Ep 3 - Build, buy or rent ===
[00:00:00] Kiri: This is part three of Arming the Rebels, a Retail Media Breakfast Club series with Anthony Rotio, co-founder and co-CEO of Growth Loop, brought to you courtesy of Growth Loop.
[00:00:14] Hello, and [00:00:15] welcome to Retail Media Breakfast Club. I'm Kiri Masters, and today I am joined with the co-founder and co-CEO of Growth Loop, Anthony Rotio. Welcome back to the show, Anthony
[00:00:25] Anthony: Thanks, Kerry. It's been a pleasure to talk to you across these first two episodes. Really looking forward to today[00:00:30]
[00:00:30] Kiri: Yeah. So this, this time around, we're gonna talk about, you know, investing in technology for a retailer. They have a couple of choices. They can build it themselves with their amazing engineering team. They can [00:00:45] go out and license technology. This is a, you know, a recurring question every time there is a, a new capability that a retailer wants to add.
[00:00:54] But where we've gone with this series so far is talking about [00:01:00] what is holy, what is really special about a retailer's audience, about their value proposition, where they sit in the market. I personally believe this is more important than ever in the age of AI, where, you know, there is some sort of flattening of value [00:01:15] propositions.
[00:01:15] You've gotta really go back to what is special about you as a retailer and why customers should keep coming back to you. Um, that is, you know, m- more important than it's ever been before. And once you know what is [00:01:30] holy about your value proposition, you gotta keep it close, and my view, maybe get rid of this, rent the stuff that is not absolutely essential to that value proposition.
[00:01:42] I know that, um, that, [00:01:45] uh, goPuff is a customer of GrowthLoop, and it recently, uh, started working with Instacart for its on-site ad tech, and that's because they wanted to focus on those big [00:02:00] custom brand activations that only goPuff can do with their audience, with their delivery speed, and so forth. So this is, that's a great example to me of, like, holding onto what's holy and outsourcing the rest.[00:02:15]
[00:02:15] I'd love to hear from your perspective as a tech vendor in this space. You probably hear this all the time. "Hmm, could we build this ourselves?" And, um, let's, let's just start there. This, this must be, uh, both the [00:02:30] most painful and the most sort of exciting question you can hear when you're talking with a retailer for the first time, is exploring whether they should or shouldn't build, uh, the capabilities that GrowthLoop has available right out of the box that you've been doing for, [00:02:45] you know, six, 10 years now.
[00:02:47] Um,
[00:02:48] Anthony: I, I actually love when this question comes up. I think there are a lot of people who don't like this question. We love this question at GrowthLoop, and this has been... You know, we've been working with enterprises since day one. [00:03:00] Google was our first customer, right, as a business about a decade ago. And, um, when you work with customers that are such a build-forward and innovative, um, partners, right?
[00:03:11] It's, of course, they're gonna wanna build things. And the, I think the principle's [00:03:15] never changed is, to your point, is build the things that will be differentiated for you and a core competency for you that you want to own, right? And, um, we've always taken the approach of, great, we offer a lot of things. Um, what are the ones that you wanna, [00:03:30] you wanna build and want to be differentiated for you?
[00:03:33] And where can we help plug in and make all those things work much better together and make the sum of the parts be way bigger than the whole and, and make sure that you're owning all the learnings from the system, right? That's, that's been [00:03:45] a principle for us, is the, the GrowthLoop system won't be a black box where we're storing things and it, it's all going back into your own data layer.
[00:03:53] And I think that's, that's why we love this question so much, is there's a lot of good technology, and right now you [00:04:00] can build a lot of things really quickly, right? With- and now more than ever, and that will only, the pace will only continue to accelerate. Um, so when you do have this, um, kind of gluttony of riches of, you know, we can build anything, that, that's where [00:04:15] strategy becomes really important as an organization and picking, "Hey, we wanna invest in our data layer because we wanna make sure that we have all of our customer learnings in one place, and that consumer relationship is really important to us."
[00:04:27] Or, "Hey, we wanna own the last mile delivery," or, "Hey, [00:04:30] we wanna own the big brand activations on site because it's something only we can do as, as Gopuff," right? Um, you have to make those choices, and GrowthLoop's always here to help you if you wanna talk about some of those choices and trade-offs. But we actually really love and, and embrace that [00:04:45] question, so we, we just released something called RMN in a Box, and there's, there's a lot of things in the box.
[00:04:50] And the, the whole point is you don't have to take the whole box, right? It's open the box, see what you want, see what you don't want. Um, and that's been, that's been really helpful, I think, [00:05:00] for, for us as a partner to be able to go into these organizations that do traditionally build. I was talking to one the other day that builds everything in the media space, and they do.
[00:05:09] They are a, one of the top tier, um, one of the top tier media networks out there by every metric that's available. [00:05:15] And to go in and have them say, "Hey, we built our audience platform, but nobody wants to maintain it anymore." Um, and it's a lot to maintain, right? Is you can build things, but choosing that, what you wanna invest in for the long term to maintain, um...
[00:05:29] And now [00:05:30] they're looking at partners to pick up some of the pieces that they decided are not something that they want to be, uh, a core competency for them as a retailer, um, or as a technology
[00:05:42] Kiri: That's a, that's a really [00:05:45] interesting point actually. There is the build it and, and there must be as, as a engineer, that's something really satisfying about, like, scoping a problem, how are we gonna solve it? How are we gonna build it? Building it and, but maintaining it [00:06:00] is, is a whole other... And integrating it, I imagine.
[00:06:03] Like, these are very sort of different kind of skills and, uh, workloads
[00:06:10] Anthony: Yeah. I mean, the, there's a saying in like the manufacturing and production side, like demos [00:06:15] are easy, scale production is hard, right? The same thing kind of goes is like once you go live and get... I had a, I had someone who's, you know, leads an AI SWAT team at, at a large organization tell me, "Oh, I can gr- I could build GrowthLoop in two weeks," you know?
[00:06:27] And m- yeah, maybe you could build a demo that looks a little bit like [00:06:30] GrowthLoop in two weeks. Um, but to actually serve thousands of audiences at, you know, 2,000 requests per second on an, on an onsite experience or personalization experience, like these things are, are much harder to do long term. [00:06:45] And that's where, you know, there are some organizations that want that to be their muscle and hire up for it and do that type of thing.
[00:06:51] But, you know, invest in, invest in your customers, invest in your products and services, invest in the things that differentiate you, um, as a [00:07:00] business rather than, um, technology that maybe you can, you can rent to your point or work with a partner on, uh, and, and learn throughout the experience too. If one day you do decide to build it, don't you wanna learn along the way from folks who have done this a little bit?
[00:07:13] Kiri: Ooh, that's a [00:07:15] good, that's a good one. Um, now I, I, I desperately want to talk about speed because that is a really big part of your value prop at Growth Loop is the speed to actually activate these audiences and share back insights to brands. But we're gonna come back. [00:07:30] We g- have a whole episode dedicated to that next time so I'm holding, holding onto that, but, um, just a little preview of what's coming up next.
[00:07:39] I've... I, I have heard this a lot, and it is in different sort of parts of the val- the value [00:07:45] chain of an RMN as well. I've, I've told this story a whole bunch of times, but a retailer that was using a sort of a general purpose ad server not built for retail media, didn't understand seasonality, and so the team, to this day, [00:08:00] they couldn't get the budget to actually swap out the ad server to a purpose-built retail media ad server, and so they have to correct by hand all of the, um, the, the traffic forecasts every two weeks.
[00:08:12] Um, so there's this sort of false [00:08:15] economy, uh, uh, sometimes of, like, using technology that was maybe built for a different type of industry or a different sort of, um, for a different time, um, where behavior was a little bit [00:08:30] different. Um, what are... I- in terms of actually sort of making an upgrade maybe from a homegrown system or something that is, like, not, not quite adequate, what are some of the, um, [00:08:45] the challenges that come with, like, making an upgrade versus coming in and, and providing this capability fresh?
[00:08:55] Anthony: Yeah. Um, man, these are, these are excellent questions and timely, some of them too. I mean, [00:09:00] the-- I'll start with the pattern that you called out, which I think is one that, um, you know, fore-forecasting these systems, uh, I'll generalize that problem has been something for a long time. If you add up all of the projected revenue across all of the media [00:09:15] platforms and partners, somehow it always adds up to more than 100%, right?
[00:09:18] So one is just like own your me- own your measurement and your-- it's, and if you own your measurement, you can own your forecasting too. So centrally do that, um, on your warehouse would be what we would, we would always recommend there. [00:09:30] But I think the, the, the challenge of, uh, having a homegrown system or, um, you know, even migrating off another platform is getting out of the lift-and-shift mentality and into the what capabilities do we have now [00:09:45] that we didn't have before mentality.
[00:09:47] And instead of looking at this transition as a f- like feature-by-feature comparison, how do I do exactly the same little tweaks to the seasonality forecast that I did over here, uh, manually over here? It's like, maybe you don't wanna do them [00:10:00] manually anymore, right? Like, think about it as a what, what goals were we trying to achieve with the audiences that we had live in that platform?
[00:10:09] And then how do we rebuild them over here in a way that's really useful to, um, [00:10:15] both our teams internally and our external partners? And that's, that's one always where you're, you wanna get time to first dollar really fast. So maybe you just grab one or two of those audiences, or 30 or 40 of those audiences and say, "We'll put these on a lift-and-shift roadmap now," so you can get time [00:10:30] to first dollar really fast.
[00:10:31] But at the same time, we're doing discovery with all the, the other teams internally to say, "What's net new? What's, what if, what's been out of reach for a really long time that you wish you can do?" And this is how we get to the example that I know I've shared with you before, Kiri, of, you know, [00:10:45] Gopuff using our MCP and APIs to build 600 audiences in the first hour with Growth Loop programmatically that were, were totally net new for them.
[00:10:53] Um, because folks are looking for more options when they go to retail media partners and, um, how big is your audience [00:11:00] catalog? I think 93%, uh, would pay more or, um, oh, 93% wish they had more audiences available from their partners today, uh, that they don't have access to. So those types of things all come together with getting out of the lift-and-shift mentality and [00:11:15] speed we'll talk about in the next episode a little bit more, like you said, but
[00:11:18] Kiri: Can't wait. Um, yes. Well, bef- just before we do that, uh, I, I know we're running out of time for this one, but one thing that is getting a lot of attention right now is the data clean [00:11:30] room capability. And as you mentioned, like, actually getting those audiences and the, the data to, um, to your brand customers is an important part of, of the value prop.
[00:11:43] But there's, [00:11:45] um, you know, some challenges with actually sending personal data to a third party. There's a little bit of, uh, sort of questioning, you know, who, who, who owns my DCR at the end of the day as well, and am, am I comfortable [00:12:00] with that? So you actually have a composable clean room capability that lets retailers avoid doing that.
[00:12:08] Um, can you talk a little bit about that tech stack, um, that capability in the tech stack that... And [00:12:15] why you chose to sort of, um, sort of elevate that?
[00:12:20] Anthony: Yeah. The first question we always have for partners when, um, they talk about clean room and clean room friction is, is this a technology challenge or a logistics challenge? And nine [00:12:30] times out of 10 it ends up being, "Hey, like I have to go through vendor onboarding and all my internal processes to like get access to let other people have access to this clean room if we're gonna own the clean room," or...
[00:12:42] So that's the problem we've been trying [00:12:45] to solve with this at Growth Loop is, um, it's a, it's more of a logistics and a, a logistics challenge for the brands and for the retailers than it ends up being for a tech- a technology problem, right? So what [00:13:00] we've tried to put together here is something that makes it much easier for folks to, say, do overlap between the, the audiences that the brand already has versus what the retailer has, or, um, looks at like excluding [00:13:15] current folks that you're already targeting and making sure that the data is actually activated in a way that preserves all of the policies that both organizations have, where it's never just sending raw PII downstream, right?
[00:13:27] To your point. And you're never, you're never kind of [00:13:30] exposing that information to a third party. Um, you know, all this is, is hashed PII for the direct uploads or using a third-party graph if the, if the br- if the brand, uh, if the, if the organization uses a third-party graph. So all that stuff, [00:13:45] you just wanna factor out the technology and make it really easy for folks to get the job done that they wanna do.
[00:13:50] So that's been, that's been our approach there is let's partner with some identity partners. Let's make it extensible so that if the retailer has an identity partner and a graph they already use, we can just [00:14:00] plug in and use that. Um, because these are the-- it's, it's always more of a organizational friction and logistics friction than a technology friction is what we've found to date.
[00:14:12] Kiri: Gotcha. All right, great. Well, we'll link up in [00:14:15] the show notes to that RMN in a box concept, 'cause we touched on that a couple of times, and the report that, um, Growth Loop funded the research into, uh, what is important to media buyers as it pertains to audiences. And Anthony, I'll [00:14:30] catch you in the next episode to talk about speed
[00:14:32] Anthony: All right. Thanks, Kerry. I'll put my running shoes on. See you then
[00:14:36]