Audio === [00:00:00] ​ [00:00:00] Phil: What's up, everyone? Today, we have the pleasure of sitting down with Lourenço Melo, product marketing lead at Snowflake. [00:00:33] About Lourenço --- [00:00:33] Phil: Lourenço started his career at an enterprise telecom company based in Portugal, where he dabbled in competitive analysis, pricing, and business development. [00:00:42] He later completed his MBA at UCLA. And then spent five years at Microsoft Azure as a senior PMM. Uh, and he also focused a lot on their data business, but obviously got really deep into the BI space. And today Lourenço is product marketing lead for the solutions team at Snowflake. [00:01:00] Lourenço, super excited to chat today. [00:01:01] Thanks so much for your time. [00:01:03] Lou: Hey, Phil, so excited to be here. Thanks so much for having me. Looking forward to it. [00:01:07] Phil: For [00:01:07] ​ [00:01:07] Phil: [00:02:00] the folks that don't know, Lourenço is the mastermind behind Snowflake's modern marketing data stack report. And, uh, this is a super special episode today because we're actually timing the release of the [00:03:00] episode with the second version of the report. So we get to have a deep dive conversation with you on the, the marketing data stack of the future and how we can kind of prepare marketers for the next couple of months. [00:03:10] couple of years, [00:03:11] Understanding the Marketing Data Stack Report Methodology --- [00:03:11] Phil: I wanted to start off maybe with just talking about the methodology of the report. So the folks that aren't super familiar with it, um, being one of the most popular solutions for unifying your data into a single platform, snowflake has a really cool, unique point of view when it comes to understanding all the different tools and solutions that. [00:03:30] Your customers are using to power their marketing data stacks. And the report that you did in 2023 evaluated, uh, different tools used by customers. You had over 8, 000 customers at the time, part of the report, and to try to figure out like what are the most common fastest growing tools, uh, that, that your customers are using. [00:03:48] I know there's a ton of work involved here behind the scenes. And, uh, yeah, congrats on getting the new report out and maybe walk us through the methodology changes and the second report. [00:04:00] Yeah. [00:04:01] Lou: you know, even the mastermind word kind of leaves me uncomfortable just because there's so many people involved with this. You know, I wish I could give everybody a shout out that participates in this, but it's like truly a team effort. You think about content, partner, uh, our, our sales and industry teams, uh, Who bring all the expertise, uh, you know, so many people are involved in making this thing a success that, you know, I think there's, I think we all are the collective mastermind behind it. [00:04:27] Uh, but to answer your question about methodology, yeah, that's, I think that's what's unique about, about this report, um, is, you know, we could, we could easily, we have partners that are, are, work closely with us. Um, we could easily do a report that just surfaces those partners that are working closely with us. [00:04:43] Just based on, you know, joint selling, uh, conversions, whatever we want to really push from a subjective point of view. The reason why this one's unique is because we actually take a step back and leave our kind of subjectivity out of it. We, uh, look through our telemetry data, [00:05:00] um, and basically surface the tools and technologies that our customers are adopting on Snowflake to get their marketing use cases delivered, right? [00:05:08] So if you think about the methodology, the first thing we do is we kind of take a snapshot about of where the industry is today and try to figure out exactly what categories make the most sense and our customers are using when we look at reference architectures, right? Um, but then the interesting thing is we actually look at Depth of adoption and breadth of adoption. [00:05:31] So starting with breadth of adoption, that's relatively straightforward. It's basically looking at the number of customers that have chosen that specific platform, right? Um, that can also be, Phil, just for the audience's knowledge, it can be customers using a connected application, for example, but it could also be a stable edge. [00:05:48] So a data sharing relationship, right? So we basically look at that and say, all right, well, how broad is their adoption on Snowflake? But you could have, you know, a thousand, uh, customers with very [00:06:00] little consumption, very little usage. So we also have to look at the depth of adoption and, and really how, how sophisticated the use cases, um, are that are being adopted by the customers with that platform. [00:06:12] So basically we take breadth, depth, we create an index based on those numbers, and we stack rank them. Right? Um, and then basically what we do is we go back to that first part of the analysis of the categories and basically categorize all these partners based on their index and their categories and surface the leaders in this space from our perspective. [00:06:35] So I think the cool thing about this and really what's been so fun to be a part of is, um, is really the objectivity of the analysis. There's partners in there that that candidly aren't are very much successful clearly because the customers are using them. But we haven't done a ton of joint go to market, and it kind of opens up that relationship for us to go and do that because our customers are asking for it, right, by based on adoption. [00:06:58] So [00:07:00] that's essentially what we do in the methodology is taking a snapshot of the industry of what categories make sense, look at our telemetry data and figure out breadth and depth of adoption, index that, and then basically allocate based on those categories, the ones that are most commonly and frequently adopted by our customers. [00:07:17] Phil: One of my favorite parts from the report last year was, uh, like reading through some of the key themes that you guys called out when you kind of like sift through all the data. And then after, like, categorizing everything, um, there was like 4 main factors that, um, you called out in the report last year. [00:07:36] Key Shifts Defining Martech and AdTech Today --- [00:07:36] Phil: And I'm curious to hear. If there's kind of new ones that jumped out or if there's recurring themes in there, but the four ones that, uh, were called out last year, uh, from the data stack landscape was the convergence of AdTech and Martech. Uh, we just, uh, chatted with Aaron, your colleague at Snowflake, uh, deeply about this topic. [00:07:54] So, um, uh, I think it's episode 138 that dropped at the end of September. So folks [00:08:00] haven't checked that one out. Uh, super fun conversation with Aaron there. Uh, but there's also LLMs in Gen AI. I think that you guys. Called that really early, uh, last year, definitely a prevalent theme for us on the show as well, but data privacy and the race for a single source of truth. [00:08:16] So there was, those are the four that were kind of called out. Um, what are some of the more interesting shifts from 2023 to 2024 in the new report? [00:08:25] Lou: Yeah, I think the way we framed it last year was we were kind of discussing them and positioning these things as trends, right? And things that we're seeing in the market that as part of evolution of the market. This year, we've taken a little bit more of a prescriptive positioning of those things, not necessarily as trends, but as just seismic shifts that are not going away, right? [00:08:46] Um, ~and there's, there's kind of subsequent replay that. Yeah, Phil, I think let me pause for a second. Yeah, Phil, I think, um,~ you know, last year when we did the report, we positioned those things you just listed out as trends and things that, um, you know, are really shaping the evolution of our of our industry. [00:08:58] I think this year what [00:09:00] we've done is we just realized that these things aren't going away. They're not trends. Right. I think there's an interesting part in the report where, uh, the writer who's awesome, uh, uh, Brian McDonough, uh, compares. It's like this. This is not bell bottoms, right? They're not. They're not going to. [00:09:15] They're not going to go away. Right. And so when you think about things like privacy, when you think about things like Jenny, I, and when you think about things like data gravity, these things are not going back to the old way of doing things. They're all going to be here for the for the foreseeable future, and they're basically just Turned the industry kind of on its head and reshaped kind of a new era of marketing So I think the first difference is kind of just taking a an admission that hey These things are they're not they're not the current. [00:09:46] They're the water if that makes sense, right? And those things also have They lead to shifts and trends, right? So one of the things that we call out, for example, and maybe Aaron touched on it in her [00:10:00] episode, is around commerce media, right? And things like, you know, you have things like retail media and even travel media where, you know, traditionally you have, I would say, thin margins, right? [00:10:12] But then you're looking at this incredible amount of, of, of first party data that these organizations are sitting on, uh, as kind of the last point of contact with the consumer and figuring out ways to monetize that and really increase those margins, right? So those are things that are being enabled by this concept of data gravity and first party data and also kind of enabled by technologies and data infrastructure evolution like data clean rooms, right? [00:10:38] That allow you to share data in a privacy preserving way. So. So those are, I think we've, we've done this year is we've, we've said, all right, there's three things that define our industry, privacy, data, gravity, gen AI, and those things are leading to things like commerce media, change in how measurement is thinking. [00:10:56] We see a lot of customers coming to us and wanting to talk [00:11:00] about MMM right now. Right. Um, And we see, again, all these customers that are looking to us for, um, you know, to, to change their strategy into becoming not only an ad buyer, but also becoming an ad seller, right? Um, and so it's really interesting how these things are shaping it. [00:11:18] I will call out one thing, which is kind of this concept of data gravity, which has kind of repercussions around the MarTech and AdTech conversions that you and Aaron touched on. Um, But really, it comes down to that part of it, right? So these tools and technologies are now using the same data source to execute, whether it's like real time bidding on the ad tech side, or personalization and segmentation on the mark tech side, whether you're, uh, you know, driving personalized experiences to your existing customers or looking to acquire new ones, the data sources are the same. [00:11:54] It's the same. It's the brand's data. The second thing that's important is, [00:12:00] when you think about data gravity, what's, what's also enabled that, again, is the kind of more sophisticated, uh, data infrastructure that's available. So an example of that is, uh, you know, Snowflake's native app framework. If you think about it, we are basically saying the application now lives where the data is, as opposed to an outdated way of doing things where you have to copy that data to get it processed. [00:12:25] Right. So like identity resolution is one that comes to mind. And I remember one of our, uh, sellers, you know, he mentions this example a lot that he was working with an account that used to take him eight days to get their identity resolved. Right. And really kind of shipping it to a provider, having that, uh, processed and then back. [00:12:45] To the data source. Now that's living natively where the data sits, and it's a matter of hours, maybe even minutes to some extent. So you look at that type of efficiency, not to mention the downstream implications of not having your data copied and moved, [00:13:00] um, which again hinders privacy. Right? Um, all of that is changing around the evolution of data infrastructure. [00:13:08] Um, and so it's a really interesting time to be, uh, in the ecosystem generally, I, uh, you know, I, I love where, where I am right now in terms of seeing the, like the seat that Snowflake has to seeing how this is all taking shape. Um, and it's really, really exciting. Like all those three things, whether it's Gen AI that lives in the platform, data gravity around things like the app framework, or, you know, privacy and things like clean rooms and how that's evolving. [00:13:36] All of that is, uh, is, is being powered by our ecosystem as well as the data infrastructure layer. Um, and leading to a lot of a lot of change downstream as well. So, um, it's very exciting, I think, to be to be in this space right now. [00:13:51] The Concept of Data Gravity in Modern Data Architecture --- [00:13:51] Phil: When you said data gravity, I was trying to figure out where, where you were going with that. And in the back of my head, I was like, uh, picturing myself being [00:14:00] a fly on the wall in the boardroom when you guys were coming up with these terms of like, what do we call this new, like, breaking down data silos and like, putting everything into one place. [00:14:10] And I feel like data gravity is a really good way of centering this idea of just like, we're Breaking down all of those silos and sitting on top of the data warehouse, I guess is like a, a cool way to think about it. So yeah, really [00:14:24] Lou: Yeah, and I think it's interesting focus like it's data gravity, but it's also to some extent application gravity, right? Um, and what we think is really interesting also is that, you know, customers, they have their preference with architectures and their tooling, obviously. Um, So it doesn't necessarily have to be native applications, but what we do recommend when it's not native or directly connected to the data is if you want to have an application that has its kind of own data store, you should at least have the data sharing [00:15:00] enabled without moving that data, right? [00:15:02] Because the moment you're going back to moving data, like think about this, like it's intuitive, right? You go through this big organizational investment, almost cultural investment of unifying your entire data stack, right? Just having a single source of truth, which, by the way, is a continuous and kind of work in progress. [00:15:21] Um, it's so counterintuitive to then say, Okay, we did it, right? We're where we want to be. Now, to get any sort of A. I. Modeling done or to have any sort of, you know, campaign execution. Let's copy that data out word and execute it right. You just went through this whole investment. Why are we going back to that? [00:15:42] And so it's a it's a really kind of interesting conversation, but the applications are coming to the data and at least at the very least, removing friction and removing copies. Um, whatever the application model is that customers choose to use. [00:16:00] Yeah, [00:16:02] Phil: and. And just like moving away from, from API was definitely like a, a central trend from, from the episode that we had with Aaron and like a huge theme from even last year when I did my whole deep dive on, [00:16:14] Navigating the Trade-offs Between Packaged and Composable CDPs --- [00:16:14] Phil: on the package versus composable CDP, like it's one of the main arguments, right? [00:16:18] Like the package CDP with all the valuable like advancements that it brings you. If you also have a data warehouse, like you're, you're, you're doing a lot of the same thing, and there's a lot of overhead and you're paying twice for the same data. And it's not as fresh in one instance as in the other. So it is one of the most intuitive elements of the arguments in both architectures. [00:16:43] Lou: 100%. I totally agree with that. And I think, look, there's, there's a lot of customers we speak to that prefer the packaged approach that, that they might not have as kind of robust of a, of a data and technical team, uh, required for that. And so they, they, they prefer, or they've had [00:17:00] success with the tooling. [00:17:01] Again, what we think it makes sense is if you're going to go that route, you're going You should have that data sharing enabled. So you're not actually doing ETL back and forth between these, um, these tools. That's where you get yourself into trouble and you go back to those privacy issues and risk of having, you know, surface area with your customer data in two different places. [00:17:22] And, um, so that's kind of where we see the market going is flexibility of architecture. I guess optionality is, hey, yeah. Whatever you would like to do, we have a couple of principles we believe in. Uh, and, and they're not our principles. They're ones we believe would set you up for success as an organization. [00:17:43] And we basically accommodate the different, um, architecture you were, you are interested in most, right? So it's really interesting to, to think through that. Um, and, and I know you and Aaron, you know, went, went deep into the, the collaboration. And [00:18:00] I, you know, if you think about where we are as an industry and, you know, the kind of the kind of murky water around cookies and what's happening there, like the reality is we are in an era of signal loss, right? [00:18:13] Um, and with signal loss comes kind of this, this tension of, well, wait a second, I, a slice of, of my data pie is gone. If you go back. A few years. So I need to make that up through collaborating with my ecosystem and making sure that we're doing that within a privacy preserving way, right? And so data sharing and particularly clean rooms are ones that allow you to do that and really kind of get fill some of those gaps you might have in your first party data strategy. [00:18:46] In a way that then can lead to a ton of different value and use cases for organizations really across every industry. So interesting time for sure. [00:18:56] Phil: Uh, let's go back to those themes that you called out in the new [00:19:00] report there. Like, how would you say those themes show up in the stack itself that you've kind of put together and what changes are kind of reflected in the categories in, in this year's report? [00:19:10] Evolving Martech Categories to Reflect Industry Shifts --- [00:19:10] Phil: Like I was just, uh, bringing up. Last year's report again and looking at the categories. [00:19:14] So there was like a split between like marketing domain specific tools and then the data stack, like the foundational tools that kind of like supported and empowered a lot of those marketing tools. But on the marketing side, there was like four main categories. You had data capture. Tools like amplitude and mix panel. [00:19:34] You had enrichment tools like zoom info, you had activation, which was kind of a mixed bag of like reverse ETL census, but also like activation and in automation tools, and you also had a measurement box. Um, we'll maybe talk about like some of the changes reflected in, in the categories this year. [00:20:00] Yeah. [00:20:00] Lou: And the reality is, Data and marketing are kind of again. The data stack is converging a little bit, right? And you think about it like we see a lot of customers that are in that foundational stage of unifying their data. [00:20:12] I'll give you an example. We have a lot of customers looking for what we call campaign intelligence, which says, Hey, I want to understand. I want to have a single kind of glass pane view of how my campaigns are performing. And so you've got to integrate all your marketing and advertising data And a lot of cases that opens the door for downstream things like MMM, for example, but from a maturity curve standpoint, we just see customers today just putting on a BI tool on top of that and having kind of a view of, of how they're performing across channels. [00:20:44] But a BI tool is not a market tool, right? It's a data tool. Essentially. It serves many purposes and many use cases. Same thing with integration tools and modeling tools like dbt. We can not call those marketing tools. But DBT, for example, plays a key step, [00:21:00] a key role in the composable CDP, for example, right? [00:21:04] And some of the modeling you're doing inside the data platform. So, to answer your question explicitly, we have these two layers still. Um, uh, how does, for example, AI, right? That's something we looked at and said, hey, like, This is going to change everything and it's already, we're already seeing it with our ecosystem building applications with AI and native AI powered in Snowflake. [00:21:29] Um, it doesn't make sense to have it as a separate category. Last year we had, I think, uh, AI and ML as a category. Now what we're seeing is the likes of Snowplow and Hightouch and Growthloop and so many others building AI applications. ready solutions, right? So omitting that as a separate category and kind of, you know, taking the next step of having that be kind of prevalent throughout the stack. [00:21:58] And that's why you see that [00:22:00] AI layer above the data, uh, core, uh, was one change in how we reflected kind of gen AI and what's coming there. Um, privacy is another one, right? Like we, we, there are privacy enhancing tools specifically. But the reality is privacy is now kind of, it's bigger than a specific category. [00:22:19] Um, and you look, you look at things like, you know, even the way the applications are built on top of the data has a privacy layer to it. Um, but we have one specifically that we see, uh, really important, uh, nowadays, which is consent management, right? And, um, and how you can actually. Uh, you know, kind of honor the kind of data subject requests of your customers and, you know, Phil, you might log into a website one time or just be at a website one time and be okay with cookies the next time you decline them. [00:22:54] We have to go chase that down. If there are data copies everywhere, right? That's gonna create a [00:23:00] nightmare for consent. And so we see that as something that is very prevalent as far as the marketing and advertising tools. So the layer on top of it. You have kind of analytics and data capture. So the ones that you mentioned, uh, mixed panel and heap and snow plow and piano analytics, we have a data enrichment. [00:23:19] So that's a really interesting one because we never know what data is being used for marketing use cases. Specifically, if you're selling sunscreen in an area of, uh, of the U S or Canada, That doesn't have typically, you know, sun year round, you probably want weather data to understand inventory levels, right? [00:23:41] Um, we also have identity resolution and onboarding, right? So, so our partners like, like TransUnion and LiveRamp and, and Merkle and Axiom. Uh, we then do have last year, as you called out rightfully so, we had customer data activation, which was kind of a hodgepodge of, of a [00:24:00] couple [00:24:00] Phil: Mm hmm. Mm hmm. [00:24:01] Lou: This year we've, we've just looked at our customers architectures and said, Hey, these are in fact distinct things. [00:24:07] And so we have a CDP category, which includes, you know, a lot of the providers that, you know, operate in the composable space, but also in the package space. Um, and then we also have marketing and customer engagement as a separate category. So the likes of Braze and Message Gears and Cordial and Attentive and OptiMove and so many others, right? [00:24:30] Um, and that includes also the, you know, Zeta, for example, that has a broader capability set. And then when you move past that, it becomes, uh, Programmatic solutions. Right? So if you think about the trend that you and Aaron went deep into a martech and adtech convergence, now using that same data layer, we see a lot of of that happening on snowflake through the likes of obviously the trade desk and magnate and stack adapt. [00:24:59] These are [00:25:00] DSPs, SSPs that are utilizing kind of the brand's customer data. Um, and using data sharing. And then lastly, it's measurement, right? And here we've also included measurement and optimization to include, um, you know, tools like Optimizely and EPO, uh, and these really exciting, uh, companies that are helping, uh, you know, organizations do things like A B testing, for example, and really kind of get to that optimization, uh, layer. [00:25:28] So the stack is, again, reflective of where we see our customers going. And where we see the industry going. And for us, it became very clear that customer data activation wasn't one category. It was actually two distinct ones as we see it. Um, and so that was a lot, but I hope all of you listening do take a look at the stack and maybe you can follow along as I'm kind of walking through it a little bit. [00:25:53] But I think that that's what we're thinking through, um, for this year. [00:25:58] ​ [00:25:58] Phil: [00:26:00] [00:27:00] No, awesome. Yeah, I'm picturing folks kind of looking at the, the Snowflake data cloud and all the different categories. And maybe they're getting a bit more curious about like some of those boxes and don't recognize too many [00:28:00] logos in there. [00:28:01] Spotlighting New Tools in the Snowflake Ecosystem --- [00:28:01] Phil: And I feel like that, that spurs, you know, some of that curiosity and diving in and figuring out and discovering some of those new tools, I think that's the benefit of like still having. [00:28:11] Uh, a data cloud map, I guess, like landscape that's small enough to still see the logo names on there. Uh, Scott Brinker and France can't say the same, uh, anymore about their landscape. You need a magnifying glass to be able to make any sense of that category. But obviously you're taking a much more focused and siloed look at the landscape. [00:28:32] And it is really, um, like, uh, an interesting point of view because it's like customers that are building. On top of snowflake. So it's obviously not a look at like all the data warehouses and all the tools that are on top of data warehouses. It's it's snowflake only, and it's forward thinking companies that are building on top of that. [00:28:51] So I like the simplicity of it, but there's also like myself looking at the, the, the, the report last year, there's a bunch of names [00:29:00] that jumped out that I wasn't familiar with, that I jumped on their sites as started following folks on LinkedIn and even had some of those folks on the show. I just recently chatted with the founder of Ruckerbox, who wasn't a measurement attribution vendor that I was super familiar with. [00:29:16] So, um, yeah, I think that that's like part of the benefit there is like, yeah, I recognize some of the names, but like, damn, like, I don't know, I don't know that name and that name and I'll go check them out and see if I need an extra tool in my stack. Yeah, [00:29:32] Lou: of all, shout out to Scott Brinker. Uh, he is, if you read the report, uh, he was, uh, a huge help and we consulted with him, he is the, the godfather of the ecosystem, in my [00:29:43] Phil: opinion, [00:29:44] Lou: Myles Younger as well, uh, providing his kind of expertise on the ad tech side and those guys, you know, uh, Franz and Scott provide. [00:29:52] A landscape of the full ecosystem right there function in that report besides all [00:30:00] the trends and research that's amazing that's done in there is it's really about talking about taking a step back and getting exhaustive on the ecosystem. Ours is more around saying. What are the ones that the tools and platforms that are that are kind of connected to the snowflake ecosystem and that our customers are opting for, right? [00:30:20] So it's exciting that you're seeing some of the names like rocker box, for example, who's there again? Um, uh, you know, our customers are choosing are choosing them, right? And and they they have delivered value. And so they and many others, right? And so it's exciting to at least get that out there. That name out there, particularly positioning it as a, as a joint relationship with snowflake, um, that our customers can kind of benefit from. [00:30:48] Phil: it's, it's fun seeing for sure the, the two breakdown of like the, the foundational layer of data tools and then, um, how it kind of supports like the marketing domain specific tools. And [00:31:00] I like how you, you've got a separate category for, for activation and, and engagement this year. But it, like one thing I chatted about, um. [00:31:07] With Aaron and a couple of different folks on the show too, is like the [00:31:11] Convergence of Martech and Data Stacks --- [00:31:11] Phil: overlap between Martech and data tech or like data tooling, like I grew up in marketing in a time when marketers had to fight for engineering resources. And, um, I don't know how it was like when, when you grew up at, at Microsoft and Azure there, but like, like most engineering resources were focused on product and to accomplish any type of technical stack, you had to like get in line and it was a long line. [00:31:35] But nowadays like marketing and data teams work. Hand in hand, like at the, my most recent startup, we worked super closely with my counterparts on, on the data team. And some companies they're even like part of the same team or in growth teams. There's like one of each, like these like tiger teams, right? [00:31:51] How do you think of this distinction between the Martek stack and the data stack and what's that overlap? And do you kind of see the lines becoming [00:32:00] less and less blurry? Like in five years, when you do this report again, you're Is there even going to be a delineation between those two separations? Like, yeah, curious your [00:32:08] thoughts there. [00:32:09] Lou: it's a great question, honestly, and I don't know where it's going to go. I do think there's always going to be space for, for tools in the data stack that are, that are applicable for, for use cases beyond marketing. And then there's going to be kind of domain specific tools. So. Um, but I think that, you know, it's interesting you said that because, um, your point on, uh, kind of getting in line and, like, growing up in the marketing era where you've got to have them submit a ticket and, hey, we'll get back to you in a week, that, that doesn't even necessarily happen in the stack, uh, creation or, or kind of, uh, design phase alone. [00:32:48] Like, that happens to create campaigns and to actually execute at a fast pace. Gen A. I could could end up changing that, right? I mean, you think about how you and I as marketers can [00:33:00] now talk to our data and get insights quickly. Um, you know, or having kind of the kind of the convert, I guess the convergence of A. [00:33:11] I. And B. I. And how we can actually get insights more quickly as marketers without being so dependent on, you know, Our technical teammates, right? That's that democratization of of access to data, I think, is going to be really powerful going forward. But to your point, To get that started the right way, those two teams need to be in lockstep, right? [00:33:36] And we actually see that kind of marriage between the data team and the marketing team or the business team be a prerequisite for success. If they're not in lockstep of, hey, these are the business use cases we want to, we want to address and reverse engineer that into what data is needed. It's, it's usually a recipe for disaster if those two are not speaking the same, the same language. [00:33:59] So. [00:34:00] for anybody listening out there that's thinking about, um, kind of diving into this kind of data driven marketing space and really kind of getting really serious about building your stack the right way. First step is to get in line with your data team, very, very close to them and thinking through what's out there, what's available today, what we need to accomplish, what are the business goals and almost have this tiger team of really thinking through that together and reverse engineer that into the data stack. [00:34:29] So I think. Um, you know, your point on, on these two things kind of maybe potentially converging, I think there's always going to be space for, you know, BI isn't, is never going to be marketing specific, maybe it will, maybe there's going to be specific BI tools that are marketing only, um, but, but I think they're, they're always going to be kind of a lane between kind of broader application. [00:34:52] Of technology versus marketing domain expertise. And I think Gen AI is going to help democratize that access to data [00:35:00] in a much more kind of fluid way where you know, Phil you're running a campaign and you want to kind of just create an audience and You don't need to go to your technical team Now you can just actually use the tooling in your natural language not sequel to go and get that and iterate on that That's really exciting. [00:35:17] I think and that's one of the things that we're we're most excited about You [00:35:20] Phil: Yeah, folks listening are just like, Oh, I'm, I'm excited for that day. Like when I'm able to just like [00:35:26] Conversational Analytics and the Future of GenAI in Martech --- [00:35:26] Phil: chat with my data and have a conversation with it and like through an LLM, like that actually exists today. Like there's, there's some tools that have like functioning products that allow you to do that. [00:35:37] Yeah. Um, we had a startup on their show called narrative BI and his founder was walking us through exactly that use case. Like he's starting with Google analytics and like moving to different tools there. And I'm actually a customer and I have it set up on a, on the humans of Martek site. And I get like Kennedy's observability type of alerts, like when there's a spike and when there's like a dip in traffic. [00:35:59] And so [00:36:00] he's really trying to democratize access to data and like dashboards. Are used by like some people, but like to really get insights out of a dashboard, you need to log into your BI tool, have the dashboard up and kind of look at it or look at the reports that you got in your email. These are like more product centric driven events when there's something you need to look at and like you might care about, it's going to be in your inbox. [00:36:23] So you're going to get a notification about it and then you can like have a conversation with it, like seeing the spike, like, Hey, you got a 50 percent spike in traffic and organic yesterday. Cool. But like, Where does that mean? Like, where did it come from? How can I make that happen again next week? Like those conversations you can actually have with your data and it exists today. [00:36:44] And it's, it's still like in its early nascent, uh, but yeah, it's super exciting. [00:36:49] Lou: Yeah. And feel like, you know, if you were, if you were to ask me maybe six months ago or a year ago, like, where's Jenny? I going? I think the honest answer for all of us is we just don't know where the [00:37:00] limits are. Um, we just we just don't have line of sight into what all could change here. What we do know today is that use case of conversational analytics exists. [00:37:12] Right. And we're seeing it with some of the providers you have in the modern marketing data stack report using that today. Right. So I think for a while when, when this kind of wave of, of Gen AI revolution and excitement happened, I think we all kind of, at least my personal belief is we all kind of struggled to sink our teeth into something concrete. [00:37:34] Right. There was a lot of conversation around the creative element. But what else? Right. And I think this kind of talk to your data. piece in your own language and things like co pilots like that's where we are today. Who knows where this ends up? You know, I think, um, you know, scott had a really good point and we called it out in the report, which was, you know, all of our focus right now and kind of conversation has been around kind of [00:38:00] the provider side. [00:38:01] But what happens when there's co pilots on the consumer side, right? And how, what's gonna happen to marketing there when you essentially have a co pilot marketing to somebody who has a co pilot. Now you're just having these two sides interact with each other. You know, we haven't even scratched the surface of that debate. [00:38:19] Um, and so my honest answer is we know it's all gonna change. I, I personally believe that. I don't know how long term. I do think right now, these are the use cases we're seeing and customers asking us for and our partner ecosystem, the likes of high touch and growth loop and snowplow and others have already built applications that are doing that, um, and using some sort of A. [00:38:44] I. M. L. Component to automate some of this and make it more efficient. Effective and intuitive for people that traditionally don't have those kind of heavy technical data skills, right? [00:38:57] Phil: Yeah. [00:38:57] The Critical Role of Data Quality in AI-Driven Martech --- [00:38:57] Phil: And I'd also say that I'm also [00:39:00] excited about Jenny eyes role or just like AI and MLs role in data quality. Cause I feel like all these investments and improvements in like orchestration and, and doing fun stuff with the data is, is important and it's fun. But if you're still not having good data to like feed those things, those things really don't matter. [00:39:20] Like it doesn't matter how sophisticated your orchestration is. If you're still not working with quality data and that's where, um, yeah, like a lot of tools are focusing on how do we use AI and NML and enrichment partners to make sure that the foundation of the data that is powering those fun tools isn't a good spot. [00:39:40] And we can trust it, especially some of the, the industries that you call out in the report, like, um, FinTech and the health tech, like you don't want to make mistakes with PHI and PII data or send like. Wrong type of financial information in like an AI orchestrated world. [00:39:58] Lou: Yeah. I mean, [00:40:00] the stakes are exceptionally high there. And the reality is one thing that doesn't change with, you know, one thing that will never change is the garbage in garbage out component of this, right? Which is if your data foundation is garbage. So too will your AI. You can have a beautiful UI, you can have the right kind of prompts in there, but the insights that's going to surface are going to be directly linked to to what you have on the back end. [00:40:28] And, uh, that's why we, we believe so much in that component of data quality and really think that like the first step to getting the proper C. D. P. Downstream is really kind of getting that customer to be 60 right, properly built. But that can't be a buzzword that requires the quality data. Same thing with what I talked about before of on measurement. [00:40:48] You can run M. M. M. But if you're sitting on a poorly constructed foundation of campaign intelligence, You're probably going to be spending your advertising dollars poorly, right? [00:41:00] Um, and so I totally agree with that, and I think that's going to be one of the biggest helps, uh, or components that Gen AI is going to help, I think, the ecosystem with. [00:41:10] Why Composability has Become a Nebulous --- [00:41:10] Phil: One of the biggest buzzwords of 2024, maybe even 2023 is composability. And it's something that, um, you featured a bunch of quotes from Scott, from Scott in, in, in this year's reports. And it's something that obviously was like a big. Big central theme on the show for, for several episodes. Um, census is a sponsor of the show or verse ETL too. [00:41:34] And like, they're all lit on the composable CDP. And like one of the things that jumps out to me in, um, the, the data cloud landscape and in the first and the second report is that you're not really picking aside on the package versus the composable CDP debate that was kind of all the rage last year, and you still see it a bit, uh, this year too, but you're highlighting both packaged and composable. [00:41:55] And composable tools in, in the reports. And I know you were kind of introduced to the [00:42:00] concept of composability early in, in your snowflake journey. Uh, we'd love for you to like unpack what that means to you and where do you see this all kind of heading? [00:42:09] Lou: Yeah, I think you're right to call out that we, we don't, we don't pick sides. I think the side we pick is the one, is the one we're seeing our customers adopt. And the fact of the matter is they're still looking for both types of architectures. That's just a fact, right? Because it's not a one size fits all thing, uh, in my opinion. [00:42:27] I think the other thing that I struggle with personally, Phil, is like when Composable, Composability first was, was, was a theme, right? A couple of years ago, the evolution of that term has gotten so murky and like, how do we actually define composability? And I think Scott does a good job of, of, of painting that picture. [00:42:45] I think we, we need to, I'm not sure we've landed on a consensus, um, uh, definition of that. And you see a lot of, um, providers that. still offer [00:43:00] packaged CDPs lean into some of that composable framing but in different ways, right? Um, and I think that's become challenging as a part of this ecosystem to get very clear on how we define it, right? [00:43:14] Um, for us, the way we see the market is basically the first branch is are you connected to the data cloud or are you siloed? That's point one, right? And we kind of just look at the side that is connected to the data cloud. There's two branches that we see there. The first one is what we call modular CDPs, right? [00:43:37] And the reason why that's important to call out modular is you have a provider like HiTouch that offers the end to end capabilities of a packaged CDP, but they do it directly on the data cloud. That's a different lens than, ~say, a snowplow~ who participates in the composable CDP. market, but do one [00:44:00] part of that capability, right? [00:44:02] So we call that modular on the other side of the kind of data cloud. A. I. Data cloud connected. C. D. P. S. Is what we talked about before of that package C. D. P. With data sharing, right? Um, so, you know, brains uses data sharing. Simon Data uses data sharing. Zeta uses data sharing, and you have these, um, essentially packaged offerings with the full end to end capability. [00:44:26] But you don't have to deal with those silos and those pain of kind of ETL ing everything and bespoke integrations and things like that. So, that's how we see the market. Um, others might disagree and that's completely fair. I think it's um, I think it, again, it's, it's going back to that kind of murky water of, of where, of, of how to define it. [00:44:45] You know, you also see providers, Census, your sponsor, great partner of ours, leader in the modern marketing data stack. They are kind of repositioning themselves, right, from something broader than composable CDP, [00:45:00] right? And I think it's an interesting, uh, it's just the evolution of the ecosystem, I think, um, and how, how providers are kind of looking at their own capabilities and seeing what gaps customers need from them and then evolving their positioning and narrative accordingly. [00:45:15] So, how Snowflake sees it is one thing, how the ecosystem sees it, I think, is fragmented. But for us, ultimately, the horse we have in the race is what the customer wants to accomplish, and that's where we're gonna favor, right? Um, and we see really kind of demand on both sides of that of that narrative, to be honest with you. [00:45:35] Yeah. [00:45:39] Phil: to the report, but also building out, um, how you're going to market it with, with partners on that. Like makes, makes a ton of sense. Um, I, yeah, it makes me think of like the, the smaller up and coming players too, that like, that are like, Just start being used more and more. [00:45:54] Cause I like that you're calling out like the leaders, but also the companies to keep an eye on that. [00:46:00] They're not as prevalent in the customer base, but like you're seeing an uptick and kind of growth from the previous year. So yeah, kudos to you guys for, for getting that report out. I know, uh, the lift was, was probably a massive and, uh, Yeah, hopefully, um, we get some, some good noise, uh, from, from this episode here to learn. [00:46:18] So I got two last questions for you. [00:46:19] Lourenço’s Unconventional Journey from Data to Martech --- [00:46:19] Phil: Um, I wanted to ask you about just the journey into Martech. Like I, obviously you're super smart guy. You speak at a very, um, proficient level with like Martech, but you're like a product marketer. What's your journey? Working for a data warehouse company, um, like your journey into Martech was kind of unconventional, right? [00:46:37] Like having come from a systems and a data background, even dabbling a bit into like IOT Internet of Things before that, we'd love for you to like, just share how these experiences have helped shape your current perspective on Martech. [00:46:51] Lou: Yeah. It's a really good question. You're right. Like, I think I would I would guess that north of 80 percent of your of your guests on this podcast. Yeah. [00:47:00] Are coming more from the marketing and martech side or ad tech side while I'm coming from the pure data side, right? And I think it's interesting that we talked about like, hey Will we get to a point where these two things are now kind of one, right? [00:47:14] I think personally when we go back to that conversation we had of customers You know having to really align these two sides of the house. I think it's important to have both perspectives In an organization and how we go to market. And so my journey was like you said, telco, uh, came to from Portugal to the US to do my MBA, um, got a, you know, got a, got a role at Microsoft on the data side and like you said, IOT, so a lot of streaming and understanding what happens there with low latency use cases. [00:47:47] Uh, and then, you know, came to Snowflake thinking around this component of initially. Our positioning was really around customer 360, right? Which at the end of the day is a data foundation use case. What it unlocks [00:48:00] downstream becomes marketing, right? Um, and that's where, honestly, I've been, I've been educated by the ecosystem. [00:48:08] That's the reality, right? That's my school is the ecosystem is the people, you know, like yourself, like Scott, like miles, like Aaron Foxworthy, who you had on the show, like my friend Teammate, Luke Ambrosetti and people like that, that have guided and evolved my knowledge, you know, and I had the foundations of it through the data side, but getting that marketing understanding and talking to our partners, uh, like growth loop and high touch and message gears and snow plow and telium and others have literally gotten me to the point where I've been able to expand my knowledge around how these. [00:48:48] How this data foundation can be used downstream. So you're right, it's unconventional, but there's, there's many ways to get there, I think, right? And look, I need to, I need to continue learning. There's so much, like if I could ever get to, [00:49:00] let me put it this way. I'm not sure I'll ever get to the level of depth and knowledge that Erin Foxworthy has, period, right? [00:49:06] She's very, very smart, and she knows exactly what's going on across the ecosystem. Um, I think we have different perspectives that complement each other, which is helpful for an organization. But from a marketing standpoint, it's just, you know, there's so many different layers to this, and it's such a robust ecosystem that, um, you got to have different perspectives on it, I think. [00:49:27] Phil: Yeah, yeah, definitely important. I think that's, that's the beauty of having a job where I get to speak with a bunch of different humans that work in different, uh, slices of disciplines, all kind of still working with each other. With tech to help marketers grow a company. I think there's no shortage of, um, like philosophies and building and trying to be more efficient in, in using data and all that fun stuff. [00:49:53] So, um, makes my job super fun. And, uh, yeah, I appreciate your, your time being on the show. [00:49:57] Finding Balance as a Martech Leader and Father --- [00:49:57] Phil: You're got one last question for you to learn. So. [00:50:00] You're a product marketing leader, a father, a former woodworker, also a huge soccer fan. You've got a ton of stuff going on in your life. Uh, one question we ask everyone on the show is how do you remain happy and successful in your career? [00:50:12] And how do you find balance between all the things you're working on while staying happy? [00:50:17] Lou: Yeah, it's interesting. You kind of answered it for me the way you phrased it, because it's about balance. Right? Look, here's the thing. Having two young kids at home, uh, is fantastic. It keeps us very busy, but it has the hidden, uh, surprise of, you know, Saturdays waking up at 7 a. m. organically to watch the Premier League, right? [00:50:37] I don't have to set my alarm anymore to watch those games and to kind of feed my desire to watch, uh, international soccer. So that's one thing, you know, I, I just, I think you're, I think you've, you've hit it on the head. It's about balance and everybody's balance. I think it's different. Mine. I found it with, um, you know, spending my time, even like learning about this stuff is interesting to me and becomes.[00:51:00] [00:51:00] It can kind of fit into the work bucket, but it's also kind of for me, um, uh, my own kind of hobby in reading about this stuff. Um, there's nothing better for me than going to the beach with my kids and reading a book if I can, if I don't get interrupted every five seconds. Um, um, but really, like you said, it's about balance. [00:51:19] And I think I've been really, really fortunate and I'm super grateful to work at an organization that's allowed me to have that. Um, and you know, I have an awesome, uh, support system at home. My wife is the true hero here and, um, and, and allows me to, to focus on, on getting that balance the way that it works for me. [00:51:39] So, you know, I haven't done woodworking or metals in years, partially because like you said, I'm a father of two, but someday I want to go back to that because it is kind of, um, therapeutic for me and kind of allows me to do a little bit of. almost meditation while I do it. Um, and so those are the things that I will pursue next when [00:52:00] I find some time for it. [00:52:02] Phil: Appreciate the perspective there. I thought you were going to start off by saying like, uh, yeah, 7 a. m. It's uh, like some days we get to sleep in like 7 a. m. For me on a weekend is sleeping in with a one and a half year old. [00:52:13] Lou: it gets, it gets better, Phil. You're a little bit, you got a one and a half year old. Mine are five and three. So it gets a little bit better, you know, but the game started seven. So, you [00:52:22] Phil: perfect timing [00:52:23] Lou: earlier, but so it's kind of like, Hey, if I'm going to watch the games, I might as well be doing it with my kids. [00:52:27] So [00:52:29] Phil: Love it. Lawrence, I really appreciate your time. Congrats again on the getting the second report out and hopefully folks find a ton of value from it. But yeah, definitely found value, uh, sitting down with you here and walking us through that, thanks so much for your time. [00:52:41] Lou: Hey, Phil. Thank you so much for having me. Uh, let's do it again sometime. Appreciate it. [00:52:44] Phil: Sounds good. We'll do cheers.