of what we did from the beginning, which is organizing and giving you a system to operate in with your data, with your data clean and ready to go for any type of reporting. Right. That was like the beginning. Right. But now with Moby, which underneath Moby is honestly, it's, it's, it's an agentic harness.

Right. So Moby has all of the frontier models. Yeah. All of the creative models, anything that's new, we get it almost a day before the day of. Right. We're very close to all the frontier models and all the labs. And, um, with inside the harness, we have the models, then we have tools, software, data, your data, benchmark data, the ability to really, and any type of experience you have with a Claude or a Codex you can do in Moby, but on top of your data and it's governed.

Right. You have control. Right. There's automations. There's, um, instead of having like all these MCPs you need to manage, we have them that we're managing for you. Yeah. Essentially. And can you also turn ads on and off? You can schedule budgets increasing and decreasing. Right. Right. You can really start to offload a lot of the work. So like to go back to your original question. So first of all, I don't think anyone's really figured out. We are seeing the brands at the top end that are super AI pilled and the common themes of the ones that are really making changes are that the founders or the CEO are so obsessed with AI and see the future that they have the anxiety and they're thinking and they're pondering like, what does this mean for humanity?

You know what I mean? Yeah. They're the ones that are really leaning in and just curious. And I think too, like AI is so great for people who are tinkerers, people who want to mess around with something or just get something. Build something. Build something. Yeah. That'll like show me the money. Right. AI. You know? And there's ways to do it. And whether it's like leaning or scaling back teams potentially or it's like what are the things that our team is doing all the time that are repetitive that we can now really automate and trust it?

Mm-hmm. And then what are the opportunities now that we can finally go and do? Right. Like, I can remember being on the merchant side and always wanting to try a new channel or a new, either it was a new ad channel or a new place to sell. But, like, you know, in between trying to figure out where my money's going right now, trying to mess with inventory and plan cash properly. Yep. Manage ads, so on and so forth, right? You never get the time to do those things.

Right. I think there's, like, an infinite opportunity to test new things. Mm-hmm. And you see that. Like, you know, whether you're an AI doomer or, you know, which I'm not, I think it's going to be, like, unprecedented opportunity for everybody.

Right? I think there's, like, that [name unclear] principle. You've heard of that? Yeah. Well, like, basically, it just creates infinite opportunity because there's always another landing page to try. Totally. You know? I believe in that. Like, you're seeing now with engineering, right, people are saying, or even internally, when we pushed AI coding, things like that, there's anxiety, you know, and I don't blame them. But, like, now most of our engineering is fully agentic. Yeah. But you have... You still have the leaders who know how to...

You still have the engineers. Yeah. You still need them. Yeah, exactly. I think that's the perception people have is, oh, it's going to replace this person entirely versus it's actually going to make this person 10 times more productive. Because the guy that knows what needs to be done can now orchestrate those things and still review and make sure they're good. Yeah, because at the end of the day, you can vibe code whatever. But, like, how many things have you vibe coded into production that have to, like, perform?

Right. You know, and you don't really do that unless you have someone who knows it down to the kishkas. Yeah. Totally. Totally. Totally. And you end up needing engineers that know how to orchestrate these things. Yeah. Can you explain... Just one quick thing that was interesting is that... Yeah, yeah. ...so all this scarcity mindset of, like, okay, there's going to be engineers to be gone, one specific example, right? And now you have engineers just in fields you never would have thought they would be in.

Yeah. Because now it's like... Like, what's it mean for plumbers? Yeah. What's it mean for, like, accountants? Totally. So I think you actually create more demand. Yeah, I mean... Someone had told me that lawyers... Yeah. The amount of lawyers you need... Because, like, you're going to go and you may get a good idea of what's going on with something using [product name unclear] or whatever. But, like, no, you want a lawyer to review that. Right. Now there's a lot more things for lawyers to review.

Right. And at the end of the day, it's like human trust relationships. And, like, you need that. Yeah. 100%. Yeah. It only really multiplies the output that's possible. AI is an amplifier, a multiplier. Yeah, totally. Yeah. And I think another just side anecdote that's been interesting, too, is, like, offshore teams now using AI, but because they're using something like a Moby, they're actually just as up-to-date in real time as the people in the New York City or LA office. Sure. You know?

Yeah. They're not waiting for data. They're not waiting to get something to them. You know, they're not waiting on somebody else. They're able to actually go and execute. Yeah. And it's been really interesting just even watching brands take these teams and help. Like, they've now turned somebody who might have just been kind of a data entry or project manager, project coordinator type role. Superpower. And now they're building landing pages. They're looking at what creative is doing best and coming out with 50 new versions of

it on their own, you know? Without having to require six or seven other people in the process. And I think that as it relates to, like, what do org charts look like in e-commerce companies? What are the actual, like, show me the money AI? Yeah. It's these types of things. Totally. And I also think that people don't necessarily know, but you now have the time and the space. Yeah. You start to experiment and figure out what does the future org chart look like?

What does the future... What is the true capacity? You know, I like to say internally that, like, AI is giving us internally, but also our clients, more importantly, like, what's it called? Iron Man suit. Yeah. You know, it's like bionic. You know, you can do so much. Right. I mean, Moby has, like I mentioned, has the creative models in it. Right. So now we have the pixel, right? We've always had our pixel. Right. Which is telling you a source of truth on customer journeys.

Right. Which ads are producing the best customer journeys, the most profitable customer journeys. Let's take those ads and then let's put it into a Nano Banana or a Seedance 2.5, you know, and then let's churn out more creatives like that. Let's attach it to landing pages. Let's tell Moby to do that every Monday, 25 different variants and publish them for us automatically. Right. That's like a growth engine. That's a dream. Yeah. It's a whole flywheel. You can do that. Under one kind of process.

Yeah.

Yeah. Yeah. Can you explain how a harness is different than what most people probably use, which might be ChatGPT.com or the basic Claude app or Claude Cowork? I would say like a harness is really just like, well, let's back it up. So an agent or like a Claude, a Claude Code or a Claude or a ChatGPT, it's really like an engine, right? It's like an engine. It's powerful. But for an engine or a car to move, you need all the other things, right?

You need the steering wheel. You need the mirrors. You need everything that makes a car move. It's not just the engine. It's a very powerful part of it. So with a harness, you have the engine, but you have all the other tools as well. It'll make it powerful, right? And then the relationship between the models, between each other that we've been able to orchestrate. So you go into Moby and you get some type of insight from Claude, pass that insight off

to Seedance 2.5 to make the creative better based on whatever inference you just derived from the frontier model, right? So an engine is really just like the model itself. The agent, or sorry, the harness is really all of it together. Does that make sense? Right. All of the things together. You can plug in the context, the memory, data set. Memory. Additional models. Yes. And what our team has done, our AI team has made it so they can work together really well.

So when you put all that together, you get this recipe for just really driving business results. Yeah. And I think too, the way that Moby is built out, it's very much built for people to use AI at a really high level, but also the early adopters and people who are looking to start implement or integrating it into their workflows too. You know, you don't have to make a lot of complex decisions to get good output from Moby. You don't. No. And it's gotten smarter and smarter.

And a good thing about being a harness is that we ride the wave of all the models. You know, you go into Moby and you can select which model. Right. So the new ones, they're always there. Now, they're more expensive, right? Yeah. We all know that. But you can kind of ride, we kind of ride the wave.

Do you find with, you know, even like, let's say somebody is like, I only want to use Astra, right? Pay for the most expensive one. People are still finding a lot of cost efficiencies and savings by using these models to execute the work that they're trying to get. Oh, yeah. Yeah. And it's work that we do also, like on routing. You know, what questions should go to which model. Yeah. Can you explain that routing a little bit? That's like, can Moby predict or tell you basically on what you asked it, which model

should it take? And that's something we're getting better and better at. You know, as the depth of usage of Moby gets more and more, you know, which model should take what? And that's how you start to optimize the costs and bring down those costs. Right. And I would say also, like, what is Moby's relationship with Claude, ChatGPT outside of Moby? Right. So we have an MCP. You can take Moby out. Can you explain the MCP? Yeah. So right now you can take the analytics that Triple has and put it into Claude or to ChatGPT.

I think next week we're going to be launching an app inside ChatGPT. So you can come in and have Triple Whale. You can just tag Triple Whale. Yeah. Or you can have your dashboard. You can build a dashboard in there really quickly. Yeah. You know, so like we want to be everywhere. You can't just really pick. It's too, it's too like every week is something different in this new world we live in. Right. So we've taken the approach of like be where people want us to be.

But there's a real advantage of doing it in Triple Whale. So the data's the freshest, the fastest. Yeah. You know, and we have a lot of the things that you need if you're trying to set up a whole ecosystem for your business. You know, you want the data warehouse? We've got that. You want the integrations? We've got that. Almost every single integration we've got. There are some edge cases. We're working on it all the time though. We want all the models, right?

We've got memory, context. We've got all that. You know, you want a computer that can check browsers for you? I'm like run. Moby's got that. Right. Yeah. It's pretty amazing. Like a lunchable. Yeah. It's overwhelming, to be honest.

And we're figuring out how to like de-overwhelm it and get really practical and useful. I think with a lot of this stuff, there's a learning curve you kind of have to go through. Like, you know, for me personally, end of last year, I would say I remember thinking like, wow, how are people tweeting that, you know, vibe coded apps are going to replace Shopify apps because I only know AI as like a chat interface. Right. What are these people talking about?

Yeah. And then in earlier this year, probably Jan or Feb, I decided, all right, let me just go all in and try to figure this all out. And, you know, I set up our Hermes agent and got to Claude Code and Codex and started using all these different tools. And, but it took me a very dedicated couple of weeks to get very immersed, learn the models, learn what model is good for what. And, you know, what does it even mean to go from a light to ultra?

You know what I mean? Like scale, I didn't even learn that until two weeks ago when I realized that it's actually just the speed of what you get versus like the caliber of the model, you know? So I think there is some sort of a learning curve. And I think, like you said, like the challenge is trying to figure out how you get people over that without getting too overwhelmed. Like the way that I look at it, we built a lot of raw capabilities.

Yeah. You can, you can code an app in Triple Whale. You could, you could code almost anything. It's got the data, right? You can code. I was on the call a few weeks ago and we were just like, I was having an experiment with a relatively large client. I'm like, what's your biggest problem right now? He's like inventory and like knowing what's in route and what's the highest, like what's the top sellers and what's the velocity. Yeah. So I had Moby just create an app.

It was an inventory management app. Boom. Right. I was able to do it. Am I going to ship that to all of our users? No. But like you could, if you knew what you're doing, you could actually take it and maintain it and, you know, and do it. There's a lot of raw capability, but the more raw it is, like the harder to get mass adoption on it is as well. Yeah. So we've had to like almost add layers of productization, which we're still doing now, make

it easier and easier. You know, you go into like a Higgsfield right now, I think you just like put an image and maybe a video and it goes into a template. Like on the other side quickly is some creative. Right now, if you go into Triple Whale, like you could put, you could dump your Google Drive in there with your creative files. And like you could stitch together whatever you want, but it is a power tool. You know, we have brands that are doing their entire creative in Triple Whale right now with

Moby. Yeah. And the more that we productize it and make it easier to get, that's where the mass adoption starts to really take place. But we've seen adoption climbing. Create like 30% growth in our Moby sessions every month. Yeah. I imagine as soon as somebody starts using it, it becomes addicting. It's addicting. For me, it's like there's a couple of brands I consult for. I can't, I can't imagine what it would be like working on their stuff day to day without

having Moby to just use. You know what I mean? A hundred percent. Like we have some of the top brands right now that are using it that are saying that it is the most intelligent agent that can employ. They want a physical Moby in their office. Yeah. We heard last week for the first time, you know, it is truly powerful. Yeah. I don't think we're that far away from some sort of a bot or robot in your office physically that like you just ask questions to, you know?

Creepy. I want to talk about media buying too. You got to embrace it. You have to. You have to. I love too that everybody's now changing like agents to characters, you know, to try to make it more friendly. Muse did the best job of that. Yeah, for sure. They took a very complex chat based, you know, product experience. Yeah. Chat GPT, Claude, right? Put a nice, you know, character on there, made it real tangible and real. And my mom could use it.

Right. You know, it's like when you did that, you know, you hit something really. Totally. And I think what they did well is they made it very friendly for like, like our mom to use, right? Like it actually comes to you with suggestions. So that's the thing. Yeah. That's what you really want. Yeah. You want someone coming to you, pushing you things, not you having to pull stuff out. Right. That's really where it becomes magical, you know?

And I think that that's a good segue into some of the media buying stuff. Yeah. That we are doing. Yeah. So you guys have been doing media buying. Well, you just kind of recently announced it right in the last couple of weeks, but they haven't even really GA'd it. Yeah. At all. But it's been something I know Bryan Cano, for example, has been telling me about it for a while. Yeah. Some of the stuff you guys were doing with True Classic.

How has that, you know, I know there was a pretty good incremental lift on some of the stuff that was happening. And you mentioned a little bit like, you know, Moby is able to actually understand every single event inside the change log, build patterns off of that, build memory, build context, understand how a brand should buy with different headwinds or different periods of time. Yeah. Can you just talk a little bit about that? Sure. Okay. So kind of playing off of the back of like, okay, proactive versus like you reacting and

pulling. Right. Like the best employees push stuff to you. They're like, ah, you know. Totally. Now you're having to go like, tell me this, tell me that, give me that, you know. And, um, on the Moby media buying front, which kind of playing with different names for it, to be honest, you know, cause it's not just media buying. It's like full funnel growth. It's like a growth engine. It does CRO. There's landing pages. Like, remember I explained that sort of engine to you where like creative to landing page.

Landing page is published on your Shopify theme, pixels and everything. Seamless checkout. We're doing that right now. Um, it does full funnel. It does retention marketing. It has Klaviyo built into it. SMS from Klaviyo built into it. And so you can come in and give it a goal and it can, you work with it. And we have a context engineering team, our forward deployed marketing team. It can give your growth team, your agency, your media buyers, essentially like that Iron Man suit.

But instead of like this sessions based sort of like chat back and forth, it's, um, sort of organized multiple agents that find hypotheses in your business, propose solutions. The solutions could be change budget over here or reallocate budget over there or try this channel. Accept, approve, accept, approve. And then it learns the lessons from the acceptance. Or if you disapprove, you say why? And it learns that. And you give it a goal and you pace towards that goal and you can work with this agentic

workforce. Um, and you can actually talk to it and interact with it. It can live in Slack, can live in Triple Whale, can live in the mobile app, can live in very soon Muse, live in our MCPs as well. And it's like a, it's a growth team. Uh, it's, it's a, it's an agentic growth engine that can work with your team. Yeah. It's pretty amazing. And, uh, I gotta show it to you. Yeah. After this, the Slack integration is also interesting.

Yeah. A lot of agents have moved themselves to Slack because that, that's where the conversation is happening. And you want it. They hear everything. They know everything. They can keep your team focused. They can be a resource for your team. Right. Yeah. Yeah. How, do you have any good examples of brands that have been testing this and, and, uh, like how they've been using it or anything that surprised you about how they approached it or used it differently than maybe how you thought it would have been done?

I've had brands and I think we have the name right now, but you have a brand that I think within Slack, there's talking with Moby to figure out different, um, opportunities. And Moby found fraud on their website. Wow. And they were able to, you know, say you need to maybe change some coupon. Codes up and that saved a couple hundred grand, I believe. Yeah. They were at scale. Um, but I think a lot of that was happening. That conversation was happening in Slack with the team, you know, or you'll have teams kind

of going back and forth with creative and just in Slack, Moby will proactively can start to recommend ideas full, you know, whether it's image, video, the whole thing right there and like a nice artifact you could download or you could publish it right from there. Right. So you see any use case that you're seeing inside Moby right now, we just bring the party to you in Slack. Yeah. That's where everyone already is. So that's kind of our, the ethos of how we built is like be where the people are.

Right. All the, all the automations that you can do inside Slack as well. Like. So automating reporting, automating budgets, automating, um, creative production messages

to your teammates. Like you can just. Yeah. Is there, um, you know, let's say like Moby is, uh, does Moby know, for example, what my margins are and how I should be optimizing ROAS or how the, you know, maybe a lower ROAS is actually better than a higher ROAS. You tell Moby your goals. Okay. Yeah. And your goals and the completeness of your data is what's, what it's going to learn off of. So if you have your COGS in there, like actually like FIFO.

Yeah. Right. Like how much it actually costs. Yep. You get the real true marginality of products and things like that. Whatever context you give to Moby, it knows and it remembers until you tell it not to. And it will ask you when things changed. Right. Or if there's conflicting people saying conflicting things. It'll proactively raise those things up to you. Yeah. That's amazing. Yeah. It's like, if the data is there, it knows. Yeah. Are there things that Moby picks up that you, for example, a media buyer or a founder would

never pick up on their own in their, in the, in an ad account or. Oh, I mean, look, it's, it's, it's a robot. So it's going to do, it's always on, you know, and it's always looking. It's, you know, it's not waiting. Right. So all day. Yeah. It's finding, maybe you didn't realize you're running low on some stock and it knows because this ad is pushing and selling a certain SKU when it knows what's going on. Right. Maybe Google's taking, you know, claim on purchases that it shouldn't be and they can

type maybe reallocate or. Yeah. Try this new channel. I mean, it's, it's always on. I mean, it's, you have like a rules are nothing new, right? You do rules inside meta, but you put rules with LLMs inside software. So there's like governed rules around it, which is basically what Moby, what, what the media buyer version of Moby is. Right. You can get some real magic out of it. Yeah. Now going on to like the ads and landing pages piece. Yeah. Do you ever feel like, you know, if, if we were to put Moby on this, like we're losing

that human touch or that approach, that special sauce that comes from a human doing the whole part. So the question is like, do we, are we afraid that might happen? Yeah. Yeah. No, for sure. Um, that comes down to like training it. Right. So nothing goes out without approval. So we have a whole approval flow built in and it's got permissions. Who has access to these approvals. Right. Um, as opposed to maybe going into a Claude where it can kind of just go.

Right. Yeah. It has some anxiety along with that as well. Yeah. People operating that thing. Moby has a real governed layer. Yeah. How does that work? The Moby approval layer? Cause I know it's actually something that's a pretty big piece of Moby. Yeah. So like there's business context in Moby where you can basically customize it and our team can help do that as well. Tell it what's green, what's yellow, what's red. Right. You know, and bake that into like, you have a rule that you'll scale all ads that are

1.7 ROAS or something like that. You put it in there, it's saved. It's not going to hallucinate cause it's also built, not only is the LLM in there, but you have software in there. Yeah. The checks and balances. It's basically the software system, right? With the LLM inside of it. So it has guardrails built into it. Right. And then you put your rules in there and like, it's, it's good. Yeah. It's good. So whatever you tell it to do, it will do.

Yeah. Let's say we struggled with hallucination at the beginning. Right. Like everybody. Yeah. Once we got more and more into this and more and more people came into the space and more solutions were out there with how to deal with these things, how to govern these things, like we were on the top of all of it. Yeah.

What do you think, um, changes for brands that start using Moby on a weekly basis? I think the, you know, if you have a founder that's AI pilled and obsessed, you know, you can come to a Triple Whale and you have everything you need to really start to transform your business in the new era that we're going into. And I think that once you put, when you marry, you know, the human strategic and human creativity, right? I think you're always going to need that stuff, the direction, the feel for what people want.

I haven't really found that AI can replace that, which I think is a good thing. Yeah. Right. But you start to put on an Iron Man suit. Right. Sort of semi-robotic. You get crazy efficiencies and then you get insights from, let's say our benchmark dataset. So you try to see like what actually is working out there. Like you can start really, you know, have some outsized results from agentifying your whole organization. Yeah. And we've seen businesses, you know, from on the low side, an average of like 10% ROAS

efficiency to up to 30%. We've seen it. Right. And that's just on the media buying side that we started. Right. The landing page creation is relatively new, but when you start to push CVR up, which I think one of the best ways to do that is through dedicated landing pages and offers. Totally. And matching creative as well. You start to see even more. These things start to compound. And then let's get, you know, your logistics data in there and getting AI on top of that

can start to really squeeze out and new opportunities start to come. Yeah. I mean, even. You could see how like the optimism around AI can just bring so much opportunity to the world as opposed to like, say the Doomer side, which is like, oh God, we have no more job. We're all screwed. No. I think there's maybe going to be some displacement or something like that until we figure it out. But that happens with every revolution. Every revolution. Yeah. Yeah. Yeah.

I think also. Gotta lean in. The other thing you get that, you know, is not talked about as much as like the speed and convenience more than anything. Right. Like you jump on X and you see that there's a new format of ads that work well from, you know, these animated AI ads that go to these listicles with those AI characters. Yeah. And previously that's a four or five week sprint from start to finish. Right. Now that's maybe a four to five hour project Maxx, you know.

And used to have a team. They used to have maybe outsourced, you know, hire actors, things like that. Yeah. There's a video we made many years ago. Funny video we did back in, back in the day. Um, it was like 250,000 bucks when you were casting scripts, actors, you know? Yeah. We were just messing with some of the newer models. I think we put, uh, was it 5.5, Sonnet 5.5 [model name/version unclear]? The newest, maybe I'm getting the wrong one. Seedance 2.5 together.

Mm-hmm. And told it to like watch that video. And it made a video in one prompt in like an hour that was on par. Wow. You can really, like the video capabilities are really coming, coming to light. And it's unbelievable. Yeah. The, the, the video stuff is next level. Um, and just how realistic it is too. Oh my God. You know, it's, it's, at this point it's, you can't really tell. Yeah. If you're like, if you're in there with it and prompting it and, and really governing

it, like you can, you can get unbelievable outputs. Yeah. Is there, um, like, what do you think, uh, for the differences for a brand that's doing,

you know, five to 10 million a year versus 50 to a hundred million a year? Like the smaller brands have even more access to like opportunity, right? In terms of systems that help you scale. The like 10, 15 years ago, let's say 10 years ago, you're running a brand. Like it's hard. You're like doing everything as a founder. You can start to really put some agents on your, your org chart. If you have one, you know, you have an org chart, right.

And start to leverage and get more out, you know. Are there specific agents you see people set up first? Like for example, maybe on the media buying side, maybe on the creative side, maybe on the project management side. It always starts with reporting. Yeah. People wanting to. Dashies. Everybody wants a dashie. And people still want our dashboard. Right. They just, they come and they even have everything set up. Like I see the brands that are super AI pilled and they've got like everything.

MCP. And they've got Claude talking to Moby. Crazy things. But at the end of the day, they're coming back to like our summary page because they want real time. Like it's always on. So there's always going to be a place for, you know, dashboards and things like that. Yeah. But in terms of like the first use case that we see that people set up, it's always around reporting. Business reporting. Yeah. Accurate. Like actually getting accurate reporting that you can share with your partner or your

team or if you're a bigger brand, like, you know, the C-suite and things like that. Right. Consistent. Consistent. Yeah. And I think to the dashboards. Then it goes to like media buying, like budget. Yeah. Budget changing and things like that. It kind of goes over to like, okay, creative. Right. Can we play with creative? Yeah. I find. Static. Because statics are, you know, they've been good for a while now on the AI side. Right. Yeah. And only getting better with Astra and the new [model name unclear].

But I think dashboards I've found are kind of the easiest way to get people onboarded to AI. Like everybody that I've probably gotten, I don't know, 50 to 100 different people really immersed in AI this year. And maybe 30 to 40 of them with, you know, a Mac mini with a Hermes setup on there. And the first thing everybody does is basically set up dashboards. You know, they want to see how to set it up. The fact that they can see they can build their own dashboards and connect multiple data

sources and all of that. And then it kind of gets them that itch to go and see, all right, what other agents can I go and make? Yeah. Yeah. Yeah. I mean, our approach to that was like, okay, we have like our, our Keystone dashboard and like that summary page. It kind of gives you, you know, the lay of the land that we brought in like Moby on top of that as a chat box and recommended prompts based on the data you're looking at and their

dynamic and like that shot usage up too. We also put that on like the attribution page and things like that. Yeah. And then that's a good, like, you know, lever. Totally. Yeah.

Couldn't agree more. Yeah. Um, you guys have the BFCM command center here coming up for Black Friday. A lot of brands are already, you know, basically switched it on and have started working with it. But, um, could you tell us a little bit about what, what's in the BFCM command center and what makes it special and how is Moby powering that? Yeah. Essentially the command, the BFCM command center has, as you onboard, it tells you what's, what's your goals. What are your plans, your goals?

How did you do last year? How do you compare it to your competitors? Right. Whether it's from Facebook ads, uh, public library, or even our, our, our benchmark data set. And then you plan it, you go, it tells you what you're missing, what you need to do. And you can essentially track how you're pacing throughout the whole weekend. Live pacing, live data. Basically you have like an NL, an analyst in your whole team. Right. Right. Like just essentially have your meeting, your stand up with that thing on the screen,

share, watching how you're doing. Yeah. And what's interesting too, is like the whole team has access to that every minute of the day versus just finding those numbers in those meetings. Correct. Yeah. Jump in, check it out. It's right there on the, in the nav. Yeah. Pin it, come back to it later. Amazing. What is, uh, what's something that, you know, people, they'll onboard to Triple Whale or, or maybe they're on Triple Whale. They haven't started using Moby yet. What is something they can go and, you know, they're listening to this episode now they

can go and do tomorrow to start to see that, uh, that ball or a snowball starting to roll down the hill. I mean, right now for free, you can go on to Triple Whale and start and just connect all your, yeah. And the free plan, you get all your data. So start connecting all your data and you can start to automate reporting. Right. But you can also start messing around with the creative stuff. Right. Yeah. Once you connect your Facebook ads, it'll look at which ads have been performing well

and you can have it make five variations of this ad. Right. Make it a flat lay. And the difference by the way too is, uh. You can do that tomorrow. Yeah. Moby has a lot of skills to make ads built in as well versus if you were to go to like a Claude, you'd get something totally off brand. This is going to come. You know, it looks like AI. Exactly. And Moby is, is really purpose built for e-commerce. Like our team is building skills, right?

To help you, to help you get successful relatively quickly. Yeah. Also when you onboard, um, we have a brand guideline that will automatically start to like basically scrape your site, get your colors, get your logo, look at your assets and make sure that what it produces looks like your brand. Right. Which is huge. Yeah. Is there anything I didn't ask you that you feel like founders need to understand or

marketers need to understand about Moby or about Triple Whale or kind of how the product is, you know, where the product is today and like what it does for brands? I mean, I really think we hit a lot of it. I want to say that like the measurement piece, the pixel piece. Yeah. It's really at the heart of all this, right? Like one of the moats that we've built for ourselves is this data, right? The way that it's different than let's just say a Claude is that first of all, it's purpose

built for e-commerce, but at the core of it is the source of truth on customer journeys. Right. Which ad channels are most profitable? And how do we know? And so the pixel at its heart, the attribution at its heart is like the heartbeat of this thing. Right. Because you can trust it. Right. And so not only that, we have this data, but as a merchant, you can leverage like our, we have custom attribution with meta. So you can take an attribution model and put it back into meta and it'll optimize off that.

Wow. Like that is showing results, big time results for brands. So like you have this whole system, you know, you have this whole system and it's like built on the back of performance marketing architecture. And it's just like, you can unlock, you can unlock. Unlock is probably the big keyword here. Like a lot of this stuff allows you to unlock, you know, whether it's a process you wanted to try, whether it's a skill you wanted to build, a routine you wanted to build, all these

things that you've maybe wanted to do at some point that Moby is, is really the unlock to get to that other side. And if you're busy and like, not everyone can sit around and try everything. Right. Talked about, our team can help. Yeah. We have a context engineering team. We can deploy into your, into your business and help. What is that number one workflow that's going to be the most impactful to start to kind of agentify it? We can help you do that.

Yeah. Yeah. I think it's awesome how you guys are actually helping brands like get on board and start to use it, you know, not just. The adoption curves are all over the place. Right. You know. Yeah. Amazing. Well, Maxx, where can people reach out to you or if they've got questions or just maybe follow-ups they want to ask about their business? Reach out to me on, on, on X. My email. I don't know if I should publicly say it. I don't care.

People can guess it. Yeah. It's not so hard at all. Maxx with two X's at Triple Whale.com. Reach out to me. I hope you get set up. I, I like to like still be in the trenches very much with brands and businesses and the entrepreneurship story is my story. It's also like what I love to do and help people grow and build their own businesses. Yeah. That keeps me, that keeps me fired up. Amazing. You can reach out to me there.

Perfect. Well, Maxx, thank you for coming on and explaining all this. It was good. It was awesome. I love our interactions and, uh, evolution of our friendship and business relationship. So. MentorPass. Yeah. From MentorPass to sending you voice memos at three in the morning. Yeah. It was a good time. Yeah. Yeah, no, it's been awesome. Thank you for coming. Thanks for having me.