Commerce Without Checkouts: Bryan House on AI, Composable Commerce, and Why Digital Commerce Is Becoming Intelligent
- 9 minutes ago
- 56 min read
Digital commerce spent the last two decades perfecting the storefront. AI may be about to make the storefront far less important.
In this episode of Signal & Noise, Brett House and Rio Longacre sit down with Bryan House, CEO of Elastic Path, to explore how AI is reshaping ecommerce, B2B commerce, product discovery and the buying experience itself.
As more consumers turn to ChatGPT and other AI assistants for product recommendations, brands need to rethink how their products are discovered. Bryan explains why rich product data, APIs and composable commerce could become essential as AI agents begin interacting directly with catalogs, pricing and inventory.
The conversation also dives into B2B commerce, where complex pricing, ERP systems and routine purchasing could make agentic commerce especially useful. At the same time, Bryan pushes back on the hype around fully autonomous shopping and explains why creating a seamless AI-powered transaction remains difficult.
Topics covered include:
AI-powered product discovery and agentic commerce
Composable commerce, APIs and product data
B2B ecommerce and automated purchasing
Commerce protocols, payments and stablecoins
Retail media and the future of ecommerce websites
How AI could dramatically simplify commerce implementation
The future of commerce may not literally be "without checkouts," but the path between I need something and I bought it is about to look very different.
Read the full transcript below.
Brett (00:01.363)
Hey everybody, welcome back to Signal and Noise. I'm Brett House, joined by my co-host Rio Longanker. And today's guest, we're thrilled, it's gonna be a fun conversation, is my brother Brian House. known him all my life, great guy out of Boston, well, west of Boston. he is the CEO of Elastic Path, which, if those that don't know who Elastic Path is, it's a commerce platform platform helping some of the world's largest brands modernize how they build digital commerce experiences.
With clients spending retail manufacturing telecom and a big focus on B2B. Did I get that right, Brian? Yeah. And brands like Orgel, Cerna Lilly, IMI, T Mobile, Johnston, there's a there's a whole bunch of really big brands that you support. And for those that don't really know the commerce media space, because we do have a lot of Martech and ad tech folks on that, you know, listening in.
Bryan House (00:36.886)
You did? Well done.
Rio Longacre (00:54.494)
it's kinda related, but yeah, but not everyone will.
Brett (00:55.727)
Yeah, it's r it's related, and we're gonna we're gonna certainly get into the AIA topics, but think companies like Shopify, not to mention your competitors, Brian. big commerce, commerce tools, plays peop you know, players like that, that certainly play in this this space. But since you know, becoming the CEO of Elastic Path, which which I didn't really fully realize this, Brian, is is a twenty-five-year-old company. I knew it was right? So you guys have you're one of the oldest sort of
commerce platforms. And it sounds to me from everything that I've read in the news and you've been talking on podcasts that you're really leading kind of a transformation, expanding sort of beyond composable commerce, which was a big topic. I heard you talk about a lot, especially yeah, on LinkedIn, to kind of more broader digital experience capabilities, and helping enterprises, you know, prepare for for, you know, AI powered commerce, which again is becoming kind of a hype term.
That we need to sort of dig into. And I think that's a large part of what we want to talk about today. Yeah. So so so Brian was also the chief commercial officer and founder of Neural Magic, which is a was a
Bryan House (01:49.006)
Absolutely.
Bryan House (01:52.814)
Great. That sounds awesome.
Rio Longacre (01:59.234)
And Brian, by the way, if you if he screws up your background you like give him a noogie at the next family dinner or something like that.
Brett (02:02.276)
Yeah, totally, totally. I want to I want people to know where he came from, right? Is and maybe I'll let you do, but Neuromagic, hardware-focused AI company that was acquired by Red Hat. so you've certainly been there. you were you were a founding member at Acquia. and you were there what for nine years? Nine and a half years, and and they had well you you you exited and and then they've they were they grew up to about 170 million while you
Bryan House (02:15.509)
That's right. Nine and a half years, yeah.
Rio Longacre (02:18.916)
Nice.
Bryan House (02:26.999)
Yeah, yeah, we were about a hundred and seventy million and then two years after I left they were acquired by Vista Equity Partners. Yeah.
Brett (02:33.444)
Vist Equity Partners, yeah. And you and you've played and then one thing that I've always taken from you, Brian, is that you've played multiple hats. And I think you told me in one of our conversations in the last sort of six months that your superpower was always sort of blending the the the sort of subject matter expertise of product leadership with this sort of commercial acumen of being like a commercial leader. And you and you made a pivot from product marketing early in your career. so we certainly want to talk about your career career trajectory as sort of a a CEO.
Super interesting. And and for those that don't know, Rhea Longacre is a diehard Yankees fan. And he's joined by the House Brothers, who are diehard Boston sports fans. And and so we might might make some jokes about the Red Sox or the Celtics. Although I did give Brian some shit because I was supporting the Knicks just because they are the heart and soul of New York City. And Brian was like, that's like blasphemy, sacrilege. Like, no.
Rio Longacre (03:09.06)
Yes, I am.
Rio Longacre (03:29.902)
Yeah.
Bryan House (03:30.263)
So much for the Lakers.
Brett (03:32.109)
So much for the Lakers. Yeah.
Rio Longacre (03:33.028)
Well, I I still Brian, I still give your brother shit 'cause there's so many teams he supports. Like he was sometimes a Steelers fan, sometimes a St. Louis like a Rams fan, I remember. And then then they moved to LA. Cardinals, right? So he he like his I'm not accusing you being a bandwagon house because like it you know, not not all the teams were good you always rooted for, but you did have a lot of teams, I'll say that.
Bryan House (03:38.283)
Yeah yeah.
Brett (03:43.442)
Cardinals.
Brett (03:54.374)
That is true.
Bryan House (03:54.784)
You know, I I think I think it is a result of us moving around a lot as kids. You know, we lived in LA when Magic joined the Lakers and they went on their early run with Kareem and James Worthy. So like I get it. But he we accused Brett of being a bandwagon jumper on her as well.
Rio Longacre (04:11.192)
Yeah, no, it's it's funny too. So I I'm a pu I'm a kind of a purist when it comes to sports. You get one team and one bac one backup once your team is eliminated that y you can't love but you can still like kinda like, right? So that that's how I feel very strongly about it. And you get to decide, you get up to around the age of twenty to decide. Like what is your team and after that
Brett (04:11.679)
yeah, I get that. I get that all the time.
Bryan House (04:24.661)
You
Bryan House (04:28.097)
Yeah. But you clearly never lived in Cleveland or, you know, one of these you know, the Washington, DC where the team sucked for decades, so you know. Fair enough. There you go.
Rio Longacre (04:38.756)
Brian, I'm a Jets fan. I mean come on, man. I've stuck with them. They're the crappiest team, the worst ownership, they make the dumbest decisions. I've I've stuck through the butt fumble. I mean, I've I had to be traumatized by Tom Brady for twenty years, right? So
Brett (04:38.876)
Well ri
Brett (04:42.906)
And he stuck with it all the time. He never jumped over to the giants.
Bryan House (04:53.293)
Jets nets Mets. Yes, exactly. That's funny. Okay, fair enough.
Brett (05:01.904)
Well so so Brian, did we cat did we did we capture your sort of career trajectory? What brought you to this point as the CEO leading well and anything that we missed?
Bryan House (05:07.051)
Yeah. Yeah. Yeah, yeah. No, I think I think you got all the high points. You know, as you said, you know, I've sort of made a a point of sort of owning different operational roles within companies. You know, I had a I had a great mentor early in my career when I was in the product marketing side and doing a lot of marketing. He was like, you know
I think he's like based on your ambitions and what you seem interested in, you should be moving towards a CEO path. And the way to get there is to is to own a sales number, particularly if you're gonna work in tech. And so out you know, we had that conversation driving from Munich to Prague after attending Oktoberfest, which was awesome, but we were hungover and all the rest. but then, you know, shortly thereafter we created a a new sales organization at Acquia.
called the global account program, global account team. And and that was the beginning of my transition into sales and had a ton of success there. And that was a really pivotal moment in my career because it's sort of, you know, one of the things that happens to all of us is we sort of get sort of on a on a railroad track and then you're just sort of going that journey and you don't it's hard to pivot out of that. and I certainly encourage people to to take more of a sort of a generalist view across the company and
Brett (06:21.682)
Yeah.
Bryan House (06:28.118)
putting yourself in different roles. It's a tremendous learning experience.
Brett (06:32.006)
Yeah, no, that's awesome.
Rio Longacre (06:32.291)
I think being general, this is maybe more important now than ever, right, Brett? I know we've talked about this a lot on different pods about how like AI, you have this like country of geniuses, right? Or sitting in a data center, you can call on for a lot of specialized knowledge, but I think being able to bring general skills, managerial skills, human skills, certainly I I think if anything maybe more important now. but Brian, we we we're thrilled to have you on. I mean, this these are we haven't gone into commerce as much, Brett, as I
Bryan House (06:35.593)
I totally agree.
Brett (06:37.297)
Yeah.
Rio Longacre (06:59.863)
think we probably would have liked to or maybe should have because it's e-commerce continues to to spike and all the changes now right if there's this big trend we talked you you mentioned earlier Brett about kind of composable commerce that was the thing I know we definitely want to dig into that Brian but maybe we can start up I'd just love to get your take on with all of the recent changes and how just discover product discovery is so different because of AI as we I mean let's face it we're in the AI era now maybe we just start there like how do you see things fundamentally changing right now?
Brett (07:02.481)
Yeah.
Rio Longacre (07:28.003)
How's it changing platforms? How's it changing commerce? How's it changing maybe even how you're going to market and and even interacting with customers and prospects? Love to maybe start there.
Bryan House (07:38.125)
Sure. Yeah, I mean, I think I agree that the commerce market is evolving. And one of the things that's been interesting in my domain, I've been at ElasticPath for six years, that, you know, if you asked me two years ago, you know, I would have told you that commerce is all exceptionally well defined patterns, right? Like we understand what a commerce experience looks like. We understand how to eliminate friction. And it sort of had gotten a little bit boring, right? You know.
For for better or worse, Shopify had sort of homogenized and commoditized most of the commerce experience, you know, and they delivered great results you know, to their to their community of sellers. but they're you know, so it was hard to innovate, right? Because at the end of the day, it simply comes down to conversion rates and you know, and time matters and friction matters and all these things to sort of puff polish off all the rough edges.
Brett (08:31.898)
Yeah.
Bryan House (08:36.108)
and then AI has fundamentally changed all this. And and suddenly now there's excitement and new things and new sort of topics in in commerce. You know, I invested in a company that was acquired by Elasticpaths, which is how I ended up here, at the beginning of sort of the composable commerce era. And that really grew out of some insights I had at Acquia. I remember talking to a customer Whole Foods years ago, and they were like,
You know, we don't want to use your front end. We love your back end technology and how you manage content and all the information and structure and all that. but front end development it moves so much faster than platforms do, right? The technologies change, the developers, you know, evolve really quickly. We want the freedom to to innovate there and use you as sort of a stable back end. and that was like this big epiphany moment for me. And I remember putting it on an executive leadership.
Rio Longacre (09:14.552)
Yep.
Bryan House (09:32.539)
topic list for an offsite because that that was like a fundamental threat to Acquia's business because we were frankly in the front end hosting business. That was our value proposition. and so yeah yeah it was it was a exactly and so and it was really at the beginning of decoupling front ends from back ends, right? And sort of the beginning of headless architectures, right? Because when they're tightly coupled
Brett (09:41.83)
Yep. A content management system, right? That's that's kind of pivoted that pivoted at least how they positioned themselves recently, but yeah.
Brett (09:53.308)
Yeah.
Bryan House (09:57.293)
Essentially to build do interesting things in the front end, you have to know the unique ways that that platform works and there's all these constraints that prevent you from doing that. and so, you know, that decoupling took time for platforms like Drupal at Acquia and for a lot of the legacy CMSs, but you saw Contentful and others come into the space and that yeah. And then that's what hap that was where this company I invested called Molten. we were doing the same in commerce. you know, but at the end of the day
Rio Longacre (10:15.363)
Yep.
Bryan House (10:26.332)
One of the things that that I realized was every commerce stack is heterogeneous. There is no single vendor that provides all the things in a commerce stack, right? You know, you know, we would look at monolithic stacks and they've got shipping providers and tax providers and search providers and multivariate testing providers and all these Martech companies, right? So they're not all coming from that one vendor. and so really
Brett (10:50.501)
Yeah. So they have like a a marketplace of of vendors that they partner with to provide other discrete services and use
Bryan House (10:55.828)
Yeah, they i i your s the stack is sufficiently complicated and you have enough things that you want to offer in it, that you need
Rio Longacre (11:03.363)
Or you're gonna product catalogue, PIM, you're gonna have payments, you're gonna have br browsing, you're gonna have all these different components and you rely on different vendors for that, I g is I guess the point, right?
Bryan House (11:07.388)
Exactly.
Bryan House (11:11.826)
Exactly. Exactly. And so so what you needed, you know, the composable, the interesting piece of it was how do you make building best of breed stacks easier, right? And it really is sort of an integration code and in and simplifying that and and frankly applying concepts of microservices and sort of independent services to a commerce stack. Right. And so that was a great thing that
This, you know, the ElasticPath is done very, very well and you know, and our customers really want to think about services that power their commerce applications. And some of them, frankly, just provide sort of an API backplane. And then other groups in their organization will build front ends for them on an as needed basis. you know, like T Mobile is a big customer of ours. Every digital experience in T Mobile, every
Every transaction, whether it happens in a store, in T Life, online, it pa it touches our commerce system on the back end. But we don't any own any of the front end experiences. And so, you know, that's a s
Brett (12:12.262)
Yeah.
Yeah. And
Rio Longacre (12:15.651)
Brian, let me let me ask you a question about that though. Just interject for a second. So, like I mean, like you've been arguing for a bunch of years. I've read in some of your things you've written and p pods you've been on. And I I I I I I agree with you by the way on this one, is that like monolithic platforms have generally been I don't know if a problem is the right way to put it, but monolithic platforms didn't innovate quickly. A lot of the things you called out, like be able to like race ahead and like s adopt some of the newer front-end technologies or there's an innovation and payments.
I think they were it was much harder to innovate with some of these monolithic platforms. There's been some really cool like companies that have that have taken off over the last few years, both in composable UX as well as composable commerce. But now that AI is becoming more of a thing, right? I mean AI agents can really interact with anything. And if everything's going headless, AI agents can interact with anything using MCP servers, like do you really need
Composable? Is it maybe monolithic? Are we back to monolithic or does it not matter? I mean, I'd love to hear your thoughts on that.
Bryan House (13:17.995)
So so interesting is the composable was grown out of a way to solve two problems. One was to decouple the front end from the back end, and two was to enable best of breed architectures. The unintended consequence of this is you've actually created the perfect foundation for AI in commerce stacks. Because in order for a a composable stack to really truly be composable,
Is it has to have exceptionally well structured APIs and everything that happens in the platform must be available via API. And so that's you know, that's exactly the platform we built. and so all of our APIs conform with the open AI API spec, they're all well documented, they're all accessible via MCP server. And so now the when AI agents are interacting for product discovery or for workflows or what have you, they're all just communicating with the API level layer altogether. And so
You know, I it would be it would be naive to say we that's what we were building for when we started building this thing six, eight years ago. But at the same time, like it is exceptionally well suited for this moment we're in right now. And I think honestly, the it's right. Honestly, some of the the monolithic and legacy technologies which have not don't have the API coverage and are not nearly as easy to use and understand by AI agents.
Rio Longacre (14:29.283)
Sometimes better be lucky than good, right? There's nothing wrong with that. Yeah.
Bryan House (14:42.239)
They're the ones that are having to do all this work to sort of modernize their platform for this moment where we didn't have it we haven't really had to do any work.
Brett (14:48.581)
Yeah, and beyond those two categories that you described, you know, sort of the decoupling and the architecture behind this, was there like is there a commercial play around microservices like easier to penetrate into a client if you're s if you're so you know, at a a potentially a cheaper price point if you're entering in at a specific use case level? They're like, We don't need the fifty five things your platform can do. We need four of them. Right? So I'm assuming there's a cool com yeah.
Bryan House (15:02.365)
Absolutely.
Bryan House (15:09.695)
Well, that a a hundred percent. So we call this the strangler pattern and and and you know, one of the things that's a a challenge in commerce and the commerce market has slowed dramatically. I had a big boom during COVID and historically before COVID there was sort of every five years there'd be some new re reason to replatform. And so there's like this ready cycle everyone to replatform. But what's happened now is the pla products have gotten so mature that organizations don't necessarily wanna wanna do
Do go through the replatforming, but that doesn't mean they don't have problems to solve, right? In a in a replatforming commerce is six months to two years and probably cost half a million to two million dollars, right? And so and when I think about our biggest competitor over the last few years, it was actually do nothing. Because they're like, you know what, we don't want to go embark on this enormous project. Like that's exactly, exactly. Right. It's it's full of risk.
Brett (15:44.048)
Yeah.
Rio Longacre (16:00.417)
Not again, right? It's gonna be expensive, it's gonna be a pain, you know. It w d don't to know that you'll do have to do it again after another couple of years, right?
Brett (16:01.925)
Yeah. Yeah.
Bryan House (16:07.463)
Exactly. So but they're like, I have a cart conversion problem. I have a product catalog problem. I have a promotions problem. And so because we have a microservices architecture where all those services operate independently, we can go be very, very narrow about the problem to solve. And what it does, it actually enables us to build credibility in a relationship with a customer. So now we can build s trust in one another, like, hey, you solved this problem really well. You've impacted our business in these ways. Now here's our next problem.
This is our next priority. What can you help us with there?
Brett (16:36.827)
Yeah. Yeah, totally. It's a classic upsell, cross sell, right? It really eliminates some of the barriers to entry from both a cost perspective but also to your point an implementation perspective, right?
Bryan House (16:45.535)
You know, but like you know, one of the things that's unique about us is as compared to some of our other players in the market is our business model allows us to sort of take a long view of customers. We care about LTV. like I'm not here to maximize your payment revenue and right, like our our take rate on the payments. I'm not here because I have to my footprint of my platform is so large that I have to start a deal at multiple hundreds of thousands of dollars, otherwise I'm losing money on the customer on day one. Right. And
These are the things we see in the market. So w we're able to start small and then grow with a customer over time and take much more long view of s of joint success.
Brett (17:21.669)
Yeah. And you've quoted that, I mean, just to to give the audience sort of an understanding of of how these how these big implementations often don't succeed. I think there's some facts behind it. I mean Forrester's had some stuff, sixty to sixty-five percent, some somewhere around there, of BDP platform implementations on the commerce side underperform, and I think you've been quoted, or at least the company's been quoted, as saying eighty to ninety-five percent, which is seems like the overwhelming majority, obviously, of BDB commerce implementations
either underperform or or fail. I mean, i i is that just because of the technical complexity and the and the lack of the decoupling that you talked about before? I mean, why is it?
Bryan House (17:57.131)
I think you know, I think I think these projects are are are complex in general and you know it's funny and I've s witnessed this and I've and we've changed our behavior results of this, you know, one of the things that happens is like you decide to purchase a new platform and then you go into a discovery process to figure out what you're gonna build. And you know, what we used to see is systems integrators will go in and effectively it's like a hopes and dreams session.
Tell us all the things that you want to do. We'll go to the whiteboard, we'll write them all off. Exactly. You know
Rio Longacre (18:28.919)
Systems integrators are great at selling a dream, right? Sell a vision, sell a dream, to have a whiteboarding session, right?
Brett (18:31.533)
Yeah, yeah. It's selling a PowerPoint. R R Rio's career for the last twenty years has been
Bryan House (18:35.498)
They're like there's nothing we can't do, right? And so suddenly what you thought was a half a million dollar implementation now is a five million dollar implementation. And that's based on a real example I had. Okay, I guess we could do that. Maybe we'll take over, you know, it's a multi phase thing and but we'll go on that journey. What has started happening in the last five years like
We're not thinking that's a journey we're not going on. We're killing this project, right? And you know, and that, you know, for for me as a software vendor selling SaaS software by and measured our business based on ARR, I am 100% in the lifetime value game. Right? That is the core, core measure for our business.
Rio Longacre (19:19.331)
Well qu qu but qu but question for for a second but question for a second. If if they're saying these projects fail, like I mean like how are they defining fail? It fail as in like it actually won't process orders or fail as in like, you know, it didn't maybe meet our expectations, which frankly could just be much or as much business or as or strategic as tech, or maybe more so. If it's not maybe have doesn't have the number of transactions, maybe your go-to-market's bad, maybe your product catalog's bad, maybe it could be a million reasons not related to tech. I w I wonder how they're defining that.
Bryan House (19:29.099)
who knows?
Bryan House (19:49.611)
Yeah, look, I you know, I am confident that that my peers in this space, their technology works. So they can process an order, they can pr present a catalog, create a product detail page, you know, all Exactly. Like so like there there's nothing along that. But I you know, I think one of the things that happens and and we're very cognizant of this, is hey, buy our new platform and you'll increase your revenue by ten percent.
Rio Longacre (20:00.713)
You don't deserve to be in business if you can't do those things, right? So
Bryan House (20:16.264)
Like or twenty percent. Like no one's buying that from the vendor, right? Like, you know, sure you can solve a problem. Like we have a customer who had a very specific cart conversion problem where their cart was failing and their conversion rates were terrible. We implemented our cart our solution solved that problem in a unique way and their conversion increased by thirty percent. That exactly. Yeah, yeah.
Brett (20:19.598)
Yeah.
Rio Longacre (20:36.611)
Right, but that yeah, that there you're solving a specific technology problem. The cart c the cart sucked, right? Which is what a terrible experience for for for the customers, but you're sorry, keep going.
Bryan House (20:44.744)
Yeah, exactly. But you know, and so so I can claim that as like here's an example of what we do, but I can't promise you, right? And I think this is a little bit of what happened in sort of the evolution of this market space is in the old days, whether it was like build your own, then it was run an on prem system, then it was moved to the cloud, then it was moved to composable. Like each one of those they could tie to growth in the overall e-commerce penetration of the market. And so it would be like, Hey, your business will grow. Well now
Brett (21:02.5)
Yep. Yep.
Bryan House (21:13.382)
E commerce is slowing as a is from a growth point of view, it's about five percent or tr or growing about five percent year over year. So no vendor can come in and say, Hey, we're gonna grow your e commerce business just because you use our technology.
Rio Longacre (21:24.803)
Well I'm sure that doesn't stop them from making bogus, like outrageous claims like that though, right? Like it you know, I mean I I have seen SaaS companies like come in with decks like this. it's gonna increase your sales or your or or this by or by some crazy numbers. Like, no, technology's not gonna do that. It's your point. E commerce is growing steadily at five percent a year, which is actually not bad growth, right? Because it compounds over a year. if you improve your if you improve the experience and you make it so people can
Brett (21:46.916)
Yeah.
Rio Longacre (21:53.827)
pri it it can process transactions and product discovery is good. then then you you you should see at least that you you should see at least that growth if you do it right. And obviously if you're solving some underlying technical problems, it'll be more than that.
Bryan House (22:02.665)
Yeah.
Yeah, yeah, right. You know, and and you know, and there's also lots of quality of life things for the people that run the commerce s system and so that way they can do more they can spend more time on the things that really add value to the business rather than sort of keeping the lights on and the trains on time, if you will. but it it
Brett (22:21.614)
Yeah. Yeah, we talk we talk a lot about that about data wrangling and the in the chain the the the that how it's really changed sort of the data analytics and data science professions where they're spending a lot less time in sort of the janitorial sort of data unification and data movement and normalization and identity resolution and they're focusing on higher value stuff because AI is managing a lot of that. You can have agents that are that are literally doing all of that. And it's changed the measurement space. It's
Bryan House (22:25.709)
Exactly.
Bryan House (22:33.917)
Exactly.
Brett (22:46.692)
But you had a really good speaking of data and the data like you you kind of talked about you had a a recent post on on LinkedIn about, you know, argued you basically you argued that teams that can spin up a ton of s of storefronts, but that doesn't really solve the problem when they need really a data model that can flex and and I think you use the John S Johnston supply exam is it John Stone or Johnston? John Stone. Yeah.
Bryan House (23:07.401)
Johnstone supply, yeah. Johnstone, yeah.
Brett (23:12.292)
where they had a million S SKUs across more than a hundred ERP systems. Did I read that correctly? I mean that that's an a you know yeah, and and you guys were able to really drive up their their AI assisted cost or order or order amount, right?
Bryan House (23:17.031)
Yeah, yeah. Yeah, yeah. They
Bryan House (23:30.131)
Yeah, yeah. So, you know, I mean, and we see this a lot, particularly in the manufacturing space, which is a big place for us, is companies grow through acquisition. And so when you acquire a company, you get what their products have, but you get their underlying ERP and PIM systems. And so, you know, organizations have different approaches with what the what do they do with that? Do we automatically migrate you to our PIM our ERP so we have one or do we leave it in place, right? You know, and there's pros and cons to both places. But you know, at Johnston we sort of actually
as like an abstraction layer for all that product information to unify it to then make it available to front-end experiences that are relevant. And so that's their online property. And then they built this really interesting mobile app on top of that as well. And that's unlocked by because they have this unified data structure now in Elasticpath, so they can build these things and build up a lot faster and take out some of the complexity because it absorbs it on the back end. and what they saw
Brett (24:22.413)
And it didn't result it did it didn't require data movement between like move it like yeah, which is kind of the old model where you would bring, you know, IBM, we were talking about this with another guest, if IBM did this in the sixties, where it's like you move data from point A to point B. we would do this at New Star, we would simplify like data that was in f you know, fifteen, twenty different data warehouses, and half the job was just bringing it into a central repository before we could do any analytics. And you're saying no, you that's all distributed and federated right now.
Bryan House (24:45.971)
Yeah.
Bryan House (24:50.449)
Yeah, yeah. And we can sort of abstract and sit on top of that.
Rio Longacre (24:50.497)
Yeah, but but
Brett (24:53.476)
Yeah.
Rio Longacre (24:53.975)
But Brian, I thought do you mention something I find very interesting there. So you know, I think rewind twenty years ago, right, or even fifteen years ago when e-commerce is becoming more mainstream. But you like a lot of the B to B use cases that were just handled by EDI back in the day, right? When you have massive product catalogs as B to B, right? For consumers, you're gonna need a front end, someplace they can look at products and explore and look at reviews. But you know, for a lot of this, heavy duty when you're ordering looking through hundreds or thousands of SKUs, mass pri massive orders, like really I
complex discounting, I'm I'm I'm guessing based on bulk discounts and complex shipping rules. That used to all be handled by EDI 'cause just it was just easier to do it, right? Is that now starting to be handled by like more e commerce? Is i that's very interesting, if if so.
Bryan House (25:35.838)
Yeah, so w certainly one of the things there's two things that we've seen in that space. Certainly there's a lot of EDI left, right? We get that. But, you know, as the the demographics of people that are making these purchase decisions and driving these purchase processes change, they're getting younger, they're more involved, they're more internet savvy, they're looking for more internet savvy tech experiences, right? And so it's very much the ki the ongoing trend of consumerization of IT. and so that's a big part of it.
and then and then two is is the as they move into sort of AI enabled pieces, they're looking for ways to to push that information out. And the problem with sort of EDI and ERP systems is they don't behave in real time, right? Like you can't be like we we have a a deal that we just won where they're
Brett (26:24.035)
Yeah.
Bryan House (26:29.971)
commerce at commerce front end on an ERP. And then if you change your color swatch, it would crash this the commerce site like because it couldn't get the call from the ERP system. Like so like they're they're organizations are trying to create sort of experiences that are aligned with their buyers and their buyers wants and needs, but still sort of meet their B2B you know sort of volume and sort of discounting requirements. And so we handle all that really, really gracefully.
From a negotiated pricing and priceless point of view and sort of providing a a product delivery, as well as sort of abstracting away some of those back ends so you can create real experiences. Because, you know, it's not necessarily like EDI is such a desktop oriented experience, but Johnstone Supply is a great example. Their buyers are sitting on their phone, they're in a truck, they're on s on a job site. They need that they need things that they can deliver, right? Or go pick up in the at the distribution center because they're on they're working a job. They need it now.
you just can't do that stuff in a sort of traditional EDI
Rio Longacre (27:30.819)
Well, by the way, I love that answer. That would I was not expecting that. That that makes so much sense that these are younger buyers now. They're gonna expect to order on their phone, order on a website. They're not there, you know, not gonna wanna mess around. A lot of EDI was done with faxes and all crazy stuff like that in order, right? So so I can see why that would but looking at you mentioned consum you mentioned like discovery and that whole thing. I think for years we've optimized commerce and just I think just digital experiences as well around the
Brett (27:37.955)
Yep.
Bryan House (27:40.498)
Exactly.
Bryan House (27:45.039)
Exactly.
Brett (27:45.774)
Yeah.
Rio Longacre (27:58.933)
a consumer journey and a customer journey, pushing people through these, you know, through these processes and through these workflows on regardless of what the what front end you're using. And I think composable commerce did revolutionize that a lot as well because you you could you could decouple the front end. But I think looking at like what AI is doing, like people are not searching or doing product discovery the same way they go into Chat GPT, they go into, you know, they're going to Claw, they're going to perplexities to these AI
E e you A AI assistance or they're going to y you know, large language models and that's where they're starting this and they're getting that's where that's where I think it starts a lot more. Thoughts on how that's impacting commerce and the whole kind of composable revolution.
Bryan House (28:41.32)
Yeah, I mean, know, there's always been this idea in the commerce space of high intent Burke buyers and then sort of browsers, right? And and you know, people used to see their the high intent buyer, like the few clicks to get through the site to put a product in a cart, go, right? And keywords drove high intent purchases, right? Keyword search. Well now, but you know, particularly again with young people and their phones, they're not starting in a browser.
Brett (28:59.405)
Yeah.
Bryan House (29:10.226)
They're starting in ChatGPT. And so the I think the high intent purchase is moving to the LLM space. And what we're what we're seeing is is, you know, the they're converting at a much higher rate because they're a motivated buyer. And ironically, they're buying long tail things, right? So in the you know, in the old days of sort of SEO and keywords, you know, you wanted to be on page one because otherwise you never get found.
Well, that's not really a a challenge in sort of with the L LMs is is they're surfacing if their attributes are done well, they're surfacing the right solution for the question that's asked, not for the keyword that's asked, right? And this is
Brett (29:37.071)
Yeah.
Brett (29:42.446)
Yeah.
Brett (29:49.731)
Yeah, that and the th and there's f there's a fact behind that that I mentioned in w in a podcast that we just recently did that thirty percent of all answers in LLMs would not appear on the first page of a Google search result. Right. And I was like I was blown away by that. I don't know like a third, it's surfacing stuff that that wasn't found before, right?
Bryan House (30:00.005)
Exactly. Exactly.
Rio Longacre (30:00.899)
Just wild. Yeah.
Bryan House (30:07.09)
Well, because, you know, the nature of keywords was it was sort of a eighty twenty rule, right? Twenty percent of the keywords delivered sort of met eighty percent of the needs. And now we don't have to make that trade off. and so, you know, one of the things that we're excited about is that we can offer effectively unlimited attributes around your products, right? 'Cause to make an LLM discovery process work well for you as a merchant.
Brett (30:15.3)
Yeah.
Bryan House (30:32.412)
You know, you in effect need like an eighty page PDF worth of information about your product so that way you can train the LLM when to recommend it. Right. And it's FAQs and it's detailed specs and it's all this stuff that's well beyond what can be grokked on a product detail page, right? Like a human's not gonna read all that information, but an it an LLM can process it in seconds. but most commerce systems don't let you sort of
store an unlimited amount of product information around that product. So we s we have that in our in our product experiences and we have this capability we call commerce extensions, which is effectively like a database as a service. So you can store any information that you want in there about products or what have you, all accessible via API. So now you're you know you're delivering the the the meat that the LLMs need to recommend your products more often. Right. So this really becomes the game is how do I flesh out
the available information about my product and get my product data models in clean and and well structured so that I can facilitate and foster these experiences and find, you know, get discovered more often by people looking for the things that I do well.
Brett (31:44.621)
Right. And if you don't get that sorry I was gonna say if you if you don't solve that problem, the data model and data infrastructure problem, this this notion of agenda commerce is just not possible. Right? yeah. I mean and and would you you know, d so what do you think what do you think agenda commerce is gonna look like? Do you think agents are gonna are gonna start operating on behalf of of either companies in the B to B world or on behalf of consumers, like a consumer facing agent that goes out and is, you know,
Rio Longacre (31:45.079)
Well you've mentioned Go ahead, Brett.
Bryan House (31:55.813)
I agree.
Brett (32:13.74)
Buying certain things on behalf. I mean, do you think that we're ready for that at this point?
Bryan House (32:18.504)
Well, I mean, clearly we're not because open AI had instant checkout and they killed it. Perplexity had buy now, they killed it. Exactly, right? So so, you know, and and a lot of people will say that they think it's a trust issue. I don't trust the LLM to do the transaction. And I'm I kind of call BS on that because you certainly call trust TikTok shop, which is the biggest algorithm driven AI oriented experience that any of us probably use.
Brett (32:20.302)
Yeah. After like five months, right? Yeah.
Brett (32:36.909)
Yeah.
Bryan House (32:45.416)
And you buy stuff on that all the damn time from like rando influencers. So, you know, it's not really a trust problem. I just think that that I think it honestly, the experiences have been poor, right? Like if you were to buy on OpenAI when they had instant checkout, you could put one item in a cart and then you could only check out one item. It's like you couldn't build a basket of items, right? And Google UCP has said, hey, you can build a basket, but now you're building baskets from multiple retailers and merchants.
Brett (32:47.597)
Yeah.
Rio Longacre (32:52.419)
What do you think it is then?
Brett (33:04.909)
Yeah.
Bryan House (33:13.98)
Well then how who how they gonna manage that on the back end, right? So I think it turns out the transaction element in commerce is much more difficult. Like you're not gonna vibe code a solution to process transactions for for sh for e commerce. And I think you know what's interesting is a bunch of people ran to it, right? Because the economics are amazing, right? Like in this is sort of the Shopify world is Shopify has proven that the business model of e commerce, one
Brett (33:25.889)
Yeah.
Bryan House (33:43.462)
really effective business model is to give away the software and and make it on the take rate on payments. Right. And there seventy five percent of their revenue is is is payment revenue. so I seventy five percent. It's all in their in their public filings. Yeah, because you you it's two and it's two point nine percent and and thirty cents on the transaction. Then if you're big enough, you can you can negotiate. Well that's what it is.
Brett (33:48.812)
Yeah.
Brett (33:53.593)
S you said seventy five percent? Seventy five percent. Wow. And it's just a take rate on on on checkout.
Rio Longacre (34:00.621)
And what's that take rate typically? Is it a couple of points? Like w how much is that?
Brett (34:09.464)
Very similar to credit cards. Yeah.
Rio Longacre (34:10.775)
You can negotiate something better.
Bryan House (34:12.687)
You can, you know, but you gotta be of a certain scale. And eighty percent of Shopify's merchants do less than a hundred thousand in a month in GMV, so they're not negotiating their payment rates. And if you don't use Shopify's payment rails, you gotta pay a penalty. You pay 16 basis points. so you it gets more expensive. but so the but the economic model is super compelling. So open AI said, Hey, we should take the transaction. Perplexity said, Hey, we should take the transaction. We'll just deliver you orders and then you can own the customer.
Ironically, Facebook has just announced it was on Bloomberg that they want to build a competitor to Shopify. Because why? Because guess what a Shopify merchant does? They go and they advertise like crazy on Facebook and then they go to Shopify.
Rio Longacre (34:52.611)
I mean, Facebook's gonna gobble up a lot of Shopify's business. There's no way that doesn't happen. I mean, so much so so many of those sales driven by Facebook. I mean, and we talked about this last week, Brett. Remember, like Facebook, like they're able to serve up such highly targeted, relevant ads. And I agree with you, Brian, 100%. It's not that people don't want to like it's not a trust issue. It's an experience issue. People will buy stuff online. If it's a good ad, if it's a good offer, if it's a good product, people whip out their credit cards. They can you know they if if they if they get
Brett (34:55.788)
Yeah.
Bryan House (34:56.067)
Exactly, right? And they're all SMBs.
Bryan House (35:13.297)
Hundred percent.
Rio Longacre (35:21.335)
Defrauded, they can just say it was they can just cancel the order, right, and get their money back. So I I think I I think you're right. I think that just the experience was crap and that's what's prevented it. that that I hadn't thought of it that way, but I think it's a great point.
Bryan House (35:25.457)
Exactly.
Bryan House (35:31.324)
You know.
And look and look, you know, there's a bunch of vendors like us who've been in this space for a long time and who've lived with the complexity, right? You know, and so we're attuned to it and understand how hard it is in a lot of these ways. And so new people come in and be like, Hey, we wanna be part of shopping. I wanna get part of the take rate. Like, you know, TikTok shop's huge. I w I wanna get into that. And like, I didn't realize it was all of this. Right? Reddit just announced they're gonna have shopping too. Like, are you kidding?
Brett (35:51.5)
Yeah and
Brett (35:56.416)
Yeah, and and they they Yeah, yeah, and they're just and they're and they're gonna end up f kind of facing some of the same friction problems through that process that that that you guys have spent decades figuring out, right? So so speaking of complexity, merchants are being asked to support a bunch of different protocols, right? And I'd love Bri Brian for you to explain some of these things. UCP, which is Universal Commerce Protocol, ACP, which is a gentic commerce protocol, and AP two, which is agent payments protocol. That
Bryan House (36:04.621)
Absolutely.
Bryan House (36:09.232)
Exactly.
Brett (36:25.281)
W i is is one protocol gonna win a across all of them and what do these actually do? What do they mean? Yeah.
Rio Longacre (36:29.667)
Does it not matter?
Bryan House (36:32.081)
Yeah, so so that's a good question. I think ultimately that winners will emerge in the protocol battles, right? Everybody wants their own view. and I think honestly Google with UCP is the best suited because you know one of the things that happened, so there are twenty-six million merchants worldwide, right? And
Brett (36:35.564)
Ha ha.
Bryan House (36:54.459)
You know, if you think if you you know, when OpenAI inst l launched instant checkout, who did they go to? They went to merchant aggregators, they went to Etsy and to Shopify and to Walmart and said, Hey, bring a million merchants to us and then we'll make them available. Because there's no way they can figure out how to onboard twenty-six million merchants one at a time, right? It would never be happen. Guess who's got a huge lead in that? Google does, because they've been doing it with Google Merchant Center and Google Shopping and Google Ads for years.
Brett (37:03.191)
Yeah.
Bryan House (37:22.437)
Right. So they have all that infrastructure built in there. So it seems to me that Google is the best position with UCP to win that that market. And you know, it's ironic because Shopify was all about open AI and and and instant checkout and their protocol for product feeds until it became clear that that wasn't working and they're like, No, no, we're U C P. We helped build that. And I was like, Yeah, yeah, you were you know, you're at all of them.
but I think the the value from a platform perspective is in merchant aggregation. How many merchants can you bring to, you know, Google, to open AI, to whatever LLM? 'Cause there's like what, ten LLMs out there in the world.
Brett (37:47.095)
So so
Brett (37:59.266)
Yeah. Yeah. And from a consumer experience that means that I can go and and you know, because they've got this massive merchant network, right? I can have a seamless checkout process be because of that network directly through MCP, right? Is that is that how from a consumer perspective?
Bryan House (38:14.786)
Yeah, yeah. So I I I think ultimately what's happening now is the the the checkout still resides with the merchant, right? Like people think that they're doing AI checkout, but they're not. They're actually they're being sent redirected to the merchant's website, whether it's a it's a like an iframe in or it's a pass through, whatever. They're not they're actually not buying with an AI agent. and
Brett (38:23.458)
Yeah.
Brett (38:32.93)
Yeah.
Brett (38:37.644)
Yeah.
Rio Longacre (38:38.359)
Right, like kinda like Rufus does with like with on Amazon, right? It'll have it'll it'll it's an AI guided agent that actually or that that it works with you. It's not honestly it's not that great, but like they claim it is actually improving average order volume and in in in some other metrics, which it which it very well might be. Yeah.
Bryan House (38:41.753)
Exactly.
Bryan House (38:53.104)
Yep. It g it leads you to the button, right? It leads you to the the buy now button and then you complete transaction and then you're off and running. you know, I think that the agency payments stuff are interesting and one of the positions that I've t taken for a while is sort of MasterCard and Visa and the payment networks.
Brett (39:01.633)
Got it.
Bryan House (39:15.02)
are gonna be the sort of gateway to sort of agents bu actually doing the buying rather than a consumer, a human in the loop kicking the button. P
Rio Longacre (39:24.003)
Do you think, Brian, that like maybe they get disintermediated? There's been a lot of talk, I don't know how real it is about, okay, like is like I know like like about okay ending that kind of oligopoly cartel, whatever want to call it, like is is will will stable coins somehow somehow emerge at a much lower take rate, right? Instead of two or three percent, we're talking like, you know, one hundred one one hundredth of that, right?
Bryan House (39:49.158)
Yeah, yeah.
Rio Longacre (39:49.284)
i are these things that you think are realistic or is it just really a pipe dream and and Visa MasterCard will continue to dominate?
Bryan House (39:56.099)
it's a great question. You know, I I think we're a long, long way off of stable coins, you know, and you know, so like let's go back to where we talked about before. Commerce was exceptionally homogenized. Every rough spot was sort of polished away so you could have the best cart conversion and the best PvP conversion and all that. Now you introduce stable coins as a payment method into this. Like, holy shit, like have you ever tried to buy a stable coin and use a a a ledger wallet or something like it?
Rio Longacre (40:23.497)
A tether or circle, one of those, yeah, it's hard. Yeah.
Bryan House (40:25.349)
my God, like, you know, like I don't have time for this, right? Like I you know, I have neither time nor patience. There's wait.
Rio Longacre (40:30.849)
Well well, I think the stable coin peop I think the crypto people underestimate the technical difficulty of using these things and like they overestimate the number of people who are willing to invest the time and have the expertise to actually actu actually set them up. It's it's it's not easy.
Bryan House (40:45.582)
No, it's a it's a pain in the ass. I mean, think you know, and you guys travel the world, you know, ten years ago, you know, if you went to Europe, they would bring out the the device to your table and you'd tap your card and you'd pay your transaction and put your ATM in. You know, in the US they you like, let me take your card and go put it in the back room and like copy it and, you know, do fraud. They still do this. Like Yeah, exactly. Like my you know, my
Brett (40:58.625)
Yep.
Yeah.
Brett (41:05.057)
Yeah.
Rio Longacre (41:06.401)
I got you with a thing, yeah.
Bryan House (41:10.32)
Canadian colleagues come down and they're like, Where are they going with my credit card? Right. Like this is still to this day. Right. So we haven't solved that problem. Suddenly stable coins are gonna r replace credit cards as the way to do payments, right? Like come on.
Brett (41:21.291)
Well, the restaurants have largely solved the problem and most restaurants have the the mo yeah. Yeah, yeah, yeah. Yeah
Bryan House (41:25.178)
Well now could everybody have toast, right? So or clover or you know, or or you know, but but like we were a decade behind Europe from a sort of privacy and security.
Brett (41:33.931)
Yeah. It was it was happening, yeah, ten, fifteen years ago. And I do remember sometimes somebody taking one of those mobile devices for checkout and wandering around trying to get a connection because he couldn't connect within the restaurant and he was walking down the street and I'm like, Where did our waiter go? And it fifteen minutes later I'm like, Do have you finished or can we leave?
Rio Longacre (41:39.301)
Well then Fred in the nineties. Like
Bryan House (41:44.198)
Yeah, yeah.
Yeah, yeah.
Rio Longacre (41:54.273)
He he came back wearing a new clothes, right? A new hat, right?
Bryan House (41:54.695)
That's right, exactly. Look at my shoes.
Brett (41:57.238)
Exactly. Yeah, he's I had to go I had to go get so they apparently had some telecommunications issues that we didn't quite have in the United States, but that's funny.
Bryan House (42:03.086)
Yes.
Bryan House (42:07.054)
So yeah, so sure, it's there. I hear lots of people talk about it. I wo I would I'm bearish on a short term impact of that.
Brett (42:16.386)
Yeah.
Rio Longacre (42:16.525)
Okay, because I mean that that would be a huge opportunity, right, to disrupt that market. But yeah, I think easier easier said than done. It's one of those markets that you I think when you look at the complexity and the and the moat that these these incumbents have, you realize why it's remained kind of an oligopoly, duopoly, cartel, whatever you want to call it for so long.
Bryan House (42:34.128)
Sure, sure. You know, the Michael Porter five forces. It's great when you have a monopoly. You can have as many forces as you want.
Brett (42:34.241)
Yeah. Yeah
Brett (42:42.485)
I remember reading those books. That was part of the the MBA curriculum for sure. so so what I didn't realize in doing research for this was that the B2B commerce or e-commerce space is heading towards it's it's almost a $36 trillion global market. Right? B2C, and B2C gets all the hype. I mean what we're talking about is is large, some of these things are largely B2C. I mean like w
Bryan House (42:46.491)
Yeah.
Bryan House (42:56.42)
Yeah, it's significantly larger than B to C.
Rio Longacre (43:05.175)
And by the that that surprises me a little bit. I didn't realise how much B to B the the B to B transactions used to be done by Fax or by EDI or some combination had actually shifted to commerce. That's actually pretty cool.
Bryan House (43:11.824)
Yeah.
Bryan House (43:15.418)
Yeah, yeah.
Brett (43:15.627)
Yeah. Yeah, and w and what do you think is driving I mean, wh why why do you think it just it kinda flies into the radar, right? Is it just is it just the the systems behind that are operating that that nobody really cares about?
Bryan House (43:25.414)
So I so I think, you know, certainly one one thing that I've noticed is, you know, we as a culture, maybe this is just the United States, but we as a culture are very focused on sort of the company success stories, the company that goes public. Everybody's talking about SpaceX and and OpenAI and Anthropic as they sort of you know, the impact on markets. you know, and Fang or whatever they call the the big seven now.
And all of the big commerce companies were all B2C companies, whether it was Demandware or Magento and Adobe or Shopify or Intershop or ATG or whatever. You know, a bunch of them went public and sort of had big splashy success stories. But what you find what we found in B2 B2B is a lot of B2B companies are like, hey, we want to go digital, we want to have digital commerce. The available platforms are all B2C platforms.
Let's try to use those. Like, crap. They they they don't they don't work for our business. And they're too rigid and too inflexible for us to customize in a way to do that. So what did they do? They all built DIY pod. Like, I'll just build my own. Right. And so there really hasn't they there hasn't been sort of a a vendor that's owned that space in the way Shopify today owns B2C. Like Shopify is the story. They're a $200 billion behemoth, right? There's no
Rio Longacre (44:23.555)
Yeah, not gonna work. Yeah.
Brett (44:34.601)
Mm-hmm. They in house it.
Bryan House (44:49.015)
Equivalent in the B2B side. So we don't have anyone to sort of absorb all of that light and energy. obviously I'd love it for it to be elastic path, but we'll see. you know, and so I think that has been interesting. And one of the things that's great for us is I love DIY places, right? Because composable commerce is a technical solution, right? You're gonna have to build stuff, right, and assemble it together. Absolutely.
Rio Longacre (45:12.003)
You don't have to replace everything they've done. You can you can you can augment it, you can build on top of it, you can you offers offer maybe better versions of some things that they've built that don't work. but looking specifically at B2B, I think that was an interesting point you met made a second ago, Brett, too. About I almost like and you've argued this. I I read a couple of things you wrote that it's a better proving or testing ground potentially for for agenda commerce, right? and then but looking at like
Brett (45:20.607)
Yeah.
Rio Longacre (45:40.065)
The reason why a lot of these companies have built their own, like done their own DIY B2B commerce platforms is because, you know, the negotiator pricing, entitlements, approvals, bulk discounts, crazy business rules that tend to go into these B2B relationships that you definitely Spotify does not have to deal with, right? So thoughts on like how you navigate that with the shift to agenda commerce?
Bryan House (45:58.778)
Yeah, yeah.
Brett (45:58.988)
Yeah.
Bryan House (46:02.489)
Yeah, so I mean, sure. So, you know, a lot of a lot of what we've been building over the last five to seven years has been really providing a foundation for that. So we built at the core of our offering is a service called Product Experience Manager or PXM. And so what we did is we we took the principles of microservices and applied it to the product catalog because the product catalog is probably the longest poll in the commerce tent.
But it's also one of the most rigid areas. Like, you know, typically B2C commerce platforms have a limit of 200 variations on an object. Well, if you're, you know, a B2B manufacturer, you might have a thousand variations or five thousand variations on an object. You can't run it there. and so by separating the catalog into separate microservices, we separated out products, pricing, and then the organization of the catalog, and then we just instantiate that in a read-only view that's called a catalog.
Brett (46:33.452)
Mm-hmm.
Bryan House (47:00.305)
so what that means is I can handle contract negotiated pricing. so if I have ten thousand customers and they all have their own price list, then I can just do that and I can just deliver a price list in real time with one set of products and lots of different options from pricing. You log in, you get the right price to the right person at the right time. you know, we you have the ability to to, you know, let's say
You know, you have a hundred products in your catalog, but you know, merch you know, customer number one gets the first thirty, customer number two gets twenty-five through seventy-five, and customer three has only access to seventy-five to a hundred. I can do that from a single catalog or single set of products and have multiple catalogs that pr pr change selection or change selection by region. and so there's all these sort of intricacies that are inherent to a catalog that historically in commerce
Would require custom off-platform development in order to meet the needs of that we can do out of the box because we sort of broke down the catalog into component microservices. So, like a good example that I talk about a lot is Serena and Lily. Serena and Lily makes custom furniture. Well, you know, when you spend you know premium dollars on custom piece of furniture and you go into a design shop, there are thousands of options to choose from. But you go onto their
Try to recreate that on a digital experience, there's a limited number of options, right? And you can't recreate that whole brand experience that you do in a design store. With our platform, we eliminated all those limitations limitations. So now you can have a bundle, a configuration for a high-end couch with 1,500 options in it. And it's all priced dynamically and and deliver that in, right? So that delivers the brand promise that people expect from Serena and Lily.
With a s you know, with an out-of-the-box configured solution that doesn't require any customization. They don't have to maintain over over years, right? And really that's been the heart of solving the unique challenges in B2B is every B2B, you know, most B2B people ha are in that scenario, right? You know, it's interesting. Shopify came out and said there's this r leaked partner memo, they're like
Brett (49:03.989)
Yeah.
Bryan House (49:10.425)
We don't want to do B2B over two hundred million in GMV that has ERP and EDI and sales people and custom pricing and all their like we don't want to do any of that because guess what? They can't do that, right? Like I mean you could, you could hack around for months, but then suddenly now that's your code. We're like that's our sweet spot. We love all that stuff. So and we can do all that out of the box. And I think that's sort of the nature of what's happened in B2B is people are like, no one's ever really built solutions for us, you know.
We've gotta solve our own scratch our own itch.
Brett (49:39.446)
Can Can you guys disintermediate the B2C s players with this type of solution because it is it can handle such complexity, right? I mean, can you does this push outside of B2B? I you know, like you have direct f you have manufacturers that are going D to C like Temu, right? That kind of kinda went by the wayside, but you know, where people are ordering direct products dropped from the manufacturer and shipped directly to them. There's huge catalogs associated with that and complexity, probably not what you just described. But yeah, are there
Bryan House (49:48.537)
You know, I mean
Sure. I mean
Bryan House (50:07.692)
Yeah. I mean, we can, like Serena and Lily's a great example. Pella Windows is another example, right? Where there's inherent product complexity in a B to C experience or direct to consumer experience. however, you know, I think the you know, the reality of the world that I operate in is I got a two hundred billion dollar gor gorilla in the B to C space. So like I you know, why why would why do I wanna fight
Brett (50:09.343)
B to see use cases?
Brett (50:29.641)
Yeah. Why try to compete there?
Bryan House (50:32.184)
Why do I want to fight on their terms of the where I can create a you know, go into another market and the I can set the terms of engagement?
Brett (50:38.847)
Yeah. Good point.
Rio Longacre (50:40.523)
Looking looking towards the future, let's say agenti commerce does become a thing. Not everyone believes it will. We've had industry analysts who I mean very famously, Brett Eric Sufer continues to call it a mirage, right? Doesn't think it's gonna happen. So and he's not the only one. There are there are many people, but let's say let's say agentic commerce does become a thing and people are both businesses and consumers, some of their transactions are happening.
Brett (50:54.56)
Yeah.
Bryan House (50:54.808)
Yeah, yeah.
Rio Longacre (51:08.139)
agentically just without them knowing and maybe they're approving things like the so let's say it gets worked out some way. Like what technologies in the current e commerce stack, or let's say maybe even enterprise IT stack become casualties, become roadkill in this scenario. Any thoughts?
Bryan House (51:26.798)
Yeah, no, it's a it's a great question. Yeah. you know, like an example of sort of agentic commerce that I think is very real and a short term possibility is you're a B to B buyer and you do reorders all the time, right? Like, you know, let's just pick a s a super dumb example. Like you're you know, you have an office and in that office they need paper and printer supplies and pens and Yeah, you know, but like very simple. Right. And s
Rio Longacre (51:48.651)
You repeat a dumb but very like very practical like scenario, right? Everyone does it, yeah.
Bryan House (51:54.435)
You know, consumables, things that get consumed. And so regularly you place reorders. well, you could certainly have a an agent that says tracks the supply and usage of those things and then understands how to automatically say, Hey, this we're ready to reorder X. We're ready to order XYZ, right? Every thirty days we're gonna do a check of levels and then determine what we need. like all of that can be handled promatic programmatically, it's in the order, goes in, maybe there's an approval step by a person or not.
know, th that's a that's an agentic commerce example that I think is gonna happen before we know it. I think from I I I think we're starting to see it. Like, you know, like an example of that is is like dog food showed up in my garage today, right? Like, you know, every I have to change like when the expected delivery date is, but it happens programmatically. Like, hey, you need dog food. and it shows up on my
Brett (52:30.687)
Yeah. But it's not happening today. P
Yeah.
Rio Longacre (52:47.459)
Well and I and I love and I love the way like Amazon solves for that. They'll say, Okay, we're gonna figure out how to just send you things and predict what you want. And you know what? If we screw it up, you can just return it, no big deal. So they they d we'll solve the return thing, like they're basically saying, don't worry about it. And and w while we tweak and work out how this agentic commerce is this these really machine learning models you're using in order to predict what you're probably gonna buy. While we work out the kinks were willing to accept some returns. It's actually a very smart way to do it.
Brett (52:59.593)
Yeah, they've removed friction for that whole process for sure.
Bryan House (53:00.206)
Yeah, yeah.
Bryan House (53:12.42)
Yep. Yep. But I think from a casualty point of view, you know, one of the things that was sort of hot at topic for a while are websites dead, right? Like if everything happens through the interface of an LLM and a chat window, do I need websites? Right. And and and I think I think the the relative importance of a site goes down. But I think you'll you know, we all have multi channel sort of experiences. Yeah, yeah, yeah.
Rio Longacre (53:38.371)
Yeah, you're still gonna need websites. I think it's just like people are gonna stop going. You'll wanna like maybe you won't go as frequently, and I agree the importance goes down. Maybe it won't be the most important place to do everything, but you'll still need them.
Brett (53:40.17)
Yeah.
Bryan House (53:45.326)
Yeah, exactly.
Brett (53:50.28)
But but L L is also driving more direct traffic, right? There's a lot less referral traffic, but there's a lot more direct to publisher A, B, C, or D, right? So not huge numbers. It's not making up for the lack of referral traffic, but but it's still there.
Bryan House (53:50.478)
Yeah, yeah.
Bryan House (53:54.303)
Exactly.
Bryan House (53:59.876)
A hundred percent.
Bryan House (54:04.235)
You know, I I also think that I think one thing that I'm I'm watching is, you know, there's like an endless array of sort of micro testing tools for conversion rates and for and for things like that. And I I wonder, you know, I wonder about the potential longevity of those, right? Because if if if I go into LLMs from a high intent purpose and then it delivers me the thing I'm looking for.
Brett (54:19.978)
Yeah.
Bryan House (54:34.529)
Right, that's gonna convert. Right. So do I really need like a multivariate A B tester, right? And am I getting enough traffic? Like so like some of that like super niche you know, that technology that emerged from like, hey, I've got an idea, we should try a thing, right? Like like I think those I think all exactly. I think all those things become probably less valuable over time. And I can speak from our own experience in terms of
Brett (54:50.655)
Yeah, multivariate testing, A B testing.
Bryan House (55:01.675)
you know, running a B to B demand gen engine that a lot of those you know, we've sort of turned off a lot of those things because they were they certainly were work harder type solutions, but I'm not sure they enabled us to work smarter and sort of we've figured out better ways to do it. But, you know, there was the
Brett (55:18.687)
Yeah. Yeah, 'cause if the people that arrive have a high propensity no matter what, then yeah, it's it's kind of like the problem with advertising. Oftentimes you're crediting channels for for incrementality or for some sort of performance. And and you know, in your model you may just not have had the right math behind what was the customer propensity to buy without being advertised to? What was the control in this experiment? And so you're overcrediting something that's not really driving a conversion event, right?
Bryan House (55:28.577)
Exactly.
Bryan House (55:38.211)
Exactly.
Rio Longacre (55:43.937)
Or or or also with go ahead.
Bryan House (55:44.29)
You know, I think one of the things that'll be interesting is is, you know, what is the impact on this as a as a greater proportion of discovery happens in AI on retail media networks, right? because, you know, in a browsing world, retail media, you know, and sponsored ads and other products become really valuable. That's valuable real estate. In a more intent driven world, which is debatable,
Brett (55:55.808)
Yeah.
Bryan House (56:11.947)
Is that is that really valuable or is that just noise that gets in the way?
Rio Longacre (56:16.195)
Yeah, it's funny. People talk about how like search is going to be decimated. And if search really were decimated by the like by let's say re LLMs and chat interfaces replacing certainly Google search, but also Amazon search and all other other types of search that would decimate retail media. You're right. But you look at Google search, revenues continue to go up. Now there are some theories I've heard about no revenue going up because they're changing kind of bidding, they're changing the way like like the floor bids are handled.
they're you know that and they're not, you know, they're actually charging higher they're actually taking higher bids now. They're not they're not doing a penny over the lowest bids. So they're actually taking they're finding short term ways to to s to suck more money out of advertisers, which may not, which might juice revenues for now, maybe aren't long term. That's another story, but the revenues continue to go for Google search and Amazon as well. I mean, it it continues to grow and then and sp and Amazon ads continues to grow, you know, month after month, quarter after quarter. So even if like at some point we start to see an impact, we haven't yet.
Brett (56:47.467)
Cho yeah.
Bryan House (57:13.251)
Sure, right. You know, but the flip side of that though is we did see, you know, All Birds is now a IT company because the the arbitrage opportunity that existed in Facebook ads went away, right? 'Cause the price went up. So so you know we ha
Rio Longacre (57:26.839)
I think they also spent way too much money in their retail stores. It was that d that no one ever went into. And expanded their product offering too much, yeah.
Bryan House (57:29.955)
Sure. Sure. But like that we saw that but the whole D to C space collapsed based on the the the the the disappearing arbitrage opportunity on Facebook ads from a CAC perspective. So so I mean who knows, right? You know, I I it's an interest it's certainly an interesting time. There's a lot more happening in commerce around with AI than there was two years ago, where this was a pretty sleepy market for sure.
Brett (57:59.465)
Yeah. Well it it sounds like an exciting journey to be on as the CEO of Elastic Path. So so I think we should move to quick hits. We've got we got back to backs today, so we're gonna we're gonna have to run. But we can certainly talk more, but we'll we'll do that in a in a part two at some point or another. so I'll start. So what's the most overrated sort of buzzword or hype theme in digital commerce right now, you think?
Rio Longacre (58:04.579)
Totally.
Do it.
Bryan House (58:09.667)
Right on.
Bryan House (58:24.471)
Probably agentic, right? you know, and it you know, and it and this idea that robots are gonna buy do all the buying for us, right? I'm gonna create an agent, it's gonna go buy all the things. Like, you know, that is definitely overhyped.
Brett (58:25.744)
Yeah, yeah. I think a git is the most overrated, yeah, or hyped
Brett (58:39.102)
Yeah.
Rio Longacre (58:40.407)
Five years from now, are consumers still visiting e commerce websites or are agents doing most of the shopping?
Bryan House (58:49.259)
I think
Bryan House (58:54.729)
LLMs are providing the mechanism to find the things that you want to buy. you know, I I think there's an open question in my mind, and this probably has to more to do with my own shop shopping behavior, is how much how much of e-commerce is really browsing versus in like, you know, hey, I I I think I need a thing, you know, for entertainment rather than for the purpose of buying. And, you know, I'm not a I'm the wrong person to ask because I like I want to go in, buy the thing and move on with my life.
Brett (59:24.724)
Typical male. We're the hunters. We're like, we're going after this thing. We're gonna go right. Right. Yeah, and I and I think with the gentic, it's the stuff that's the it's the CPG or F FMCG type of purchases that are the recurring subscribe and save. It's a perfect example. It's happening, you know, powered by AI. but you have human agency over what you subscribe and save to. There's a an incentive because you're getting a discount for that purchase, right?
Bryan House (59:24.763)
exactly, exactly. So
Rio Longacre (59:25.217)
Most guys do.
That's how I shop too, yeah. And anywhere, like the mall.
Bryan House (59:31.427)
Always
Bryan House (59:41.473)
Yeah. But
Bryan House (59:50.563)
But the the place the place the ironically, the place that I go to learn is not commerce sites. It's actually content sites that are in the commerce space. Like a good example for me is I'm a big golfer, I go to a site called My Golf Spy, which is like the consumer reports for golf. And so I learn about stuff all the time. I'm on there all the time. And then I go buy it somewhere. and so I think, you know, the question of the place I go buy, is that a website in five years? I don't know. But in the the content's still really valuable.
Brett (01:00:20.126)
Yeah, and L LMs can serve your can be that path to get to that type of content, right? so so composable, we talked a lot about composability, composable and or monolithic, right, software. Which conversation do you think the industry's still getting wrong?
Bryan House (01:00:23.828)
Exactly. Exactly.
Bryan House (01:00:35.82)
I think I d I think it's the part that people get wrong is it's not a binary choice. I think there is no such thing as a true monolith, and we talked about this. Every stack is heterogeneous. But I think one of the things that's changed is we went from monolith to pure composable and it be sort of became this religion. And now what's happening is buyers are want more from their vendors. And so they're willing to
They want the I want you to give me the capabilities. It doesn't have to be best of breed. Good enough at the right price works, but the flexibility that I could swap it out later. and so we're definitely seeing entering categories i in the sort of digital commerce ecosystem that we resisted for years because buyers want more from us. And so that doesn't make us a monolith.
Brett (01:01:21.362)
Yeah. And and they and they want lower switching costs, right? If they want to switch something out, they want to be able to rotate it out without having to go through a two year development cycle to Yeah.
Bryan House (01:01:25.036)
Yeah.
Absolutely. Right. Exactly. Like in the old monolithic days, like you could use our search or you could not search, right? You know, like I want to use Algolia or whatever third party search tool. Now they wanna be like the expectation is flipping those things in and out is easy.
Rio Longacre (01:01:45.443)
Makes sense. What matters more in AI era? Better data or better architecture?
Bryan House (01:01:52.791)
I think better data. Data i i you know, wins the day here. and I think this is this is the moment for all the people who've always been claiming like you need to get your data house in order. Like now you now you have a rebellion reason. So that's right, exactly. I knew my day would come.
Rio Longacre (01:02:05.571)
They've been vindicated finally, yeah. That makes
Brett (01:02:10.92)
Yeah, but but I think the definition of getting your data house in order, to your point about abstracted layers that can they can plug into a hundred different ERP systems. It's no longer we gotta move all the data into one data lake house and then we've gotta normalize it and all that stuff, which is how the world is largely operated by. That's kind of the snowflake ecosystems and the you know and now you're getting these abstracted layers, right, that are enabling right?
Bryan House (01:02:37.056)
Yeah, you know, I you know, I still think there's a tremendous amount of value from a data lake point of view because again the the reason data matters is there's so the depth of data is what becomes valuable in the future. And and so you have to have that. Whether you have it in one ERP or seventy five ERPs, it remains to be, you know, that's a that's a you problem kind of thing. You gotta figure that out.
Brett (01:03:03.038)
Yeah.
Bryan House (01:03:03.98)
But you still have to have the depth of data to to make it valuable to an L L from a discovery point of view. Exactly. That's a right better way to say it, yeah, for sure.
Brett (01:03:09.864)
Yeah, yeah. Whether whether it's federated or centralized, you still have to have access to it. Yeah.
Brett (01:03:18.376)
So so what's one prediction of commerce that sounds crazy today but you believe will become obvious? And I'm not even gonna say the next five years, 'cause like the next two years, 'cause things are happening so fast.
Bryan House (01:03:30.402)
So I think actually I would say that implementation of a commerce platform is becomes goes it goes from an implementation project to an onboarding task. That actually is the crazy thing, right? I think you know we gave I gave the my company a sort of a B HAG goal to chase, which is six weeks and a hundred thousand dollars for an implementation.
Brett (01:03:43.807)
Yeah.
Bryan House (01:03:55.518)
without sacrificing our value proposition. You know, the easy way to to reduce the cost of an implementation is to cut all the things that are hard to do. and so then you're like, hey, well, we neutered the product, but it was easy to implement. Like that doesn't solve the problem. and I and I think that's because the patterns in e-commerce are exceptionally well understood, and AI can is is great at pattern recognition and pr pattern execution. And so
Whether it be front end development, back end integrations, data management and migration, all those things can be handled exceptionally well with AI and that is where the world's going. And I think that's a huge thr
Brett (01:04:32.125)
Yep. Has it been successful? Has has the team fulfilled on that request by the CEO?
Bryan House (01:04:36.822)
We're we're well down that path and and it's been great. And I think the I think what it does is it putish put puts the traditional SI ecosystem at risk because like hey, I don't want to spend two million dollars and take two years to do this.
Rio Longacre (01:04:50.487)
Yeah. Well, yeah, the the the AI model is under pressure and it should be. I mean, the you know, the relying on big offshore teams, taking forever to do things, overstaffing them. I mean, blend it's and a lot of the things they were building for, frankly, could be done by AI much better, right? When you look at a lot of the roles in a team. So I think it's a good thing if if that happens. you know, there there'll still be a huge need for systems integrators because it like these the the specialized technology n
Bryan House (01:04:55.596)
Yeah, exactly.
Bryan House (01:05:07.925)
Exactly. Ex
Bryan House (01:05:15.391)
yeah.
Rio Longacre (01:05:18.893)
Like the knowledge of the specialized technologies and then the scale delivery for sure, but it's gonna be a very different model. So I think the sooner we get to that, the better. No, there's gonna be more need, but the thing is, but like the comp but like no, but I think no but I think that the the legacy ones, Brett, they're in big trouble.
Brett (01:05:24.711)
Rio, are you shilling are you shilling being being a systems integrator? No.
Bryan House (01:05:26.955)
Yeah.
Brett (01:05:32.862)
Yeah.
Yeah, no to navigate a lot of these things does require yeah.
Bryan House (01:05:36.31)
Yeah. Well, I mean suffice to say the the where you add value changes. And that in in in you know, and and where every action happens in a market, there's gonna be an equivalent reaction. So that's what I think is gonna change in this space and I think that's great.
Rio Longacre (01:05:52.459)
Yeah. Yeah. And it's funny too, like you look at like the like the forward deployed engineer is a consultant. It's a a systems integrator, right? But it's like someone who's so so ever people are investing more in services now, which is interesting because it's not this dirty word within tech anymore, right? Because you know, the margins on forward deployed engineers are actually more similar to what we're seeing in AI than than the SaaS business was. but it's certainly gonna be I d I think a lot of the incumbents will probably
Bryan House (01:06:06.273)
That's right.
Rio Longacre (01:06:18.209)
have to radically restructure or th or or they'll disappear. There'll be a lot of new people, like you know, what like the Palantir's business, for example. So
Bryan House (01:06:24.275)
Exactly. Well, and I think the reason it matters from my perspective is the current pain gain formula of switching platforms is out of whack, right? You're like I'll just stay with what I'm doing. I live with the devil I know. and and so fundamentally I'm trying to change that pain gain formula is like, Hey, you fundament you change T C O in a way that makes sense for us to act. That's that's I think the super interesting part for me.
Brett (01:06:36.775)
Yeah.
Rio Longacre (01:06:47.649)
Yeah, if if the switching cost is quick, the pain the pain threshold the pain is low, right? then then then why not, right? So that makes that's a good point. This is great.
Bryan House (01:06:52.522)
Exactly.
Brett (01:06:55.837)
Yeah. Well this was awesome. Thanks, Brian, for the time today. and go Red Sox, right? We're half game behind half game behind the Yankees. Both deeply flawed teams, neither of which are gonna beat the Dodgers, I don't think, in the World Series, if if that's possible. But
Rio Longacre (01:07:03.041)
It's a close race now. What what are they you ha how many game half game behind? Yeah, yeah. Yeah. yeah, it's gonna be a bloodbath, whoever goes against the Dodgers this year, but
Bryan House (01:07:14.763)
I I'm sorry, but the Red Sox just swept the Dodgers in LA. So yeah, I I don't think necessarily overconfidence.
Brett (01:07:17.701)
Yeah, we did, we did, yeah. Yeah. Br yeah.
Rio Longacre (01:07:19.949)
Well, sometimes being actually sometimes being hot at the right time is more is the most important thing. So you never know. Yeah.
Brett (01:07:24.263)
Yeah. Yeah. And and be being hot at the right time and being hot for a long time, which proves that it's not just a flash in the pan, which I think the Red Sox are sort of doing right now. So anywho, we'll end with that. Thanks, Brian. It was awesome. everybody that made it this far in this episode, visit us at wdebdeb dot signalanoise.ai and we are all over all the platforms that you short and long form, listen to content or watch video. so we'll see you next time. Thanks.
Bryan House (01:07:24.766)
Absolutely.
Bryan House (01:07:51.828)
Awesome. Thanks for having me.
Comments