Signal Break: Tejas Manohar, Co-CEO of HighTouch Talks LiveRamp Data Purchase Proposition (from Publicis)
- Jun 19
- 35 min read
For weeks, reports that Publicis was exploring a relationship with Hightouch following its acquisition of LiveRamp circulated across the industry. Until now, no one involved had publicly confirmed it.
In this special Signal & Noise Signal Break, Hightouch co-founder and CEO Tejas Manohar confirms that Hightouch has been in discussions with Publicis. Rio Longacre, Brett House, and Tejas unpack what a potential Publicis–LiveRamp–Hightouch relationship could mean for identity, activation, clean rooms, composable CDPs, warehouse-native activation, and the future of marketing infrastructure.
To be clear, Hightouch is a sponsor of Signal & Noise, but sponsorship played no role in our decision to cover this story. We believe this development deserves broader industry attention.
Topics discussed:
Tejas confirms Hightouch has spoken with Publicis
What a Publicis–LiveRamp–Hightouch relationship could look like
The future of composable CDPs, identity, and activation
How AI and agentic workflows are reshaping marketing infrastructure
If you're trying to understand where the modern marketing stack is headed, this is a conversation you won't want to miss.
Watch the full episode and join the conversation.
🔑 What We Cover💡 Key Takeaways🎯 Why This Episode Matters
Read the full transcript below.
Brett House (00:01.326)
Hey everybody, welcome back to Signal and Noise. This is Brett House, joined by my co-host and favorite person, Rhea Longacre. and today we have Tejas Manahar, this is co-founder and co-CEO of High Touch, who's also a sponsor of of our podcast, but we promise to do no shilling or or advertorial type episodes. Yeah, but yeah, yeah. Awesome, that's great to hear. Tejas, it's great to have you on the episode.
Rio (00:21.769)
The editorial decisions are separate. We would have had on regardless, just to be clear.
Tejas (00:25.425)
FONS for your new fan, personal fan.
Brett House (00:31.406)
certainly, you know, first time we've met. I don't know, Rio, if you've met if you guys have met before. good, good. Yeah. So so I so it's good for me to meet you. But so I just wanted to give a little background. Prior to joining a co-founding and co CEOing HiTouch, you were the engineering lead at Segment, right? which is another cut customer customer data platform that people don't know that was bought by Twilio for three a cool three point two billion in twenty twenty. So
Tejas (00:36.701)
for sure.
Rio (00:37.045)
We know each other, yeah.
Tejas (00:51.293)
That's exactly right.
Brett House (01:00.578)
That must have set the stage for some of what you went on to do next, right? you're also you're also you know, and I thought this was pretty interesting about your background just from a personal perspective. You're a visiting scholar at UC Berkeley. you worked at the NetSys lab under Dr. Scott Schenker, who not only came from kind of a theoretical physics background, right? Chaos theory and stuff, but was a is a professor of computer science at the school, obviously a very highly rated.
highly recognized school so that must have been pretty pretty an interesting experience
Rio (01:29.407)
Yeah, my my father is a nuclear physicist, Tejos, and he went to Berkeley. So definitely appreciative that.
Tejas (01:33.809)
That's hard.
Brett House (01:33.888)
Yeah. Yeah. So when we see those things combined thematically, we're like, and if you've ever met Rio's dad, he is a great guy and and loves to talk about theoretical math and and and quantum physics. Yeah, quantum physics and particle theory, and you're like, my god and he he poo-poos the the non serious scientists, like the ones that are out there writing books versus the researchers, which I was which I thought was
Tejas (01:44.292)
nice, nice, way smarter than me.
Rio (01:56.233)
he hates them. Yeah, it's it's so funny.
Tejas (01:56.987)
What is your computer scientist? What does he think of that?
Rio (02:01.075)
It's funny. So he uses a lot of the same programs. So within you you have experimental and theoretical physics. So he's a theoretical, right? But then a lot of the experimental guys will actually use the detection equipment in a particle accelerator, you know, where they're like it's photon beam, they're accelerating either photons or in the relative ex heavy ion collider in New York to actually accelerate gold nuclei. So he worked in that for a long time. So it's funny. So within the detection equipment, he would use a lot of the same programs written in COBOL, believe it or not, from like the late sixties, early seventies.
'Cause they're not there was this they're not object based, right? So so he says they're cleaner, they're better, and he says they actually work. They're he he it's funny, he's an you know, he's an old guy. I mean, any any in any innovation that comes after a certain age, people usually like are not super excited about it. But he swears that they're they're actually better than some of the the the new programming languages they use, which is kind of interesting.
Brett House (02:39.18)
He's an old timer.
Brett House (02:50.806)
Yeah, so so we can we can geek out, Tangis, as as you can see. But just a little bit about high touch for those that don't know is is you know obviously one of the one of the big companies in customer data, a leader in composability, with your composable, you know, the composable CDB category c category, which you kind of created, right? and you've challenged a lot of the assumptions in the marketplace about how marketing technology is built, right?
Tejas (02:54.749)
I'm good.
Brett House (03:15.646)
And for those that haven't been reading the news about high touch, you guys are fresh off a hundred and fifty, yeah, hundred and fifty million dollar funding round, with a valuation approaching three billion. That's that's phenomenal for you, Tejas, and your co-founder and the in the whole company. yeah, you guys are leading the charge with AI agentic in how it applies to to marketing tech and the advertising ecosystem. And even more recent is this live this live ramp
Rio (03:20.073)
And there's a lot of news.
Rio (03:29.355)
Yeah, congrats.
Tejas (03:30.781)
Thank you.
Brett House (03:43.598)
push that you guys have made which has popped out in the news recently.
Rio (03:45.099)
Yeah, that came out just this week. We definitely like to dig into that about like, you know, is is it true number one and get your thoughts on it? But but yeah, I'm I'm interrupting. So but I think we wanted a little did Brett miss anything about your background, by the way. I mean that was a detailed di deep dive, but anything like that it was worth calling out we missed?
Brett House (03:51.948)
Yeah. Yeah.
Tejas (04:02.555)
No, I think that's pretty comprehensive, I'd say.
Brett House (04:07.616)
All right. So so yeah, so maybe we so well so so Rhea, why don't you start with the thesis of the episode and then we can dive into some of the live ramp and publicis details and stuff like that.
Rio (04:07.659)
Sweet.
Rio (04:15.135)
Yeah. So so so so Tasia, this has been a long time coming. You know, we've done a bunch of stuff on C D Ps, on data, on AI, on a lot of the stuff that's related to this. But you know, the one of the reasons why we wanted to have you on and particularly to talk about high touch is we like you completely innovated and transformed, I think, the whole like what initially was CDP, but it's clearly a much broader category. As we even question, is that a distinct software category, right? You've you're really disrupted this area which you as as a pioneer in a composable space.
Tejas (04:15.399)
Let's do it.
Rio (04:43.967)
And now what you're doing with connecting first party data, connecting it to AI, building that control pane for marketers. I mean, that's it's so it's it's very interesting what you're doing. I really feel like you're a company that's quietly at the intersection of several of these massive shifts, and you've been driving all these massive shifts. So as we start to rethink how is customer data stored, how is it used? How does how does it like activated for AI? And now even with the live ramp stuff and ramp for how is it activated for paid media? I think with with match booster and some of the
the features you've launched over the past the past year or so. I I think you're you're it's clear you're planting from at least from what we're seeing, you're planting your flag in that space as well. So the timing's great with can right around the corner. So welcome to the show.
Tejas (05:27.441)
Thanks for having me. Would it be helpful if I just shared a little background on Hightouch and perspective on the market? Cool. So as you mentioned, I'm Tejas, one of the founders of Hightouch. We started the company about six years ago, but it's been growing incredibly fast. It's actually not even six years. It'll be six years at the end of this year. And so we've recently crossed a hundred million in revenue. We announced growing over a hundred percent year on year. We work with some of the largest enterprises in the world.
Brett House (05:32.096)
Yeah, totally. Yeah.
Rio (05:32.447)
That'd be fantastic.
Tejas (05:57.283)
know, the PetSmartz, Warner Music Groups, NBA and DoorDash and so on and so forth. And really why I founded the company is, you know, I was an early engineer at Segment before. I was sort of in the, you know, rise of the original CDP category. But when I looked at these enterprise companies, it was very clear that the way marketing teams were able to use the data was very broken.
Brett House (06:14.733)
Yeah.
Brett House (06:22.967)
Yeah, yeah.
Tejas (06:24.189)
know, CDPs never really solved this problem. They promised the single source of truth for marketing. But you look at the enterprise organization, the marketing team had no ability to actually get all the data into a CDP like segment or Salesforce or Adobe in the first place. Like it's just a very onerous.
Brett House (06:39.906)
Yeah, it it it might yeah, it might live in five or ten different data warehouses across multiple groups within an organization, especially the big brands.
Rio (06:45.016)
I and Brett, that's why mo that's why many of them failed, many of these C D P implementations.
Tejas (06:49.213)
That's exactly right. I think Gartner or one of the analyst firms had a stat that like 50, 60 % of CD implementations didn't even get off the ground. And then there's the ones who got off the ground that could only serve a very subset of the use cases across the enterprise because the data was simply not there. Honestly, my thesis on the market after working at Segment from when it was a 30 person company all the way through a year before the acquisition was that it's just impossible.
Brett House (06:58.915)
Yeah.
Brett House (07:06.915)
Yeah.
Tejas (07:18.333)
for marketing teams to solve the problem of centralizing data. And that has to be a company wide initiative. And so what gave me the confidence to go start Hightouch with my co-founders in late 2020 was that I saw enterprises flocking towards these cloud data warehouse systems and Lake house systems. So, you know, the growth has even been rapid since then, but in the year of 2020, we started Snowflake had the largest software IP of all time.
Brett House (07:36.906)
Yes, yeah.
Tejas (07:47.249)
Google Cloud, BigQuery was growing like crazy. And Databricks, you know, wasn't even on the block yet as a data warehouse. And it's now come to the occasion and is a major player of the space with billions of revenue. Yeah, exactly. They were just in data science and Python notebooks at that time. So the market was just exploding. And I realized that isn't it crazy that, you know, companies are actually going to have centralized data, but then when it comes to using that data, all they have is Power BI.
Brett House (07:50.178)
Yeah.
Brett House (07:53.963)
Yeah.
Brett House (07:58.169)
But they're an they're a new entrant at that point, yeah.
Tejas (08:16.381)
and the same report for 20 years.
Brett House (08:17.556)
Yeah, yeah, yeah. Yeah, for for analytics and dashboards, right? They've got i or or the or they're trying to log into these these independent sort of monolithic SaaS platforms off to the side that don't, to your point, capture all of the data from the from the entire enterprise. They might it might be a sub segment, so it doesn't really serve the purpose. Right? Like
Tejas (08:34.969)
Exactly. And so then I was like, we got to start a platform for marketers that sits on top of the new centralization of data in these warehouses. The first market we took on was the CDPs. We created what we call the Composable CDP. It's a major term in the industry now. It wasn't a term before. But we realized this, you know, we always had this division of let's build an application platform for the business on top of these data warehouses and clouds. So data onboarding and ad tech was the next challenge, which
Brett House (08:52.184)
Yeah.
Tejas (09:04.775)
brings us to today. And we've launched a match booster product, is highly competitive with the likes of live ramp in the industry. We've replaced major live ramp of implications at Fortune 500 companies with our product. And then we're now powered.
Brett House (09:16.876)
Yeah. And and that's competitive with ramp ID in terms of matching and synthesizing data into kind of one token or
Tejas (09:21.565)
It is, yeah. I mean, on that point, I think when RampID started, the cookie graph was a huge differentiator with all the browser and platform changes and everything. It's hard to imagine. It's old. Yeah, it's an...
Brett House (09:32.096)
Which is hard to imagine, right? The cookie graph was a huge differentiator.
Rio (09:32.723)
It was, yeah. Well
But even rewinding, but t just even rewinding a couple years, I remember I can't tell you over the years. Well, I think the timing of this is good. This is kind of what I'm getting to here. Is over the years, I can't tell you how many companies I worked with, and many of them in the CDP space, right, who were say, we have a live ramp takeout strategy, right? And it was just like come and they would try and they it would they wouldn't get any traction. They didn't understand the paid media space, they didn't know how to work with agencies, and they didn't even understand the difference in Martech and ad tech. Totally. And then so
Tejas (09:58.173)
See you next time.
Rio (10:03.689)
I've seen that happen so many times, but it's interesting. You rolled this out. It's I don't know, I don't know how much of this is been, you know, is great strategy versus versus luck, but the fact that live ramp gets acquired, they're no longer Switzerland, right? And I would argue the value of ramp ID totally changes give given this, right? So I think that's correct. So the timing is good for this.
Brett House (10:10.05)
Why didn't
Brett House (10:18.754)
Yeah. Because because of the lack of neutrality, because of the fact that yeah. Yeah. Well, and did did live I mean, so the ramp ID was built also on the back of this massive publisher ecosystem that they had, right? Twenty-five thousand publishers, right? Which gave them a natural advantage in moat over anybody that was coming in. We tried to New Star tried to launch a a an ID, but but we didn't have that sort of validation and data ecosystem to power it. To build the public publisher ecosystem takes time, effort.
Tejas (10:22.223)
Exactly. Yeah, exactly.
Rio (10:38.859)
Yeah, they had the pipes built. I mean good yeah, good luck displacing them, yeah.
Brett House (10:48.482)
B D and it was just like we just gave up. It was called the Fabric ID. But but what how did you guys think about this?
Tejas (10:52.497)
Yeah. Yeah. I mean, at the time, live ramp came to market. You know, think, I think meta didn't even accept first party data at this time. Like live ramp was the way to do this. The ramp ID was the way to do this. It was the backbone of the advertising ecosystem. Since then, you know, the, API connectivity of a lot of the major providers and even the long tail, like we can send data to the LG TVs, the Samsung TVs of the world, the Roku's of the world. We've struck partnership with these companies without open APIs.
Brett House (11:05.144)
Yeah.
Tejas (11:21.733)
The ecosystem has really evolved from a technology perspective and we are actually able to send first-party data and enriched and hashed first-party data to these different platforms. Obviously, we still don't have the number of sheer connectors when you think about the long tail that a ramp ID has, which brings us the news of this week. And it did. It took them years and years to build that up. And I think when you look at the channels that
Brett House (11:38.2)
Yeah.
Rio (11:42.059)
But it took them years to build that up though, right, yeah. So
Tejas (11:49.761)
an enterprise is activating on, you have to wonder how valuable is the whole long tail as well sometimes. But I think we've built a really substantial base of hundreds of connectors at this point with most of the most popular channels. And yeah, lot of it is market timing. I did see companies actually, really small startups try to do something like the Composable CDP three, four years before us, the market wasn't ready for it. Companies that probably try to take on live ramp.
when the cookie graph was very important, probably couldn't succeed. So some of it's market timing and then some of it's also just having, I think we have always had a good perspective at high touch on where the market is headed. So, you know, obviously we couldn't predict the live rep news from three weeks ago with them being brought into publicist when we launched our match booster product a couple of years ago. But what we could see is that everyone relies on live ramp.
there's a lot of frustration with the solution. Like the actual process of onboarding your first part data to the solution is.
Rio (12:51.379)
And the cost, the cost is so high.
Brett House (12:52.21)
yeah, and the tax for both data in and data out, you're being c charged both ways, so it's extremely expensive, but there were no other alternatives. Newstar Newstar tried to play that's basically.
Rio (12:59.093)
Well a a tax on the internet. Yeah, I've had so many clients call it a tax on the internet. Yeah, yeah, sorry Ted, just go ahead.
Tejas (12:59.965)
It's a percentage model. Yeah, it's a percentage model and we believe that.
Brett House (13:04.322)
Yeah. And that frustration is not that frustration is not just your opinion. That that is a frustration that I've heard from multiple executives across the ecosystem for years. Yeah.
Tejas (13:14.269)
Exactly. yeah, where, where I think my personal perspective is we're always just trying to go wide. We're trying to talk to the customers and be like, Hey, what other challenges do you have? then continues to look for, for when it's something we can solve. And it was actually a partnerships person on our team who had the realization that, you know, the graph wasn't, wasn't that unique and we could build it up with other partners. So these ideas come from anywhere in the org too.
Brett House (13:27.693)
Is that
Brett House (13:36.31)
S Yeah, so well so w that's great. That's great. Yeah, you get a little bit of the the the bubbling up, right? The you know, pull the conden cord, right? And and you know put s put some some recommendations in a little box next to the the manufacturing facility. So
Rio (13:49.107)
That's awesome. So the so the news is the news is true that the news is the news is accurate, Teja. So the news that we that that was reported this week about it.
Tejas (13:50.671)
It was my first time doing
Brett House (13:54.635)
Yeah.
Tejas (13:55.803)
The news is accurate. We did reach out to Google assist to see if we can strike a more novel offer with them where where we own the identity spine and the data onboarding piece of live ramps are not their overall product. I think that represents from public.
Brett House (14:03.832)
Yeah.
Brett House (14:12.8)
Yeah, it's the data matching, it's the identity resolution play, right, which is so critical, but not not the ramp ID. Yeah.
Tejas (14:18.129)
Exactly, right. Yeah, and we see it as a win-win-win, right? I think that's why we reached out. For us, it's a win. We get like a 10-year head start on go-to-market, channel partnerships, all of these customers who depend on the ramp ID today who are in need of a better onboarding solution that's built, you know, plus five, six years, not 10 plus years ago. And then, two, you know, I personally think that
Brett House (14:36.875)
Yeah.
Tejas (14:46.205)
My personal opinion is what I'm seeing from customers, they're not happy with it being owned by a single party in the ecosystem. Customers want this to be neutral. even if, regardless of what happens in this kind of situation, customers are going to be looking for alternatives anyways. it's better for... Yeah, exactly.
Brett House (14:54.231)
Yeah.
Rio (15:06.389)
They are. I can I could confirm they are. I mean I I've had many customers and account execs on the media side are basically say how fast do we get them off live ramp? I'm hearing not quite a bit. So
Brett House (15:09.164)
Yeah.
Brett House (15:15.628)
Yeah. And would you be using that if you're if you're composable and you're plugging into where the data is, the data warehouse, you the data bricks, the snowflakes, et cetera, right? Does this add sort of a it's a data validation layer, I'm assuming, but it's also a data enrichment layer, right? I'm assuming because you've it's an enrichment layer almost entirely where you're enriching their consumer data to enable them to just create more adjustable audiences, to know a little bit more about their, you know, to have better segmentation models and audience profiling.
Tejas (15:15.984)
Exactly.
Tejas (15:29.541)
It's an enrichment layer, yeah, largely. Exactly.
Tejas (15:41.859)
Exactly. And we already have our belief from empirical evidence with these large customers is that our match booster enrichment layer is already as good or better than LiveRamp on the major ad platforms. But obviously they have way more customers than us. If we can get a head start on that, agency relationships and longer tail channel relationships, that we can then bring a better technology that can serve more use cases across the enterprise of these customers instead of ad tech.
Rio (16:05.653)
Well, you know, Tay, just one thing I think you could add to them is like they've never been good at actually building out relationships with systems integrators, consultances. You guys didn't done a much better job at that. I think it's been really core to your business and why a lot of a lot a lot of like why you've been easy to work with, right? That you've never competed really on pressure services, you've relied on partners, you've used them for scale, you've continued to innovate. So I think that like and they frankly that's been a limiting factor in my opinion on on their growth and why, despite having
Tejas (16:18.47)
It has.
Yeah.
Rio (16:33.971)
a quasi a monopoly when you really look at what the the role of ramp ID and how it's used across the ecosystem. Brett, your point about all the publisher integrations, the fact that, you for Meta they were the were the only the it was the only way to really onboard first party data for a long time. I think they could use that. so it actually would make sense. Now how's Publicys responded?
Tejas (16:51.175)
Yeah, I can't comment on that. Bye.
Brett House (16:52.68)
Yeah. Well and and well let's just talk about a little bit more. So so doesn't a lot of this stuff when you talk about data matching, data enrichment, right, doesn't that help power the sort of data collaboration use cases, how real they were is up for debate with Librant. like you know, so so would would LiveRant be willing to give that up? Would that break something within their current solution set without having, you know?
Rio (16:52.778)
Okay.
Tejas (17:13.893)
Yeah, it's great question. At least, it's difficult to say how the solutions can truly untangle. But what we're proposing is, I believe, a win-win-win where we could still strike a license agreement where the publicists could keep using the technology for the rest of live-rem ecosystem. But the actual spine remains neutral, which, in my opinion, benefits us. We can go to those customers and build new value-added solutions, deliver them more with high-touch and better solutions.
Brett House (17:21.208)
Yeah.
Tejas (17:43.887)
And, then also it's not a churning asset under PupilSets because no one wants to use a spine that's completely controlled by PupilSets. It's any agency, any, any Holdco, like no customer wants to use a spine that's controlled just by a Holdco. The whole point is to have it, have a neutral system and have your own technology. So we thought it's possible to strike a win-win-win. And that's why we proposed this to LiveRamp, a win-win for us.
Brett House (17:49.58)
Yeah, it's good for the ecosystem.
Brett House (18:05.634)
Yeah, and I yeah, it's good. I think it's good business to ecosystem practice. And I think the question is is whether somebody like a publicist is doing all of this specifically to build data advantage and to build a moat, right? A a walled garden as a hold co to compete with meta, to compete with
Tejas (18:09.765)
Exactly.
Tejas (18:22.289)
But even if they are, I would argue that the identity graph of live ramp, you know, it has maybe a mode like properties in terms of customer traction today, but not in terms of the actual data. So you could probably see reactions on, you know, on LinkedIn all around the ecosystem from the ID fives of the world to every data provider. Well, I'll tell you myself, like, yeah, I thought I'll tell you myself. I think the, you know, with our match booster product, we've been able to partner with a number of folks out there and build a graph that's.
Rio (18:22.959)
Brett House (18:33.965)
Yeah.
Brett House (18:40.461)
Yeah.
Rio (18:41.887)
They're happy.
Tejas (18:51.683)
as good or better for match rates. Actually, one of the reasons we found it's better is because in the ramp ID, the process is like you upload your first party data to live ramp and only if it matches to the ramp ID, does it make it to downstream platforms. Whereas in high touch, it's Richmond layer, right? It's like you take your first party data in your warehouse, you hash it, we hash it. And we try to send it to platforms, which now except first party data wasn't the case when live ramp came out.
Brett House (19:08.256)
yeah.
Tejas (19:20.143)
And so at a minimum that HPEI is going to make it to platforms at a maximum. It's exactly it's it's ramp ID plus your PII plus plus. And so I think sure, maybe one of the motivations we can speculate from the outside was to have a data mode, a data advantage. But I would argue that that doesn't really exist anymore in the industry because of the demise of I grams cookie graph. Like we don't think that actually exists anymore. So we think it's a win win win.
Brett House (19:20.696)
Yep.
Rio (19:23.933)
It's more than just ramp ID, right, yeah.
Brett House (19:42.061)
Yeah.
Rio (19:50.197)
Yeah, funny one of the theories I've heard for why they did it. I I don't agree with this by the way, is they did it just to take a chip off the board and create chaos for everybody, right? As everyone tries to untangle from it. So yeah, just because they got I don't think that's the case. I mean and I think they actually are trying to build kind of a walled garden.
Brett House (19:58.851)
Just because they could?
Tejas (20:02.637)
Interesting. I think there's another reason people aren't talking about here, which is that if you're an agency right now, it is absolutely critical that you inject software DNA into your company. Right? Like the future is revolutionizing. Like agencies are here to stay. Exactly. Yeah. They're here to stay, but they got to know how they work.
Brett House (20:13.9)
Yeah. Yeah.
Rio (20:14.771)
Yep. W We're we're doing it here. Omnicom, yeah. We're building Omni. I mean I
Brett House (20:20.578)
Yep, yep, and
Yeah, yeah. It's and it's yeah, yeah, and it's and it's something that that like one of my big clients at at high high signals is I I won't name them, but they are a independent that's doing exactly the same thing. They had whole bunch of solutions. It's it's it's reflective of the entire agency ecosystem that were all point solutions that were sort of built by sort of you know, some R and D, but it was all dis it was all distributed across the agency. They had no central product organization, no centralized data layers or or creative layer for.
Tejas (20:51.238)
Yeah.
Brett House (20:52.718)
creative versioning. And so th the the time has come for these guys to bring all this stuff into one place. And the product lifecycles are so fast that they can replatform and bring together and connect everything around a common infrastructure. and make it fully composable in like six months, eight months versus, you know, what might have taken years in the past, right, in terms of total build time. So it it the independents are moving fast in that direction. The Hold Coast are trying to, they're a bit slower moving because they're so large. So acquisitive becomes a part of their play.
Tejas (21:20.775)
But nothing helps as much as injecting a software company into your company to help with that transformation, right? So I think that's the reason that people aren't talking about as much on the market too. This is a software company not valued that highly on the public markets, reasonable multiple value created asset for them to get out in their books. It's not some super overvalued AI startup over here. It's something that they can have on their books responsibly.
Brett House (21:36.886)
Yeah.
Tejas (21:46.19)
with the revenues helping as well as the DNA helping in the companies is the way I see it.
Rio (21:52.021)
How does these how do these recent announcements, whether it's matchbooster or depending on what happens with us, how does that change your strategy? I mean, I think historically you sold mostly to, you know, let's say to marketers themselves, to marketing departments, to heads of Martech. Is that changing? Do you see yourself selling through media more? I'm assuming that's part of the reason why you don't match booster. Is do you see bigger partnerships with agencies and hold codes as opposed to just systems integrators? Comments on all that?
Tejas (22:16.285)
Yeah, great question. So yeah, regardless of what happens here, our strategy doesn't change. are focused, know, MarTech was our original persona and data teams, but we're very focused on ad tech. We have a whole division within Hightech that works on ad tech from a partnerships perspective, from a product perspective, from a product marketing perspective, from a sales perspective. And that's both serves advertisers and brands, but it's also to serve
you know, the sellers, the publishers, like the retail media networks or the big streaming providers, many of them were finding their top CDP use case, customer data use case, is actually now ad sales, right? They want not their marketers in the CDP building audiences for marketing, but they want their ad sellers in the CDP building audiences, sizing them out, pricing them and shipping them and activating them on-site and off-site for...
Brett House (22:48.801)
Yeah.
Tejas (23:11.943)
for advertising sales. And we're seeing this across, know, RMS, retail, know, streaming providers, just publishers in general. So this is already actually a pretty big use case of ours. Obviously not a number of customers like we see on the marketing side, but in revenue and in strategic relationships that we have with, you know, very large companies, it's actually already a big use case for us. And we're dialing it up regardless of how things pan out here. But...
If there's a way to accelerate it, we're highly interested in that.
Brett House (23:43.979)
Yeah. So going back to that question of the like the composable CD C D P movement, right, that you guys sort of spearheaded, right? And this sort of architectural debate about, you know, there was a lot of talk when I was at New Star, for example, about interoperability and b and building native within the with the cloud data warehouses, right? Because that's where everything was going. It wasn't quite there yet. There was some market timing that you mentioned before that that really you saw what was going to happen. There was going to be a centralization of data. Why not plug into that? So what do you think you got right?
in that in that decisioning that kind of and and what kind of what was the drive behind taking that that big leap of faith to say like we're gonna do this and we're gonna go all in on it, right? 'Cause a lot of other people missed it.
Tejas (24:26.333)
Yeah, a hundred percent. So one, I think we recognized it fairly early. Like how fast the momentum was growing in these large enterprises. Like when we started Hitech and we talked to large enterprises, a lot of them told us, no, I don't have all my data in the warehouse. Like, sorry, I need to buy an independent CDP or Martech or ad tech platform that can house my data. But we, we, we saw where the puck was headed. So we were like, okay, talk to you guys in a year. And then we talked to those companies in a year.
Rio (24:44.992)
Isn't it?
Brett House (24:53.08)
Yeah.
Tejas (24:56.207)
And so that was one thing I think.
Brett House (24:57.388)
And they're like our bartend investment failed. we we need
Rio (25:00.011)
Or a lot of times we're we're still building new CDP a year later, right? That's what you would hear all the time, right? We're still trying, right?
Tejas (25:00.367)
Yeah, exactly.
Tejas (25:05.468)
That's exactly right. And so we've got, have a lot of boomerang customers. We call it here at HiTouch. actually run an analysis on our sales pipeline about a year ago. And it's like, you know, our sales cycles are not too long because we can pilot solution pretty fast. But sometimes we talk to customers, like we have like three or four opportunities with the customer before they convert of just educative cycles that result in nothing. I actually feel bad for customers. We need to find a way for them not to do that because it's not a good use of that. But.
Brett House (25:34.669)
Yeah.
Rio (25:35.435)
Yeah, I loved your marketing kit, hey friends don't let friends buy C P so it was something like that, right? That was that was funny. Yeah.
Tejas (25:35.644)
Get out.
Exactly. More people need to read that and then they don't have talk to us three years straight. They could just get it the first time. anyways, one I would say is we were early, but we were convicted. So there was a lot of nos. You know, our win rate was literally bad from a sales perspective as we were scaling the company in the early years because the market wasn't ready for the sale. Like how bad? yeah, maybe in the first year or two of seriously focusing on enterprise CDP.
Brett House (25:56.981)
Like how bad? Like ten percent, fifteen percent, twenty percent?
Tejas (26:07.197)
10, 15 % just because a lot of.
Brett House (26:08.79)
Yeah, kud kudos for sticking through that. That must have been stressful. Yeah.
Rio (26:09.163)
Okay. But but but but but but when you first started, I think the only composable CDP back then w remember Action IQ, like they were like that was something. Yeah, semi yeah, and but I remember the complaint people always had was it was so hard to actually implement it, right? You you needed like heavy engineering resources, so it ended up being and they they had so so you obviously I think timing was part of it, but also yeah. It was.
Brett House (26:16.823)
Yeah.
Tejas (26:17.959)
Semi-composable, semi-composable, I'd say, yeah.
Brett House (26:20.343)
Yeah.
Tejas (26:25.401)
implement.
Brett House (26:25.44)
Integrate or implement, yeah.
Tejas (26:32.337)
And services too, heavy, heavy services, yeah.
Brett House (26:35.328)
It was heavy heavy managed services. Yeah, so so that's a good like from a strategic perspective, because I think of composability. I mean, t tell us how you architected to do this, right? It it allows for faster time to value, right? Faster time to implementation because you're only potentially implementing around a specific use case and a specific core component of of the capability versus the whole suite of solutions you have.
Tejas (26:44.935)
Yeah.
Rio (26:55.549)
And a much lower entry cost depending on on how much you want to bite off to. I think that was the flexibility there was huge.
Brett House (26:59.852)
Yeah. How did you think about that? 'Cause you're you you're a computer software guy, like how did you architect and think about like, Hey, let's make this into bite sizable pieces that can be implemented and then you can just you can add you know, faster time to value.
Tejas (27:01.383)
Yeah.
Tejas (27:10.877)
So there's like a clear statement that kind of grounded our product strategy, I would say was that we don't believe the CDP is a product. It's just a collection of products and different people need different instruments. It's like, it's a little weird. It's like kind of a bag of tricks category. Like everyone needs something in the bag, but it's a little different. Like you need to sync 20 audiences because you're just selling the same ad segments. Your company is not that complex and you need a hundred different because you have a really...
Brett House (27:22.113)
Yeah.
Brett House (27:31.309)
Yeah.
Tejas (27:37.617)
heterogeneous business, you have tons of different use cases across the enterprise. And so I think people have different, different use cases. Some companies have solved identity in house. Some companies need a CDP to solve it. Some companies need data collection. Others, actually, we said, you know, Google Analytics, you already have that data in your BigQuery. Use that. Or Adobe Analytics, let's use that data. So I think the way we architected the software was around that. So that idea that the CDP is not a product, it's a platform.
and as a set of capabilities, similar to like an AWS of marketing. If you think about it, where you log in, you get all the features on the console. And so each capability, identity resolution, data collection, audience activation, journey orchestration, same session real time, it's priced independently and differently. Whereas all the other CDPs in the market had one big price on the number of users your company had. Exactly.
Brett House (28:08.129)
Yeah.
Rio (28:27.027)
It's a bundle, yeah.
Brett House (28:28.78)
Yeah, and you might only use it for yeah, yeah, and you might only use it for thirty percent or twenty percent of its total of its total platform size or capability size, right?
Tejas (28:37.789)
Exactly. And the way they price too is there's the number of users your company has, which is really bad for publishers because they have a lot of low value users. So it like makes no sense. There's a debate on who is a user with Anonymous. And then two, it's like, it's kind of like that tax mentality. It's like, just like live ramp, you're a big company, spend a lot on ads. Okay, here's a, here's 5%. Uh, it's similar with number of users. You might need to start small. You might be serving one part of the business, but you have a big tax because you're a big company and have a lot of users. So.
Brett House (28:43.137)
Yeah.
Brett House (28:51.33)
Yeah.
Tejas (29:08.145)
We price on each.
Brett House (29:08.588)
Yeah, it's the it's it's the death of that seat model, right? So that's a that's a pricing sort of evolution from a timing perspective.
Tejas (29:11.93)
Exactly.
Rio (29:13.035)
Well well that could be bright. It could be very expensive. Like when it was like cost per cost per record, right? And that could be I mean, I think with some of them it was up to ten cents. I mean it's incredibly expensive to your point, if especially and prohibitively expensive for a publisher. I agree with you.
Tejas (29:27.129)
Exactly. And so if you combine the fact that you're paying a lot because you're a big company, you have a lot of users, you can't actually utilize the solution because you can't actually get all your data into it because you're a marketing team and that's a cross-functional CIO level initiative. And then three, these solutions were... Go ahead, go ahead.
Brett House (29:40.301)
Yeah.
Brett House (29:44.544)
And it and it's sorry, go ahead. No no yeah, yeah, yeah, and and then I I would say I was gonna say three being it you you're under utilizing for the price that you're paying, right?
Tejas (29:54.371)
Exactly. Right. And then three, and then on top of that, I'd say is that they were very inflexible. So you mentioned services, you mentioned a lot of implementation. There's a core technology that we built at Hitech we call the kind of marketing semantic layer. So basically it's a technology that allows the data analysts at a company.
So we don't need any engineers, you don't need APIs, you don't need SDKs to implement our CDP to literally go into solutions like Databricks and Snowflake and say, I want this table in the CDP and I want to rename these columns and I want to change it around a little bit. And then I want it to be joinable with this other table. So you can say, hey, I want my users to be joined with credit cards, to be joined with households, to be joined with bank accounts, to be joined with this. And you could take your tables as is from the data warehouse and say, I want them exposed to marketers.
Brett House (30:33.453)
Yeah.
Tejas (30:47.569)
Whereas every other CDP in the market, whether it's a Salesforce or Adobe or segments, still works this way today where they have this cookie cutter definition of what customer data should look like. Some are more event-based, some are more account-based, but that cookie cutter definition just doesn't work for the enterprise. Everyone's a little different and it's way too much work to translate and to keep translating over and over and over and over.
Brett House (30:58.637)
Yeah.
Brett House (31:07.574)
Yeah. And so the and so the semantic layer acts as that my favorite word, Rio, we talk about this all the time with tracer tech and everything, but but it acts as that it it just allows for what let's say tailored or bespoke solutions, based on how they're how they're defining their entire data ecosystem as a brand.
Tejas (31:11.421)
convention.
Rio (31:11.689)
Yeah, we d we do.
Tejas (31:22.363)
Yeah, think of it as you label your data and the way you want and we generate our app around that, our marketing facing app around that versus we have a marketing facing app that's static for audience building, journey building, and you need to tell us how to put your data in the high-touch format. Like we don't do that.
Brett House (31:29.805)
Yeah.
Rio (31:36.093)
Yeah. It i i i instead of relying rely instead of relying on the schema that the C D P is gonna give you ahead of time, you can actually say this is what we want and with there's different tables. Yeah.
Brett House (31:37.046)
And it and it natively and it native
Brett House (31:42.7)
Yeah.
Tejas (31:43.075)
Exactly. And they're like, you know, a lot of people want to do product catalog ads. Let's add products to our schema. Whereas in high touch, anything support in the schema. And so that is a big differentiator. It allows the company to really easily set it up so they can point it at any which tables in the enterprise without services. Like companies said they could do this, but it was with services. Like they send data engineers to your office to wire it up. We don't need services.
Brett House (32:04.726)
And and and does that semantic layer is that does that semantic layer power all of the different use cases? Right? So you've got this kind of yeah. So yeah you add that and then you can start just selling use case by use case, and they're only paying from a pricing model perspective. They're not paying a per seat, they're paying like on a use case. Is it a monthly or a flat rate? I mean how how how does that pay
Tejas (32:10.361)
all of high-titch. That powers the whole platform of high-titch from.
Tejas (32:24.925)
Exactly. Yeah, it's based on use cases. It is an annual committed kind of like SaaS project, et cetera. But it's not pure consumption, like we just pay a dollar and tomorrow you pay $2. But it is paid by use cases and we have a very land and expand type model. So we have very good NRR as a business as well through expanding over time.
Brett House (32:30.797)
Yeah.
Brett House (32:38.488)
Yeah.
Brett House (32:44.694)
Yeah.
Rio (32:46.763)
Nice. Yeah. Yeah. So that and then the semantic layer, so like the AI pivot you're making, like that that must make it very easy for for clients to use then tool use these AI, these di this agentic layer building top of it to do different marketing functions, it sounds like. Is that is that correct?
Brett House (32:47.726)
Yeah, it's an important important metric. Net recurring revenue.
Tejas (32:50.109)
haha
Tejas (33:02.331)
That's exactly right. guess we should touch on that too. you know, I, on top of all this news that happened recently in the last six months, we've launched what we call the agent tick marketing platform. This is, this is actually a really big leap for the business. Like for the last five, six years, we've been fully focused on the problem of activating data for Martech ad tech and digital kind of product use cases. Now we're saying that, you know, it's not just data, the next wave's AI and, fundamentally.
Brett House (33:13.762)
Yes.
Tejas (33:31.907)
All the work around these use cases that we power is going to change. Like as a marketer, you're not just going to be asking your creative services team for every single creative and variation. And that's not going to limit what channels you can launch on. You're not going to need a tick ads manager to launch on tick-tock with expertise. You are going to be able to think about strategies and scale them out through teams of agents. And
Brett House (33:44.76)
Yeah.
Tejas (33:55.613)
Or those can be internally with a brand. Those can also be through agencies too. think outsourcing and the concept of agencies is not going.
Brett House (34:01.271)
Yeah. The creative versioning agent, the audience audience the audience build agent. I mean they're they're different skills that are very specific. Yeah.
Rio (34:03.391)
Yeah, d it'll depend, yeah.
Tejas (34:05.69)
Exactly.
And our belief is that there needs to be a purpose-built platform that really does this well for marketing. So if you think about in the engineering world, I think it becomes really straightforward. Before Claude code and cursor, engineering AI adoption was tiny. Like I wasn't paying a big bill for my engineers on GPT and Claude at that enterprise level. Prior to cursor and Claude code, no one was consuming. So I think that's what we're seeing in marketing.
Rio (34:30.985)
Yeah, why would you?
Tejas (34:36.967)
Marketers aren't consuming that much because they don't have a harness and app that fits their workflow. And what I've also learned about marketing is it's very heterogeneous. As you were saying, Brad, there's so many workflows. Performance Marketers are very different workloads, Lifecycle Marketers, then Calendar, on brand and ad tech.
Brett House (34:47.479)
Yeah.
Did brand marketers, yeah.
Rio (34:53.161)
And it bifurcated between agency and in house, like it's it's very it's it's very different, yeah.
Brett House (34:56.79)
Yeah. Chan channel breakdowns as well, like channel differences, yep.
Tejas (34:57.597)
It's not.
channel breakdowns, but they all need full business context. They all need full marketing context. They need access to all the data that we've been labeling for CDP and ad tech purposes for the last five years in these enterprises. And on top of that, they need more data than what's in the warehouse. So this kind of is different than what we've been talking about for five years.
Brett House (35:11.021)
Yeah.
Brett House (35:20.588)
Yeah, w which is yeah, which is that it's like third party data it's the data enrichment play, right? They need data from outside their ecosystem, outside their prospect and and customer ecosystem for incremental reach and knowledge of those customers.
Tejas (35:24.999)
Well yeah, they need data enrichment. That's right.
But they also need new types of data that we couldn't process before. So AI actually benefits from looking at every single image and video that you've run on every ad platform before and understanding why it works and why it doesn't, something that a SQL database hasn't benefited from. No one's putting their images and videos into a SQL database. AI benefits from the next set of product brief announcements, product brief documents, 30 page PDFs you have as a brand.
Brett House (35:58.166)
Yeah.
Yeah.
Tejas (36:01.421)
your products you're releasing in the last six months. So we've built integrations now, not just with the warehouses in a composable way, but with dams, with figmas, with all the ad channels, not just to write audiences and capis to them, but to read all the images and videos out of them. And basically we're going from that semantic layer for marketing to now a context layer that covers more than structured data, but also all of this unstructured data across every ad channel, et cetera.
Rio (36:10.443)
Nice.
Tejas (36:29.829)
And then we think that is our of motor secret sauce to build these agentic platforms on top for life cycle and ad tech and so on.
Brett House (36:38.104)
Did you guys talk in in depth with Scott Brinker because he did a paper with Databricks that the about the composable canvas that exactly defines that thesis you just laid out, right?
Tejas (36:42.343)
We did.
yeah. Well, look at our website. We've had the rings for about four or five years now. I'm a big fan of Scott and I told him, was like, he stole our rings, but he added a couple more.
Rio (36:55.177)
He was on the pot. Yeah, we interviewed him on the pot a couple of weeks ago. We we talked about that, among other things, yeah.
Brett House (36:55.212)
Yeah.
Tejas (36:58.685)
I'm joking. We're being dead. You were in the office a days ago. Yeah.
Brett House (36:58.88)
Yeah, it's and exactly the same theory where Yeah, right. And you just described how how like your that that that data operating layer has an expand and yeah a data to context, a much much mu more broad, expanded definition of what data actually means, which is the context of everything from from video to to content to et cetera. So that's
Tejas (37:06.929)
It goes from data to context. Yeah.
Tejas (37:19.461)
And it's funny because the same advantage we built the CDP around, really good tools to understand your data, really good tools to connect to every player in the ecosystem, and really strong back in engineering. Now people call it context engineering. It's the same thing. You need to be able to search really high scale of data and images and pre-tag them, label them. It's all just data engineering again. All the stuff we built our company around still applies. And so we basically repurposed the company to...
over half of the engineers to be working on the marketing context layer, which we believe is our moat in this agentic marketing platform. So it's really expanded the use cases. We have large customers actually generating now ads, launching them across platforms, generating full lifecycle marketing journeys, all through high touch content and data. And we really think there's going to be disruption across that part of the stack.
Brett House (38:13.912)
That's awesome.
Rio (38:13.973)
Well, super exciting. So Tajus, we know that you do have to head out to the data bricks con conference in a couple of minutes and like you are a little s pressed for time, but this is great. This thank you. I would we'd love to do this again. We know you're busy at some at a certain point, maybe do a longer form one. This has been an amazing discussion. Appreciate it. Br I don't know I don't know, Brett, I've been I mean decided right up our alley in terms of stuff we'd like to discuss here. So
Tejas (38:18.151)
That's right.
Brett House (38:34.442)
Yeah, for sure. So so let's let's set up a part two and get even wonkier, right? Because it's it's it because people want to hear on how this stuff is working and you're kind of on the bleeding edge of stuff that we're hearing about from Scott Brinker that we're seeing in the industry. composability has Yeah, and it's got broad implications across various types of tech, not just ad tech. And so this was great. And thank you, T Jess, and enjoy the Databricks conference. One of my partners in my business is actually there.
Tejas (38:45.371)
Yeah, we should do a full episode on the agent side and context.
Tejas (38:57.213)
100%.
Brett House (39:03.639)
right now it should be exciting.
Tejas (39:03.965)
Thank you.
Rio (39:05.085)
Yeah, yeah, I'll be meeting with the whole Databricks crew again next week. You I've I they got a couple of meetings with them. They're great company.
Brett House (39:08.141)
Yeah.
Tejas (39:08.605)
Perfect. I'll be at 10 next week too, so see you all there. Well,
Brett House (39:12.31)
Yeah, yeah, we'll we'll be we'll be housed in the Hearst House for a lot of that time, doing a partnership with Hearst. But so yeah, so for everybody that joined us, thanks for joining us. Visit us at dev.signal and noise.ai. You can find us on YouTube, Apple Podcasts, and Spotify. And thank you, Tejas. It was a pleasure, and we will talk to you again.
Rio (39:12.511)
Definitely see you there. That'd be awesome.
Tejas (39:16.637)
All right.
Tejas (39:31.057)
Thank you, Rio. Thank you, Brett.
Rio (39:32.629)
Thank you.
Comments