[STEP BY STEP] The Secret to Smarter Meta Ad Spend: Guess Less With A Winning Mega Data Platform

Phillip: [00:00:10] This episode of the Step by Step podcast is sponsored by Triple Whale. Unlock better ecommerce performance with solutions like Sonar from Triple Whale. Learn how brands like Zox, Allen, and paw dotcom increase conversions and ROAS by visiting Triple Whale today. Go to triplewhale.com backslash meta to learn more. Hello, and welcome to Step by Step, the podcast by Future Commerce, presented this season by Triple Whale.

Brian: [00:00:44] In this 14th season, we're exploring how to optimize your meta ad campaigns for maximum impact. Whether you're working with a high GMV brand or just striving to amplify your marketing strategy, this season of Step by Step is here to guide you through creating meta campaigns that actually drive real profitability.

Phillip: [00:01:01] That's right. And in this season, brought to you in partnership with Triple Whale, we dive into the tools, the insights, the strategies with real people who are helping you and brands just like the one you run to leverage Meta campaigns and Meta's platform more effectively. So over 3 episodes, we break it all down into its essential parts from enhancing customer data to refining ad spend, and we help you to grow the bottom line in your ecommerce business. Brian, who is this podcast for in this 14th season?

Brian: [00:01:33] If you're an ecommerce professional managing a $10,000,000 plus GMV brand, this is probably a good thing for you to listen to. Or if you're responsible for any operational marketing or data driven strategies within that brand, this is probably for you. And honestly, if you just need to know how to convert your data into actionable growth through Meta, this is the right podcast for you.

Phillip: [00:01:56] And I'm really excited because this first episode kinda gets us, uh, started with talking about a big shift in strategy and a big shift in technology. Uh, of course, we are, uh, talking about, uh, how, you know, Triple Whale has grown up out of just an attribution platform into a very capable set of suites and tools that help you think around, uh, think about what the future of your business is gonna look like from a paid acquisition strategy perspective. So, uh, this season, we're we're giving you, yes, practical insights from industry leaders, and we're gonna kick it off with Kellett Atkinson from Triple Whale with this first episode. Uh, Brian, what's 1 or 2 things we need to listen out for in our time with Kellett?

Brian: [00:02:37] Yeah. This episode is called why megadata platforms are more effective than multiple tools. And this is super, super important, uh, as we look at how Triple Whale's sonar integration with Meta streamlines data collection and minimizes conversion loss. That's that's a huge part of this. Advantages of unified platforms for improved customer experience, cost efficiency, and why a single source for data is better than just random customer insights out there that you have to compile and put together and figure out what to do with.

Phillip: [00:03:10] Yeah. And and sending the right data to Meta, uh, so that you can attract the right customer, uh, and send the right conversion types of of data, I think, is really important. Also, we learn about CAPI. And so I I think there's a whole lot here, especially if you are a senior leader and you're trying to get your head around what your 2025 planning is gonna look like. It's a great episode. So without any further ado, uh, let's go into our conversation with Kela Atkinson from Triple Whale about why megadata platforms are better and more effective than multiple tools.

Brian: [00:03:48] Hello, and welcome to Future Commerce, the podcast about the intersection of culture and commerce. I'm Brian. And today, we have a special season of Step by Step Ahead for you. Uh, and we're about to jump into, uh, one of these episodes with Kellett Atkinson. Uh, and I cannot wait to get into it. Welcome to the show, Kellett.

Kellet: [00:04:10] Hey. Thanks, Brian. Thanks, Philip, for having me. Super excited to be here.

Phillip: [00:04:14] Yeah. Absolutely. We, uh, uh, where you where you calling in from today? Where are you based?

Kellet: [00:04:19] I'm in Raleigh, North Carolina. Sunny Raleigh, North Carolina.

Phillip: [00:04:23] Is it is it sunny this time of year?

Kellet: [00:04:24] It is right now. We had our false fall, and then we came back to the winter or to the summer. Yeah. So now it's 80 degrees outside.

Brian: [00:04:33] Fall at all. Your fall is

Phillip: [00:04:34] Not happening. We might we might have skipped it

Kellet: [00:04:36] Over entirely. My, uh, my lawn will be happy about that, though. Maybe I'll get some grass this year.

Phillip: [00:04:42] Nice. Nice. Good. Good for you. How long you've been in Raleigh?

Kellet: [00:04:46] Oh, I've been in Raleigh for 20 years, maybe, since high school. So been here.

Brian: [00:04:52] North Carolina is beautiful.

Kellet: [00:04:53] It is, especially this time of year. I think people are discovering that. It used to be, like, a good secret, but now it's it's not so much a secret. So we got a lot of people coming from all over the place in the Raleigh. It's growing pretty fast.

Brian: [00:05:05] Is it Apple coming into Raleigh, or is it somewhere else?

Phillip: [00:05:08] Or They

Kellet: [00:05:09] They committed to coming to Raleigh, like, I wanna say right before the pandemic, but, uh, and they, like, carved out land and maybe they even started building stuff, but I I don't think they've actually planted their roots yet.

Brian: [00:05:23] They didn't Villageite Dome or Yeah.

Phillip: [00:05:25] I I wish they, uh, if they could take some of the Raleigh energy and put it into the Apple Vision Pro that I paid a lot of money for, that would be super cool. There's not a few

Kellet: [00:05:34] Of those floating around when they first came out. I haven't seen them recently.

Phillip: [00:05:38] Well, no. It's a that's a whole other story in and of itself. Um, you your your roots go deep then in sort of the the tech space, and, uh, these days, you're, uh, in product management right over at Triple Whale. Give us a little bit of the the background in a nutshell. Um, is this your first go at Ecomtech?

Kellet: [00:05:59] Ecom, yes. And that, uh, I can tell you coming into Ecom was a, uh, an eye opener for a lot of reasons. But prior to ecom, I did analytics. So I actually started my career on the supply side as a working with a publisher, building mostly ad tech and marketing tech to help connect advertisers with audiences. And then I moved into, uh, product analytics. I worked at a company called Pindo, which is a big product analytics company that serves more SaaS. Then my I had a buddy actually who, uh, uh, joined the sales team at Triple Whale, and he's like, you gotta come over here. This is like this company is blowing up. They're doing really cool software. You'd really enjoy it. And I was like, okay. Why not? Looks looks fun. And I got in and, like, I'd never been in ecommerce before, and it, like, it was mind blowing. Like, just how how active the ecommerce scene is, how, uh, like, you know, many people terminally online in ecommerce is, like, a whole a whole culture that I had to, like, assimilate to pretty quickly. It was super interesting, though. I mean, I learned it's cool, because I feel like I'm learning every day all sorts of stuff that I wouldn't have if I, uh, hadn't moved into e commerce specifically.

Brian: [00:07:21] I mean, that's the thing about commerce. It's it's so deep. There's so much to it. There's so many moving parts. There's so much going on. It changes every day. It's like a constantly moving target across every industry, and that makes it really fun.

Kellet: [00:07:38] Yeah. And I think one of the things I really like about this industry is there seems to be, like, a really strong openness to sharing what people are learning. Like, their people are super vocal and, like, really eager to share what they're seeing and learning. And so there's, like, there's no shortage of opportunities to learn from other people, which is really cool.

Phillip: [00:07:58] The, uh, you're couple years now over at Triple Whale. Right? Yep. Yeah. We Yeah. That's right. Yeah. My my sense of what Triple Whale is or does these days might be pretty narrow. Uh, I know you all, you know, have attribution Right. Software. So you you can help people that spend a lot of money in a lot of paid channels figure out where to direct that spend, how to spend more effect effectively. Uh, is that the gist of it? Is that too simplistic?

Kellet: [00:08:28] That's probably yeah. I'd say well, that's probably what most people know us for today. Um, you know, if you look back to when I joined Triple Whale just a couple years ago, the company's changed a lot and the product has changed a lot. But the when I started, you know, Triple Whale did 2 things really well. Like, 1, we had this this product that really was the origin of Triple Whale that we called the summary page, which was really just a data aggregator. It was like a way for people to connect all their marketing channels and their ecommerce channels, and just bring all the data into one place. And right at right before I started, Triple Whale had launched the triple pixel, which was kind of our response to the, you know, the the privacy landscape changes, and, like, giving people a way to do first party tracking on their website. And that was really a catalyst for triple l, and, um, we layered on top of that a bunch of cool novel attribution products. People could look at their, you know, marketing data in different ways and make decisions about how to optimize their campaigns, but since then, um, say over the last year or so, Triple Whale has kind of pivoted into being a more unified data platform. So something that, um, while attribution is still our bread and butter, people come to us to make marketing decisions. We've invested a lot into incorporating more holistic data sources into the platform and providing kind of, like, a new set of capabilities on top for analytics, the predictive models, you know, ways to easily access your your ecommerce data through chat or through, um, automated workflows. Mhmm. Um, it's really like the it's been a big shift over the last year, and I think we're kind of ray riding the AI wave too just like a lot of other SaaS companies are. Yeah.

Phillip: [00:10:25] There's been this, uh, thing that Triple Whale has done really, really well in creating sort of a cohesive ecosystem of solutions. Willie, I think, was the the former name of the of

Kellet: [00:10:40] Of Yeah. Yeah. We had to we had to change that, by the way.

Phillip: [00:10:43] No. Yeah. I don't know if this gets redacted. I I wrote this, um, um, about a year and a half ago, I think, when the product was originally announced. I wrote a piece sort of ingest, uh, about the personhood of AI. Uh, you know, Blake Lemoine was a a Google researcher who, uh, left Google and, uh, had this moment where he felt like he had this crisis of conscious that maybe that the AI was sentient and he was, you know yeah. So there's really interesting, like, uh, story in there about the way that we, like, personify these machine learning systems, especially when you're chatting with them and, like, that becomes part of your job and now you're, like, developing some kind of relationship with it. To call it Willie also, you know, implies, uh, some nostalgia from a certain type of a a movie. So it's some type anyway, I created an online petition on, uh, on change.org to free Willie, um, from triple from Triple D. If you know that I did this

Kellet: [00:11:38] We did it.

Phillip: [00:11:38] We freed him. Yeah. Yeah. You never really freed Willie.

Kellet: [00:11:41] Yeah. He would took the petition seriously. We knew we had to do it.

Phillip: [00:11:45] Yeah. We did it. Um, but I think you guys have done a a really good job in sort of world building in the product sense of there's a growing need to engage with data in different ways that serve different facets of the market in different roles. And the the needs are changing because the platforms are changing. And to be honest with you, like, privacy and, uh, app tracking, uh, policies are changing too. So to your point, you know, the product has change and evolve dramatically, and that's what we've seen over the last few years, I think.

Brian: [00:12:18] I think attribution is such a great place to start from when it comes to evolving into a platform like this because, uh, your data's gotta you got it's gotta be clean. It's gotta be right. In order to have, like, a good view of it, it's gotta start from a good spot. And so there's a lot of other people trying to approach this dashboarding from a different angle, but I feel like having the clean, like like, correctly identified and attributed data is, like, the best place to start from when it comes to analyzing that data. Garbage in, garbage out.

Kellet: [00:12:55] That's right.

Brian: [00:12:55] Like, garb like, good stuff in, good stuff out. So I think that, uh, what a what a what a move I found.

Phillip: [00:13:04] I like that. Good stuff in, good stuff out.

Kellet: [00:13:06] It's especially true when you consider I mean, besides maybe inventory and headcount costs, like marketing expenditures, typically one of the top expenditures for a brand. So, like, being able to really understand the effectiveness of that, What what you're investing there is a super critical pain point for most businesses. So it makes sense as, like, a as a good place to to get your ducks in a row as a business first. And so Exactly. That's where it makes sense for Triple Whale too.

Phillip: [00:13:37] Let let's say that, uh, I don't I it's kind of like nobody likes their accountant. Um, also, uh, no, I don't think anybody feels like their data's in a place where they're, you know, they can take that next step. Uh, what are the kinds of things that you see are are problems and and blockers for folks, uh, that are adopting this kind of software these days, especially in that, like, mid market to enterprise, uh, range? What are those what are those blockers that you guys encounter a lot?

Kellet: [00:14:09] Yeah. I think well, they're probably similar to any business, not just commerce, but I think one of the things is, um, when you get to a certain place on, like, the data driven maturity curve, you're like, maybe you have a lot of data, maybe you're centralizing your data, but there's this, like, effort to even if you even if you have dashboards in your business, and you you're you have tons of data streaming in, it doesn't always solve a problem just by doing that. The problem like, there's still this problem of, like, what do we care about? How do we build a unified language about how we think about this data and talk about this data to make decisions? And that's I see a lot of even mid or larger sized businesses struggling with that, which is, like, looking at the data in the same way to make good decisions. And I think that's, you know, that's one of the things Triple Whale tries to solve is to to create, like, almost this semantic layer on top of your data that gives it a singular lens, like, a definition. Like, it's we're kind of opinionated in the reports that we offer. We make them easy to to consume, but I think that's like, as a business, that's always the the most challenging part is, like, making sure that people are looking at the data in the same way and deriving the same type of insight and making the same decisions off the same set of data. So that to me is probably one of the biggest challenges that even a large scale businesses have.

Phillip: [00:15:36] Brian, that's a language game. That's that's the thing that you're always talking about

Brian: [00:15:39] Is 100%. Yeah. Yeah. That's good.

Kellet: [00:15:43] Uh, I mean, we like, I can speak firsthand. That's even a problem we have, you know, in Triple Whale as we grow is, like, one of the main problems we have is, like, we we have 2 different people looking at ostensibly the same data and coming to different outcomes because they're looking at it through a different lens.

Phillip: [00:16:04] This season of the Step by Step podcast is brought to you by Triple Whale. Are you struggling to make sense of your data? Well, don't worry because Triple Whale offers a range of solutions that are designed to streamline your ecommerce operations. For instance, with their product, Sonar, you can transform your brand's meta campaign performance with just a few clicks. And if you don't believe me, you can learn more about how brands like Zox, Allen andpaw.com, and some of the others here in this very series got increased conversions, accurate attribution, and higher ROAS. Find out more today by visiting triplewhale. Go online to triplewhale.com backslash meta to learn more. That's triblewhale.comback/meta. Give Triple Whale a try today and let them know that you heard about it on the step by step podcast from Future Commerce. Go to triple whale.com backslash meta.

Brian: [00:17:05] Subjectivity is is inescapable. Right? You can't you can only view things through your context. I think the interesting thing is what happens when those 2, uh, like, sort of subjective moments come together to to Right. There are people bringing their perspective or their observations. And, like, what happens when those observations interact? Uh, do they result in in in a new perspective, or do they do they just continually crash? I think that it this is where people really come into play, uh, and understanding how to work with other people is actually essential for good data analysis.

Kellet: [00:17:47] It's a really good point. You know? And it's it's like one of the things I find as we're trying to, you know, do things like build AI AI insights for people. It's like, every person brings a whole bunch of context that that it's really hard. I I don't know if AI is there yet, or when it will get there, or if it will get there, but being able to assimilate, like, the context that a person has that sometimes bringing unrelated perspectives into into play to, like, to draw conclusions about stuff. And I think it's like there.

Brian: [00:18:20] Big hurdle. Like it's been there.

Kellet: [00:18:21] You you feel like it is? Yeah.

Phillip: [00:18:24] Yeah. An optimist at heart. I have the I think you are.

Brian: [00:18:29] No. Not a a big fan.

Kellet: [00:18:33] There, but I'm I'm skeptical that it's there yet.

Phillip: [00:18:36] Kelly, I'm I'm on on the on the same page with you. Let me let me say why I think Brian's an optimist is because in the 10 plus years that I've known you, Brian, you every single person you've ever met ever, you've you've been pretty bullish on on their, uh, on

Kellet: [00:18:52] Us being

Phillip: [00:18:53] Able to work together in some capacity. Like, every person.

Brian: [00:18:56] Well, that's because I'm arrogant. That's really the answer to that. Yeah.

Phillip: [00:19:03] I I think that it's an interesting conversation in this ecosystem that it keeps coming back down to context. Right? Giving advice requires understanding of greater context. Right? And that's so fascinating to me because the AI conversation has everything to do with context. But we tend to give a lot of advice in this ecosystem, especially on the social media side of things. It's like a lot of the context is lacking in giving of the advice, but as you gain greater context, like, we ran an agency for many years, and Brian and I spent, um, I was 10 years at the agency that Brian and I were at before Future Commerce. And everybody would ask us the same question. It's like, what are you seeing with your other clients? They wanted they wanted to understand our context because we could see more than they could see.

Kellet: [00:19:54] That makes sense.

Phillip: [00:19:55] Right? And and I see now that, you know, Triple Whale is the kind of company that has greater context for what's happening in the whole ecosystem. But now that you have such context, you realize how much more important the that your overall perspective of what's happening in one business can't necessarily inform what's happening in the context of any given business. And I I think that's that is where the challenge is is, like, is it getting better, or does it actually make us realize how independent we are from each other?

Brian: [00:20:27] Yeah. And this gets into the phenomenology sort of realm because can you actually take the experience of 1 business and reinsert and reapply it into the context of another business? Because all moments are all happening individually at all times, and nothing is ever the same. And so

Phillip: [00:20:48] Kelly, that's a good that's a good question. Like, what are the kinds of things that, you know, that you can provide insight on that could be somewhat universal or targeted to the context of an operator?

Kellet: [00:21:00] I think one thing that is a truism across all the businesses that we work with is that the ones that are more successful tend to leverage the data they have in a in a meaningful way, not just for analysis, but actually for creating cohesion across their tools and for activating data within within their other tools. What do I mean by that? Like, specifically, you know, one thing that this top of mind for me is we've we've recently released a a product called Sonar. And the role of this product is really to take the data we already aggregate on your behalf, um, and push it into your ad platforms. Now people who advertise on, say, Meta know that, really, your performance on Meta is driven by how much data the Meta algorithm has. It's really, like, you feed the beast, and it and it gives you better results. That's really what it comes down to. Um, and so I think companies that have applied smart strategies and making sure that the the the tools that they're outsourcing work to, like Meta, have are working with the most complete information. Those are the companies that tend to succeed the best.

Phillip: [00:22:22] Wow. The when when you're talking about is is there some sort of an industry wide shift that we're we're looking at that's gonna push more people towards that as a requirement? For instance, um, 10 years ago, most people would have said, well, we have an organic SEO strategy because people search us every day.

Kellet: [00:22:45] Right.

Phillip: [00:22:45] Um, on the Internet. They go on google.com and they type in, you know, my brand name. And so, you know, we we can invest in making ourselves more discoverable. Then, you know, 5 years ago, people had to bid on branded search terms just to be up at above, you know, in that number one spot. Right. So that changed dramatically. Uh, what else is happening that's kinda pushing people more toward that direction of having to understand, like, your job to acquire customers requires you to spend in certain channels?

Kellet: [00:23:18] Yeah. So, uh, I mean, I think, obviously, one of the ones that's been around for a while and I think is part of why triple Whale exists in the first place, um, is that there you know, the the Internet used to be the wild west, and people could it it wasn't hard for a company like Meta to track you across the Internet. It wasn't hard for them to know what you're doing at every step of the way. Um, that's obviously changed in the last couple of years. Well, probably a little more than that, but last few years. Um, and, you know, the you know, when you when you take something that you're you're investing a ton of money in, and there's and and it relies on the quality of the data that they're using to target audiences and things, um, the just the changes to the privacy landscape and the the Internet tracking, the ability to track with 3rd party cookies is really cut like, originally really, like, crippled the ability to do that. Um, and I and I think, you know, the push will continually be to make the Internet a more privacy conscious place, and that's that's why, you know, SONAR is, at its core, a an out of the box connection with with the with, in Meta's case, Meta CAPI. And CAPI CAPI is Meta's answer to this change in the privacy landscape.

Phillip: [00:24:46] Oh.

Kellet: [00:24:48] It is Say

Phillip: [00:24:48] Say more about that. That's what's changed. Yeah.

Kellet: [00:24:52] So, um, specifically, the ability for for a third party to track and understand who's doing what on your on your website. I mean, the restrictions to third party, um, tracking, you know, iOS changes to iOS, and, like, the ability to to track users across apps, and and websites, and devices have really made it hard for a third party like Meta to really keep track keep tabs on who somebody is across the Internet, and what they're doing. So Meta's response to that again was to release what's CAPI, which is their conversions API, which is a a server to server connection to allow you to pass first party data to meta to really complete their picture, where where all these gaping holes exist now, because of the the browser restrictions. Um, and and each of the major ad platforms have also, like, pushed towards this direction. I think at some point in the near future, uh, they still, like, push on us to couple our a web pixel implementation with an API implementation as the best, like, tracking solution,

Phillip: [00:26:05] But

Kellet: [00:26:05] I think at some point in the future, we won't need web pixels. Like, we'll just be pushing, you know, like, anything that's happening on your website that you need to feed to meta so they know who's visiting and how the conversions are happening. I think it'll all be server to server, because that that's, like, the most hardened way to ensure that everything's being captured, that is being communicated, that Meta can use it, Google can use it, anybody else. So I think in the future, that'll be a really pivotal part of a marketing strategy is making sure you have the right data infrastructure to feed these platforms, the data they need to to serve your ads.

Brian: [00:26:45] How will that change how people plan their marketing spend? Because I feel like that sort of that level of of data, uh, passing well, I mean, it's it should it should change things.

Kellet: [00:26:57] Yeah. I mean, it, like, it it it implies a need for an investment in in data infrastructure, which, you know and today, you can go on on Meta, and you can sign up for an ad account, and you can place their web pixel on your website. And that's you know, it's a 5 minute exercise, and it's just putting some JavaScript on your website. Um, but if you have to have a server to server connection, and you gotta be doing ETL on your data to prep it and send it into meta the way they wanna receive it. It's like a you have to build the infrastructure to do that. And not, you know, as particularly in commerce, I don't know. Well, in ecommerce, in the DTC space, um, specifically, techno like, having a having a full team of developers to do that is not something that a lot of brands have. Right?

Phillip: [00:27:50] Especially now because if sorry to cut you off, but, like, there's there's a timeline here where ecommerce required an IT organization. Sure. Yeah. Especially when it was it was, uh, hosted, deployed software. Yep. You would like, talk about privacy. A lot of these organizations saw companies like Amazon or Meta as having insight into their business that they felt like was, uh, a potential risk or some sort of, uh, realm of competition for them.

Brian: [00:28:24] Which is not even a narrative anymore. Like, that's not

Phillip: [00:28:27] Even a thing anyone talks about anymore. Right? But, um, which is, like, a dramatic cultural shift inside of this because it that's just 10 years ago. 10 years ago, people did not want to give these systems insight into the the size, scale, or, uh, the data in their business. So that that that's changed dramatically. But the the way that people used to build the software required a heavy IT investment, which did have developer savvy, which did think very deeply about this type of stuff. But as we've moved through the cloud era and now a lot of that are these black box solutions that provide a customer experience and just do things in the back end, We not we also have hollowed out those IT organizations that actually understand how this all works.

Kellet: [00:29:09] It's true.

Phillip: [00:29:10] Have these generalized solutions that you pay a lot of money for that accomplish a certain end goal, but you've also hadn't never had less visibility into how all of these systems are working in effort together, and you'd also aggregating that data back to being owned data within the brand also creates this fragmentation that I think lends to that thing we were talking about earlier where everybody feels like their data house is not in order Yeah. At least in this in this echelon of of the the type of, uh, you know, an operator we're talking about.

Kellet: [00:29:40] I think that's right. And I and I agree generally. I mean, it's definitely the case that it's it's never been easier to stand up, at least online, a brand

Phillip: [00:29:51] Mhmm.

Kellet: [00:29:52] Or an arm of your brand and start selling stuff. And it's all abstracted away to a level where, you know, somebody who's not a developer can do it or somebody who's this, you know, semi tech savvy person can do it. But it's definitely true that I think and this is probably a trend more generally just in across all business that everything's being abstracted away to a level of, uh, who knows? Maybe there'll be a point where nobody knows how any of it works.

Brian: [00:30:23] Black box, baby. This is, uh, this is where we're at

Phillip: [00:30:27] Now. Got it. I guess that's AI.

Brian: [00:30:29] Into the AI.

Kellet: [00:30:30] Yeah. That's AI.

Phillip: [00:30:30] That is the AI. That is the AI.

Kellet: [00:30:32] Hopefully, the AI knows how it works, and it's working, uh, and it and its motivations align with ours.

Phillip: [00:30:38] Well, I I think we are there, um, that I am pretty bullish on, especially in the context of how businesses would deploy some intelligence within their organization is trying to as

Brian: [00:30:51] Bullish as I am on this.

Phillip: [00:30:52] I don't know. I don't know about that. I I actually feel like, uh, there's a type of we're terribly uncreative about the kinds of things that we could explore within our business because there is a technical friction to being able to explore those and ideate on those ideas. So creating hypotheses within an IT organization, never mind an ecommerce organization, requires a lot of like, they think of things in terms of Power BI. That's how they think of things. They think of things in terms of data lakes. So that they're not thinking in things in terms of what if I have 15 ideas of the ways that my customers are chatting or or are using with my data and maybe I can have a deeper conversation with my data as opposed to having to give very specific briefs and requirements for this the set and outline of a report or systems reports and dashboards. And that is the friction. So what if that friction goes away? What how creative could we be on the other side of that?

Kellet: [00:31:54] I think that's that's like I think you're hitting the nail on the head for, like, the that's like the $1,000,000 question right here is, like, particularly with AI. And I think the the promise of AI and I think the direction that even triple oil is heading is, like, being able to do that well within a a vertical. Like, I I don't I know there's a you know, we we talk a lot about AGI and but, like, I think the strength the ability to do that will come faster in some vertical where you can, like we talked about earlier, build context around the the AI that's domain specific. Like, this is one of the things we work on at Triple Whale today. We have you know, we recently released a product called workflows that really is just like it's it's like it's not quite an AgenTic workflow, but it is it's a pseudo agentic workflow where you can string together a whole bunch of, like, pre contextual questions through an LLM to go fetch data and and then do run a generative analysis on top of that data. And so, like, we're, like, asking the question internally, like, what are the right templates for these things that we could dispatch on behalf of a customer?

Brian: [00:33:04] Totally. No. And and I think this is why I'm bullish. We've never had technology been able to assess context better than this before.

Kellet: [00:33:16] Right. Right?

Phillip: [00:33:18] I do agree with that.

Brian: [00:33:18] Yeah. I did just ask perplexity if Triple Whale was a person on a podcast, and I asked that person how Triple Whale could help businesses share feminological experiences with each other, what do you think that that person would say back to me? And here's what they it says they could say. Okay.

Kellet: [00:33:33] I'm super interested to hear this.

Brian: [00:33:34] Triple Whale empowers businesses to share phenomenological experiences by providing a comprehensive view of their operations through our unified dashboard. By leveraging advanced AI and data visualization tools, companies can gain insights into their customer interactions and operational efficiencies. This shared understanding allows businesses to connect over common challenges and successes fostering a community of learning and growth. Our platform not only tracks metrics, but also helps businesses tell their unique stories through data, enabling them to share those narratives with others in the industry. This exchange of experiences can lead to innovative solutions and collaborative growth opportunities that benefit everyone involved. And I would say that's about as good of context as you can expect from technology at this moment. Yeah.

Kellet: [00:34:20] I was gonna say, I'll just I'm gonna I'm gonna log off now. You guys can interview perplexity, triple will.

Phillip: [00:34:28] Uh, I was gonna say, Brian, if if things don't work out at future commerce, you might have a job at triple whale. Yeah. Yeah. From the back side.

Kellet: [00:34:34] Exactly. But no. But it is, like, uh, I do, like, I think one of the things that has been this is why I'm not so bullish, maybe. In Triple Whale so we do, like like, Triple Whale, if it for the ideal customer, we have access to pretty much everything about that store. You know? Like, what what their inventory is, what's going on on their website, what their marketing efforts are, you know, what their costs are, all that stuff. So in theory, you have enough context to to to make a decision if you know where to, like, where to look and what to assimilate. But, like, you know, asking something as simple to an AI that has all that context today, asking something as simple as, like, why did my profit drop today? That is, like, a really, really difficult question to unpack, and it actually just opens up a dozen other questions that you need to go answer that have there's, like, a tree of questions that happen there that you have to follow. And, like, humans have these pathways that they built in doing their job over time that they know, and they can just go do it. But, like, trying to put the guardrails on an AI to follow that, like, logical path is is cut is really difficult, actually.

Brian: [00:35:51] Yeah. I think that's like asking AI to write a story. It's like you it's like write me a story today. Like, it's it it doesn't it doesn't really work. Uh, you know, you have to know how to treat it like it's a junior resource, to be honest. Yeah.

Kellet: [00:36:06] And that and that's kind of the approach we've started to take is, like, we're taking these heuristic maps that we have. Like, we take, like, how would a human think through this? And then we piecemeal it into, like, different way different questions that we might ask in AI in AI. And then we string all those together so that it's, like, piecemeal finding the insights out of the data, and then assimilating all those insights and then asking the question. So why did my profit drop today?

Brian: [00:36:33] Mhmm. Right. That's right. Exactly.

Phillip: [00:36:35] So, uh, if I had to bring it back to my own personal experience, which in the age of AI is literally the only thing that I have that I have going for me is my own experience. Like, everything else is can be generalized, I think, um, or replicated. So that's fun. Yeah. And now with, like, notebook LM, like, the pot like, we we're existentially doomed in PodRocket.

Kellet: [00:36:57] Right. Right.

Phillip: [00:36:58] You're like, like, everything's done. We're over. But in my experience, I my first real ecom business at scale that I worked at was in, uh, the mid aughts in 2007. I joined a a high growth, uh, startup in the supplement space, um, that was sort of riding the coattails of a overnight success. Supplements were were really exploding online on the back of the the blogosphere. You had mommy bloggers and a lot of other people that was, like, driving these affiliate businesses to 100 of millions of revenue with very little investment on behalf of those brands. So you could spin it up overnight. It was really interesting. Google PPC was another big, uh, push there. And then analytics were very, very rudimentary back then. But the owner of that business had not even a 6th sense, like a 7th sense about what was happening in the business from day to day and knew the the seasonality in and even sort of the the flow, uh, of how their business operated on an hour to hour basis. And his name was George. I'm still very good friends with him. He would send me an an instant message. Like, this is 2,007. Alright? But he'd send me a message at, like, 2 PM, and he's like, orders are slow. Something's wrong. And I I thought for years, he was just paranoid. But it almost in 99.9% of the case permission.

Kellet: [00:38:28] Yeah.

Phillip: [00:38:28] He knew. He knew. Like, he he had such a a an intuition about the the working of that business because he built everything from the ground up, and he had all the relationships with everyone. He knew that there was something wrong. It was often technology, but not always. Sometimes it would be something was misaligned. There was a deliverability issue on some email. Uh, somebody that promised that they were gonna do some post at some time didn't. Like, he had so meticulously come to a contact contextual understanding of every facet of that business that he could predict on a day to day basis what the ebb and flow from hour to hour of the business would be like and model it in his head. And maybe this if we had to get there, maybe that sort of intuition about the business can be captured by a tool today. Maybe.

Kellet: [00:39:22] I think it has to be imparted on the tool today. Like, you have to take the framework out of his brain. Like, if he had something that is he there's probably tons of data points he was assimilating to assimilating to get, like, this to get his feel, but he knew which ones really were, like, the key levers to the business or whatever were the key drivers. You could probably codify that to some degree. Maybe you can't. Maybe it's, like, too intuitive, but I would think there's, like, some heuristic formula there that could be extracted and at least make it easier for him to do and maybe buy buy someone like that more time to do other stuff. I don't know.

Phillip: [00:39:59] I I think it worked out okay for him. I think he, you know, did they exit to GNC, I think, did really well for them. But I I think there's that sense, though, is if you were gonna codify that, that happens over a longer period of time than just plug and play analytics.

Brian: [00:40:15] That's right. I agree

Phillip: [00:40:16] With that a 100%.

Brian: [00:40:17] It's almost like it needs to be there from the start. Like like, maybe maybe not. Maybe time is, like, irrelevant now. You can actually because we have such great tracking, you can actually look backwards now, uh, and feel feel it from the start. Um, but, like, there's a certain element of, like, learning from the ground up that's required to get to the point, like, to tell the full story of, like, how you got to where you are. Yep.

Kellet: [00:40:44] Yep. I agree with that.

Brian: [00:40:46] Anyway, uh, I I I think this is so cool. I I I know that, um, you know, there's the tool that's, like, blowing everyone away right now at triple whale is SONAR. Uh, we talked a little bit about it before. I'd love to finish the conversation just by Sure. Kinda nailing down, like, uh, you know, like, what what can people do with sonar? Like, how do you, like, how do you get going? What's the, like, what's the pathway to get started with sonar?

Kellet: [00:41:17] Yeah. Well, um, I I think the cool thing about Sonar is it is a plug and play, um, CAPI provider solution. So really it's, like, I think it for people in marketing, they should understand that. For people maybe marketing adjacent, they might not understand what that means, but basically, it's a it's a really easy plug and play integration with your ad platforms to impart first party data, to convey your first party data back to those platforms, that that they can then use to improve the targeting of the campaigns that you're paying to run on those platforms. It's really that simple.

Brian: [00:41:54] Incredible. And, uh, we're gonna get into this a little bit further, but, like, that level of visibility and, you know, that level of of, you know, being able to, uh, to have that first party data is is providing drastic improvements in markets where it's actually really difficult to get that data.

Phillip: [00:42:13] Let's talk about, like, what are the the desired outcomes just quickly, Kelli, is, um, I think everybody obviously wants to better spend money, um, and have better performance. But, like, what what specifically does Sonar and sort of CAPI help promise to deliver in addition to those?

Kellet: [00:42:29] Meta is where where, at least from my vantage point, we see most of the marketing budgets going. Um, uh, there there's a couple of, like, really tangible benefits to having a solution like SONAR. Now there are other solutions, but having a solution like SONAR in place, um, the first one is that Meta uses this kind of they have this grading of the they have this score, essentially, that grades the quality of the data you're feeding the algorithm.

Phillip: [00:42:59] Mhmm.

Kellet: [00:42:59] It's called an event match quality score. And so this is an upstream kind of signal of, like, how well you're informing their algorithm. And SONAR, when when people turn on SONAR, it improves those scores almost immediately. Um, and that's that's so that's Meta's, like, measure of the quality of the signal you're providing them. Um, but what are the downstream implications like the outcomes? Generally, what that means is that meta is telling you it can make better decisions on your behalf when optimizing your budget

Phillip: [00:43:30] Mhmm.

Kellet: [00:43:30] And your budget allocation across campaign. So, um, what we've seen just generally across, we did kind of, like, a sample comparison of our some of our shops that have sonar and some that don't. For the benchmark, people that didn't over the time period we measured, they saw about a 4% increase in their ROAS, so their return on their ad spend. Um, the ones that did turn on sonar, the average was about 17%. So that's just gosh. That's just, like, just out of the box, just by turning this thing on, they're already getting a better return on their ad spend. Now I can go into, like, more nuance and some other ways you can apply Sonar, but I think that's the big takeaway is, like, you you just get more effective. Meta's gonna target the outcomes that you want better, and they're gonna drive it. You spend less money to get those outcomes.

Phillip: [00:44:21] I think that's, uh, a great place to leave it. Kellett, appreciate you. Enjoy your your fall, uh, here in Raleigh.

Kellet: [00:44:29] And Hopefully, our real fall this time.

Phillip: [00:44:32] Yeah. For sure. I appreciate it. And, uh, we'll continue to, uh, explore more about what these changes and the the maturation of this, uh, this sort of new way of thinking about data within an organization and ecommerce looks like as we continue with the series step by step. Thanks for joining us.

Kellet: [00:44:50] Thanks, Philip. Thanks, Brian, for having me.

Phillip: [00:44:52] Thank

Kellet: [00:44:52] You. Really appreciate it.

‍