← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
EXTE: Contextual Targeting, Optimization, and Expanding into the U.S. Market
00:19:51
Loading summary
Transcript47 lines
- [00:01]
Ari
This podcast is brought to you by audiohook, the leading independent audio dsp. Audiohook has direct publisher integrations into all major podcast and streaming radio platforms, providing 40% more inventory than what could be accessed in omnichannel DSPs. What's more, audiobook has full transcripts on more than 90% of all podcast inventory, enabling advanced contextual targeting and brand suitability. Audio Hook is so confident that in addition to CPM buys, they offer the industry's only pay for performance option where brands can scale audio and podcasting with peace of mind. Mind knowing they are only paying for outcomes. Visit audiohook.com to learn more. That's audiohook.com welcome to Marketecture, where you can get smart fast with in depth interviews of leading technology executives. I'm joined today by Matthew Dougherty, who is the CEO of North America for xt. Matthew, thanks for being here.
- [01:04]
Matthew Dougherty
Thanks for having me. Ari. Psyched to be on.
- [01:06]
Ari
So I was saying before we started recording that I was excited about this because I don't know very much about xd, so I'm going to learn too. So XD is based in Madrid. Tell me about the company and what it does and the history here.
- [01:20]
Matthew Dougherty
Yeah, yeah. You're not alone in being unfamiliar with the name. It's actually the name itself is new as of January of 2024. So we actually went through a rebrand at the beginning of last year and the reason for doing so is because we actually had just completed a round of M and A activity in the year prior and it sort of made sense once we had integrated these different businesses and assets together under one roof to rebrand and rename. So something, you know, something new.
- [01:50]
Ari
What was it called before?
- [01:52]
Matthew Dougherty
So the most prominent of the three companies that we acquired and merged to form XT were. And they all kind of occupied sort of a unique, a unique space within the digital ad tech landscape or ecosystem, if you will. I'll give you a little bit of history on each of them just because I think it'll help contextualize the offer.
- [02:07]
Ari
Sure, sure.
- [02:08]
Matthew Dougherty
The largest of the three, which was based in Madrid, was a company called Sun Media. They were very much on the sort of demand side, focused side of the business and think of them as like a DSP ad network. Video and display did a lot of like rich media, high impact. So, you know, had a pretty prominent presence over there in Europe and really focused on like the iberia region, the EU5, and then a lot of Latam. Significant Latam presence as well.
- [02:40]
Ari
Right.
- [02:41]
Matthew Dougherty
Another One of the companies that we acquired was a company called Rich Audience which was one of the largest SSPs in Europe. They on the sort of flip side of what Sun Media did, they were very much focused on direct publisher relationships, publisher enablement, yield management on behalf of the supply side. Of course they had some solutions of some solutions as well that were really interesting to us. They had some contextual technology that we found to be very interesting that they had been developing over the course of their 10 year history. Then you also had a third company that was based in Barcelona called Adpone. And Adpone was also sort of publisher focused, had a bunch of different publisher enablement tools as well as also had a contextual technology and contextual capability as well as a creative studio that kind of went alongside or married up pretty nicely with the creative studio that Sun Media offered. So when we were looking at those three assets in totality, there was an opportunity to kind of create and build a marketplace that we identified. Right. You've got demand side tools with the Sun Media side of the business, you've got direct access integrations with the publishers via of the SSP Rich Audience and then you've got these contextual tools, significant like patented contextual tools as well as a pretty sophisticated creative studio for custom post production creative optimization. And so it made sense to kind of bring all that together under one roof. And once we did all of that work, which was pretty much the entire calendar year of 2023, which fortunately predated me because I think that was probably a tough year.
- [04:17]
Ari
Smart to join after that.
- [04:18]
Matthew Dougherty
Yeah, you know, we had an offering that was now no longer just kind of a series of point solutions, but more of an end to end platform that we could take to market. And that was when we decided to rebrand to XD as well as make some strategic decisions to move into what we would consider expansion markets, North America being the biggest of them. And that's when I came on board.
- [04:37]
Ari
Okay, and how big is the company overall? Like round numbers of headcount?
- [04:41]
Matthew Dougherty
We're about just over 400 employees. Employees and in 32 countries I believe.
- [04:49]
Ari
So what is the basic offering at this point since you do a little bit of everything, but what do people call you and ask for?
- [04:55]
Matthew Dougherty
It's funny because I think there's different components of the offering overall. When I am asked to give a very short description of what we do, I consider us really like an outcome based media optimization platform and we use a kind of series of different tools and mechanisms that are proprietary to us, mostly driven by agentic AI that we've developed over the course of several years to drive just improved media performance across the marketing funnel, both behalf of the brand side, like the marketer side as well as the publisher side. So you know, the word curation gets thrown around a lot these days. I think everyone could kind of like fit into that bucket. Curation in my estimation is not anything new. Like I mean we've been doing, you know, supply decision, supply decision data and data applied decisioning on the supply side for years. But I think that might be like an easy ish way to, if you had to put us in a box. But the reality is there's, there's a lot of different tools that our clients lean into, some more than others that you know, are part of the offering. You don't need to use all of the components of the offering to ultimately, you know, take a full advantage of what we do. Right. There's some flip clients that use this just for supply. There's some clients that use this for supply plus the creative optimization. There's some clients that use this for supply plus contextual the main reason or the secret sauce, what clients are using most clients are using us for most often is this agentic optimization logic that we apply to every impression that we serve that's outcome oriented against media KPIs to help deliver better performance against those KPIs.
- [06:32]
Ari
Okay, so let's break it down a little bit. So on the buy side, are people, are customers generally using your sort of DSP and login system or are they using a different DSP plus your agentic AI?
- [06:45]
Matthew Dougherty
Yeah, that's a good, great question. And actually it's kind of an interesting pivot that we made because you know, there's a lot of learning entering a new market, especially as a sort of not known entity when we entered the marketplace. So we, I think to our credit were willing to have an open mind about what was going to make the most sense and fit most neatly into sort of the existing workflow and infrastructure of the ecosystem. And so we do, while we own and operate our own ssp, what we were finding was that when we were working with clients, a lot of bigger clients on the demand side or the marketer side had just finished up these like multi year, what they would call, you know, SPO exercises or consolidation exercises where, you know, going through the process of getting another credited SSP integrated into, you know, whatever their existing stack might be, could be a multi year process. Right. So while we own and operate an SSP that was already directly integrated into many of the big DSPs that are out there. The challenge is more so like working within and seamlessly fitting within the workflows of our bigger demand side clients so that, such that they didn't have to go through another like vetting cycle and vetting process. That could take years. So what we ultimately ended up doing was integrating a lot of our tech, like you know, pre, like the curation technology, pre bid a lot of our pre existing inventory into some of those major SSPs that had already been vetted and essentially using those guys as, you know, pipes, for lack of a better term.
- [08:16]
Ari
Okay, so you're creating, you're using your data and your contextual and your AI to create deals in the major SSPs and then the customers are using whatever DSP they want. Is that the typical.
- [08:28]
Matthew Dougherty
And they could connect directly in RSSP too if they were. So, you know, if they felt like they wanted to do that. But yes.
- [08:34]
Ari
And so when we talk about the data and what's proprietary there. Tell me about the contextual engine and the AI and what makes it different.
- [08:41]
Matthew Dougherty
Yeah, what's really cool about the contextual tech is there it's sort of a multi stage analysis that we do of every new domain that we see that gets published on a daily basis. On average about 20 million unique URLs that we see on a daily basis. And what we do is, as I said, we go through a sort of multi stage analysis, contextual analysis of these different pages. The first step of the analysis is sort of we look at and scan all embedded media that would exist on that page. So video, any video, any existing imagery, any embedded audio, any existing sort of media elements that would exist on that page. And we scan and we look at it and analyze it for contextual signals or clues that might inform step two of the process, which is obviously the sort of semantic or editorial analysis of that page. We've used, you know, this technology, the contextual technology that we've built and that we've developed over the course of, As I mentioned, 10 or so years now, you know, before, back in 2015, before contextual was like everybody's the apple of everybody's eye, like it is today. We started training natural language processing and large language models to basically understand the sort of syntax between words at a more, let's call it, fundamental level than what we've seen. Traditional contextual technologies that are more like topic based.
- [10:02]
Ari
Right, More semantic technology.
- [10:04]
Matthew Dougherty
Semantic, exactly. So we look at the actual relationship between the words on the page, the syntax we're able to unearth sentiment tonality. We could also obviously understand the differences between words that could have different meanings in different contexts. You think of the word fly, it could be a mosquito, it could be an airplane, it could be the fly on zipper on your jeans, right? And so you need to be able to understand the sort of surrounding environment to be able to make the most accurate contextual classification possible. And so we'll then marry up sort of the semantic analysis that we've done, that's again syntax based with the media analysis we've done on the page. And we'll arrive at essentially a, you know, a contextual classification of that page that we feel strongly is, you know, as accurate as it could possibly be. And we actually ascribe a confidence level or confidence score to that contextual categorization so that our marketers know that are working with us and may want to deploy this type of contextual targeting on a given campaign. Hey, we're 85% sure, based on the resonance between the embedded media that we've scanned in the editorial that we've analyzed as well as the editorial itself, that this is an article about sports or music or whatever it might be.
- [11:17]
Ari
And how do you give control to the marketer? What are their inputs into this process?
- [11:23]
Matthew Dougherty
So we basically have in terms of a confidence score, like we kind of have our own standard there. Anything that's below 80%, we just, it's not non targetable entity for us. Meaning if we're less than 80% confident that the contextual classification we've given a certain page could be inaccurate in any way. It's not even target, it's not even a targetable entity. But what we do give sort of control over on behalf of our clients is the ability to do fully customized contextual segmentation. Right? So when we're doing that multi stage analysis of these different various 20 million plus URLs or domains that we're scanning every day, what we are doing is we're ultimately arriving at a, you know, IAB standard taxonomy sort of segment that we're classifying the page to. So in the example I gave, gave before would be sports. But we're also, while we're doing that analysis, picking up on any sub entities that may exist and we're caching those sub entities that we've picked up along the way during that analysis. So maybe it's an article about sports, but we've identified that it's, you know, it's about baseball and the Yankees, we've identified the Yankees and the All Star Game and baseball and Yankee Stadium. And so all of these different characteristics are then cashed with that page with that analysis. So if we have a client come to us next week and say we want to run a campaign that's targeted against any page that has contextual relevance to the All Star Game, for example, we've already made that sort of that identification and we're now able to offer up that page as one that would be contextual and the targetable entity.
- [13:02]
Ari
And you mentioned there one company that was merged in here was more like the rich creative side of things. How does that product continue to exist?
- [13:09]
Matthew Dougherty
Yeah, that. So that product is kind of the idea behind that. That product. I think creative in general is a little bit commoditized these days. And when I say creative I mean like post production studio work. Seemingly a lot of different folks offer that. And I don't think that necessarily the, the like production work that we do is going to be altogether that different than what some of the other players that are doing, I think also doing great work within the studio world are offering for us. What we do is we, we have sort of a. It's more of theoretical and conceptual of built for screen and fit for format. Right. And so what will happen is a lot of times we'll have clients that again, because they're mostly using us for our optimization logic to inform the best possible placement to best yield the desired outcome or media KPI of a given campaign. They'll come to us and they'll say, hey, here's a video asset or here's a display asset. Run it in all the different formats and channels that you want to give your optimization logic the best possible chance at driving the outcome of, let's say it's a click, right? And so what we'll often do is we recognize, hey, if we just get one piece of creative, the way that that creative renders on a mobile in app inventory versus mobile web versus a desktop versus like a C CTV impression, the way that we render that image or that creative is going to have vastly different results depending on the screen and the format that we serve it within. So what we try to do is take the creative assets that we've been given on behalf of the advertiser and then sort of optimize them for maximum efficacy based on the screen and format that they'll be delivering in. We have some tools that automate that process with some clients who are comfortable with allowing for that automation. And then we also have the ability to just kind of Create it on our own and then, you know, share it back to the client for approval right before we actually deploy it. And that's a team of two, fully staffed in house, team of 40 plus engineers and designers that build everything in house. So capabilities are like, you know, as simple as overlays and end cards to as sophisticated as AR, VR, 3D, chatbot, you name it.
- [15:26]
Ari
And what's the business model in general? Are you charging for data or for media? Is it disclosed? It's, it's media. So it's on, it's mostly through a deal id and it's a, it's a media fee.
- [15:37]
Matthew Dougherty
Yeah, It's a deal I.D. or you know, there's a managed service component of our offering too, But I'd say 70 to 80% of our business is programmatic deal IDs. Yeah.
- [15:44]
Ari
Okay, great. Let's, let's transition to our lightning round. So I'll ask you relatively quick questions, hopefully relatively quick answers. What is your number one competitive advantage?
- [15:55]
Matthew Dougherty
The optimum optimization logic, for sure. Ten years building a database of hundreds of millions of impressions that look at various signals that go into a given impression, time of day, screen Type, Domain Creative, etc. We query this database and are able to predict, historically speaking, how that data has most effect, has driven a given media KPI or outcome. And that's what decisions, whether or not we actually move forward with an impression or not.
- [16:22]
Ari
All right, and what, what is, how would you protect yourself against Google, Facebook, Amazon, etc. Ultimately crushing you? And I asked that of everybody.
- [16:35]
Matthew Dougherty
I think they have the capability to do that to almost anyone. I think is sort of the reality of the fact. At the end of the day, I think consolidation is coming no matter what. I mean, if you take a zoom out and look beyond the walled gardens, right, themes that we're seeing in the marketplace. SSPs are building demand side technology. DSPs are building their own private supply marketplaces. We're going to end up meeting in the middle and I think that the players that are going to be the ones who ultimately survive and can protect their business are going to be the ones that can show that they can be accretive to these mergers and still add some value to the supply chain. So I feel like we do that in a number of ways. But again, I would lean on the contextual capabilities that we have and the predictive optimization capabilities that we have as working very nicely with some of those tools. So our optimization logic, which is supply side driven, theoretically, if you're buying out through DB360. They're going to have their own optimization logic in the dsp. Our optimization logic combined with theirs is going to lead to improved performance.
- [17:39]
Ari
Got it. And what is your biggest challenge?
- [17:43]
Matthew Dougherty
I think right now brand building in the U.S. honestly, it's, you know, we're still relative newcomer to the space. It's. There's a lot of. It's a crowded space. There's a lot of folks out here doing great things too, with technology that I've seen and been exposed to has been exceptional. So I think it's just getting the name out, continuing to drive trial with our. With clients. We're fortunate that we do have, you know, a history of working with major brands and have been an established sort of entity in admittedly separate forms and different names over the course of a few years. But there's a lot of historical experience with a lot of brands working with XD in various forms. And I think that's helped us really hit the market and start to build some momentum quickly here in the US but that remains the biggest challenge, is definitely just continuing to build the brand.
- [18:31]
Ari
In the U.S. certainly a crowded market in the U.S. so, last question. If XD was an animal, what animal would it be?
- [18:38]
Matthew Dougherty
Oh, man, I love that question. You know, let me say honey badger.
- [18:44]
Ari
All right. Why. Why honey badger? I don't think that's the first honey badger we've gotten.
- [18:49]
Matthew Dougherty
Is it the first?
- [18:50]
Ari
No, I don't think it is.
- [18:50]
Matthew Dougherty
It's not the first. Okay. I'm not original enough.
- [18:52]
Ari
I guess that's pretty original. As long as you don't say lion. So why is any better?
- [18:56]
Matthew Dougherty
You know, we're like it. We're still relatively small, but I think that we have a tenacity both in our approach to business. The folks that we have here, you know, on our team, the boots on the ground in the U.S. our approach to building software and technology, which is move really fast, you know, be very quick and, you know, iterate as much as you can as quickly as you can. I think it's really based on that sort of tenacity. And the size of the animal isn't necessarily representative of its potential. More so what's inside, you know.
- [19:27]
Ari
All right, good answer. So thank you so much. Matthew Dougherty, the US The US CEO of xd. Thank you so much for being here.
- [19:35]
Matthew Dougherty
Thank you. Ari Foreign.
- [19:48]
Ari
Thank you for listening to the marketecture podcast. New episodes come out every Friday and an insightful vendor interview is published each Monday. You can subscribe to our library of hundreds of executive interviews at marketecture tv. You can also sign up for free for our weekly newsletter with my original strategic insights on the week's news@news.marketing. and if you're feeling social media, we operate a vibrant Slack community that you can apply to join@adtechgod.com.