← The Monopoly Report
The Monopoly Report
Omar Tawakol, founder of BlueKai, talks about his new start-up Rembrandt that just raised $23 million for AI product placement
00:32:06
Loading summary
Transcript41 lines
- [00:00]
Alan Chappelle
Foreign welcome to the Monopoly Report. The Monopoly Report is dedicated to chronicling and analyzing the impact of antitrust and other regulations on the global advertising economy. If you are new to the Monopoly Report, you can subscribe to our weekly newsletter at Monopoly Market. And you can check out all the Monopoly Report podcasts @monopoly report pod.com I'm Alan Chappelle. This week my guest is Omar Tawakul, who is the founder and CEO of Rembrandt and the former founder and CEO of Blue Kai. Omar and I go way back. He was actually my first client when he was at a company called Revenue Science, which is one of the early behavioral targeting companies. And we'll talk a little bit about his experiences at Blue Kai and what it was, right, like running, I think, the first data exchange. Hey, Omar, welcome to the pod.
- [01:01]
Omar Tawakul
Thanks for having me, Alan. Great to be here.
- [01:03]
Alan Chappelle
Well, thanks. So I, I want to start the discussion back in 2007, because I remember you were out there socializing the idea that eventually became Blue Kai. And I wanted to get a sense of what your thought process was back then because you really stress, at least to me, the importance of the privacy and regulatory environment, even back then, as you were, as you were building out or thinking about the Blue Kai exchange.
- [01:29]
Omar Tawakul
Yeah, and I'll focus on the consumer privacy angle of this, not just the whole business, because if you remember, the data business existed offline in a pretty significant way. I mean, you had catalog companies and credit card companies that had some pretty rich profiles about consumers, and consumers just didn't have much awareness of it. If you could somehow allow any consumer to both stay anonymous, but control what data is known or used about them, then what you could do is you could both get relevant advertising to consumers, but give them the control over what data was used. That was kind of like the founding insight. And that led us to do this thing that we called the Registry, which was, you know, a webpage that you could go to that would show any consumer, even though the service didn't know who the consumer was because it was anonymous, it saw what cookies were available and would let them control it, clean it, change it completely, opt out. And that is something we launched roughly at the same time that we launched the exchange itself.
- [02:38]
Alan Chappelle
I'd forgotten about that. But that's a really good point because then, just to put this into some context, this was 10 years before, you know, CCPA and GDPR were actually mandating that you provide some level of access to consumers around the profiles that you guys were collecting. And, and to this day, you know, not to blow smoke here, but to this day, there's very few implementations of this access requests that the ad tech community has provided. You guys were sort of ahead of the curve and, and sort of remained there.
- [03:10]
Omar Tawakul
Thanks for saying that. I mean, we, we had some hope and some of this hope didn't materialize, by the way. We had this hope that if consumers would engage and control their data or opt out, that what would happen is that we could run a rewards program that would be completely consumer focused which would say, hey, you know, get reward for doing this. So when we launched, we started out by saying, hey, we're just going to give a percentage of our revenue to charity based on, you know, this activity. Thinking that was a placeholder from the future, we'd switch it from charity to actually direct rewards. There just wasn't enough engagement. What happened on the privacy front, which is a little bit unfortunate, is that there's two narratives that we thought could drive consumer behavior. There's actually three narratives. One is fear, which is, oh my God, this is so bad. Turn everything off and run to the hills. The second one is fame, which is, I don't care. I'm going to share my data. I want, I want people to know about me, right? I'm going to go into the platforms. I need fame. And like, you know, this is the cost of fame. Let me feed all these systems as much about me as I can. Not because I want to give up data, because this is just secondary. And the third is greed, which is, hey, could I get something for it? And we were hoping that the, I hate to say greed, because green I don't think is a good thing, but in this case it's a motive which says, let me at least pay you some reward. And we were hoping that third one would be enough to attract people's attention. And as it turns out over the last 10 years, it didn't. It ended up being either fear or fame. Are the motives that, you know, you want fame, so you give data, you're afraid, so you turn it off. And those are the only twos that dominated the airwaves.
- [04:58]
Alan Chappelle
So what can an ad, like an executive team in ad tech, learn from, you know, your failures and successes at Blue Kai? What would you like to impart? Because it's been a couple of years, but I think a lot of the things that you learned were pretty relevant even now.
- [05:12]
Omar Tawakul
Yeah, I'd say there was, like, on the success side, there was one thing that stood out strategically from what we did on Blue Kite. When we started Blue Kite, at that time we were a pure data company not selling ads. And with very few exceptions, everybody else who played in the data space started out with an ad network as a way that used data to monetize ads and then years later shed their ad network and became a pure data play when it became popular. But at that time, 2007, 2008, the vast majority of people didn't do what we did. We did a pure play. And our insight was very simple, that we did not think we could become the number 1, 2 or 3 ad company. Google and Facebook and Yahoo and all these companies had to host so many consumer interactions. We didn't think a B2B company could be number one, but we did think that we could become the number one data company. So we said in order to get there, we didn't want to compete with all the people who would use us by having an ad business. So we were really true to our strategy and we never varied from that. We stuck to this idea that we're going to be a pure play data company, we're going to do the best job we can. I think that served us very well even when all these people thought that we were crazy, that we could have a lot more money selling ads. So I think that stubbornness and sticking to our strategy made us trustable by partners because they, they didn't have to look over the shoulder saying, tomorrow we're going to compete with them. We didn't compete with them.
- [06:43]
Alan Chappelle
That's a great point. Now, I heard a similar comment from the Katz Brothers at mparticle where I think when they first launched there was a lot of pressure to, well, you know what, just do this arbitrage thing and then have your CDP as well. And they didn't. And one of the reasons they didn't is that it would be a distraction. It took you away from the core thing you were trying to do. I'm sure you and Alex and Grant could have made money on some flavor of arbitrage and setting up your own ad network. But query whether that would have gotten you where you wanted to be.
- [07:16]
Omar Tawakul
Not only that, I think a few years later we doubled down on that by saying, hey, we're also going to become a platform so that you can manage your own data and do your own second party deals and not even have to do the data exchange part of business. And that part of the business took off quite well. So that would have even been hurt more if we were in any sort of like, you know, ad serving piece of the business. So that was on the success side and I'd say it was a double success of sticking to what we're doing and then even going deeper with software on the failure side. I think the thing that I should have done now looking back was at the time that Oracle was acquiring us, we had a plan that basically said, hey, there's a future where we do not become data controllers, that we only become data processors, where we become a platform like somebody uses AWS to do something with data. Nobody's telling us what's your position on how that data is sold or consented or stuff like that. They're a compute platform. So we wanted to make this move where we're going to become a pure DMP and basically be the processor, not the controller. And we had a plan on converting our entire business to that. When we got acquired by Oracle, there was a moment where I really wanted to push that and you know, the signal was given to us of hey, we're going to split Bluekay into the Oracle marketing cloud and this new thing called the Oracle data cloud. And we want you to go build out this data cloud and go out acquire companies and build. At that point I should have pushed harder. I, I believe on, hey, I'm okay with that, but I want to only do it after I convert the exchange into a place where other people can do these transactions. But we're not taking a piece of the transaction, were just a software. I floated that it wasn't, you know, nobody bought the idea if I had pushed harder by telling them, I'm telling you, some years from now you're going to suffer if you don't do this.
- [09:27]
Alan Chappelle
So that's a brilliant idea. However, I'm going to maybe challenge that a little bit and I will concede that, you know, the cd, well, not just the cdp, but the data clean room space has sort of effectively done that. You can argue that Snowflake and similar companies have done that. Problem though is that gets really weird because it really handcuffs your ability to monetize the data and make decisions on the data because you can't just be a processor on Tuesdays and Thursdays, like you're a processor every day of the week once you're a processor. And it gets really hard with what California has done with the service provider role because it's really hard to even have a unique ID across multiple companies as a service provider. And so I sort of both agree with you and disagree with you a little bit.
- [10:18]
Omar Tawakul
Look, it is very hard for sure we would have foregone revenue, there's no question in that. But the time to make a change like that is at a point where you've done this acquisition, you've pushed the software revenue into the marketing cloud. The data revenue is less than the whole. You haven't acquired five other companies yet. You bite the bullet there, you take a reduction in revenue, but you build the long term vision that says, okay, this can go on for 20, 30 years. You get it right from day one. We can build this up and we can become the foundation for future data, clean rooms for second party data deals and really be the infrastructure for the future. You would have made less money for a little bit of time, but over the long run you would have had a much more sustainable business. And there's still some problems that I'd love for us to discuss. Like you got into it, which is how do you handle identification across companies, data sharing across companies? I'd love to lean into that. So there are hard problems to solve, but they're worth solving. And the reason we know it's worth solving is they ended up shutting the business 10 years later. So obviously you know, and, and made a lot of money on the way, but all of a sudden now they're going to make no money. So I think, I would argue it's worth it.
- [11:37]
Alan Chappelle
Fair point. So let's lean in on that. Like, I mean, and this may be hard to do over a, in, in a three minute sound bite, but how do you, how do you achieve that goal where you, you need to have, in order for your base business to function, you need to have a uni, a unique ID that effectively gets used so that you can recognize the user across multiple customers. Like, how do you solve for that?
- [12:00]
Omar Tawakul
Yeah, well, you know, I happen to be on the board of a company called Liveramp, which I don't want to speak for them, but obviously they have a perspective on how to build identity that allows you to weave second party data deals. And in many of these second party relationships, both parties have some sort of identifier that they would use to hash to be able to communicate. And if each side has opted in information from the consumer, then that could create a handshake.
- [12:34]
Alan Chappelle
Fair. Yeah. I'm familiar with LiveRep with some of their processes, so that makes some sense to me. And maybe that's a route for larger ad tech to head down because we need to figure some things out given how confusing right now the regulatory environment has gotten. But before we go too far, I want to point to one Other thing that you guys did really well, and I haven't been able to find another venue to share this with, so I wanted to share it with you. Now, maybe everybody here doesn't know, but I was effectively your privacy guy for a number of years and, and, and I got brought in on a whole bunch of sales calls, some with very, very large companies, with multiple decision makers. And the thing that you guys did really well, and I'll give you credit, and Grant Reese and, and certainly John Sedlak, is every single time I went in, I understand who was in the room, what their objections were, what the questions they wanted us to answer were. And having that type of information allowed me to walk in there and sort of make it look easy because I knew exactly what buttons I had to push. Contrast that to how it typically works in an ad tech setting where they literally bring me or some other privacy person along with somebody's dpo, and it's almost akin to throwing two kitty cats in a small room and hoping that it's going to work out. And usually it does it without a fair amount of pain and a lot of hard work. And you guys did that one thing. Again, not to blow smoke, but you guys did that about as well as it's been done. And as a privacy person, boy did I appreciate that.
- [14:06]
Omar Tawakul
Yeah, I appreciate that. I think we thought of you as an extension of our team. And so it's just the kind of prepping your team on how to get stuff done. I didn't really even realize this was kind of differentiated in any way, but privacy was so important to us, being able to do what we do. So we had to spend a lot of time with it. And I kind of personally enjoyed the category because there was a bit of a geek inside of me hoping that we'd get to a better place where what we're doing is stuff that you can be proud of. Privacy is very difficult. Needle the thread.
- [14:43]
Alan Chappelle
Yeah, sure is. So I wondered if you had any thoughts. We saw what happened to Oracle over the last. What is it, six months ago. Do you have any thoughts for what made them close the Oracle marketing cloud? You alluded to a couple of the issues, maybe, but was it, in my view, as I read through that class action complaint that seemed to be at some point as a. When you receive one of those complaints and see you litigate them for a couple of years, at some point you just sort of lose the will to live and you just decide to move on. But was it that simple or is there something else going on in your.
- [15:16]
Omar Tawakul
Opinion, I wasn't there anymore. So all this is going to be speculation on my side. But I think if you're Oracle and there's this huge cost on the litigation front, but more importantly, Oracle has to have the trust of the world because they're the database. So it's the, the impact to your brand isn't worth it. It's not just the finances. That's why I made the statement I said earlier on I wish I had pushed the controller thing because I think that would have been more consistent with their brand, which is secure, trusted, high volume enterprise, something that all of a sudden gets lawsuits and money associated with it. Even if you can make money off of the business, it's not worth it.
- [16:04]
Alan Chappelle
That's a fair point. So I, I probably should have started with this but, but congrats on closing the round with, with Rembrandt. I was really, I was excited to see that because I love that, that you've moved on and every, every company you've been involved with has been successful. But I was also pretty psyched to see you partner with Tom Vivek because I, I had worked with them as well and I remember Tom posted this on LinkedIn a week or so ago, but I remember talking to him about you. This is after you had moved on to, you know, been purchased by Oracle and, and, and he could not have been more respectful. Was like, man, that guy was a tough competitor, you know that son of a gun kind of a thing. But like, boy, there was mad respect to see you guys sort of partner. Really excited.
- [16:50]
Omar Tawakul
I'm really excited about this. I also have that mad respect for, for Tom and Vivek. They have a chemistry that is amazing and they've, you know, built teams, they've sold multiple times and you know, I, I know what that looks like. Having worked with, you know, people like David and, and Corey and John Sudlak across kind of multiple companies that, that have sold. And so I know what that looks like and they were a great competitor and I actually knew and respected Tom even before Crux. He had a company that sold to Microsoft that cracked a very interesting problem to me personally. It was strategic pricing and so I knew that this was somebody who was a very deep and thoughtful person. And I saw this twice in my career where I had a competitor that I was fiercely competing with. But as soon as the competition was off and I had a chance to drop the gloves, we became friends and I considered them a mentor. Dave Morgan was the other person who in the revenue science Days we were competing with him and then we became friends after that. And I've always respected and followed, you know, what he's done in Tom Chavez and Vivek. And so I'm going to be working a lot more closely with those guys. Love to have their help. What we're building Rembrandt, is just so interesting and new and has a very different set of, like, challenges and really nice market.
- [18:18]
Alan Chappelle
Well, can you frame what is Rembrandt? Because I don't know that everybody in my audience is going to know that.
- [18:24]
Omar Tawakul
Before I tell you what we do, let's tell you a problem that we solve because it's a very cool one. Programmatic industry has been growing like gangbusters, you know, since, you know me and you started in it from the very beginning. But the one thing that has happened as we've gotten bigger and better at getting the right ad to the right person at the right price at the right time is consumers have figured out how to run away from ads. So, like, a lot of the people are going to be listening to this podcast. They're going to go home at night and they're going to watch Netflix or Amazon prime or some of them are going to pay for, you know, YouTube, the premium, so they don't see ads. So many people, particularly in the ad industry, who spend all the money allocating the budget for advertising go home and don't see the ads. And the younger folks who might not want to pay for all the premium services, they hit the skip button as soon as they can. So the long term sustainability of this means stop betting against consumers. How do we find a model that consumers are actually willing to engage in? That's what Rembrandt's about. Rembrandt's about taking content that people love, like a movie, like a podcast, like a, a TikTok video, and inserting the brand into the scene using AI such that it looks like it was filmed in the original. So it's taking the product placement industry that's existed for decades, where you have to spend, you know, two years before the movie gets out. You negotiate to get, you know, a Pepsi can in that scene. Here we've basically flipped it to say, you don't have to do that anymore. You create beautiful content, great content that the audience loves. And right when you're about to, to launch it, you can insert a brand in different places at different times to different audiences, and do this with the scalability of advertising as opposed to the slow bespoke movement of sponsored posts. And product placement movies. We can do this at the speed of compute.
- [20:25]
Alan Chappelle
So how do you see that space, just the broader AI space evolving and then what's your role within that broader space?
- [20:32]
Omar Tawakul
I think Terry Kwaja does a great job talking about everything until now in our ad industry has been about efficiency, has been about, you know, squeezing in better optimization, better targeting, better pricing and bidding and header bidding and all that stuff that we've been occupied for years and years and years. But what was unaddressable on software was the creative itself. People go out and they spend all this money and time creating creative and then they launch it. Now with AI, we can go into not just efficiency, but effectiveness, which is a lot more space for us to perform better by changing what ad gets shown to be person and changing the format. Like in our case, the format is the actual content you're watching. How can we put a product in there? How can we get it noticed? So our focus in AI is opening up the creative space so you can scalably alter how a brand appears to an audience and do it a way that drives performance and attention and positive engagement.
- [21:36]
Alan Chappelle
Makes sense. And I think I know the answer to this question. But just to get this off the table, are you guys, do you have any plans to move into like the, the profiling and data collection space or are you just purely an ad insertion?
- [21:49]
Omar Tawakul
Yeah, no. At Rembrand, I'd say we think of us in three different ways. We are really spatially aware augmentation of video and we happen to be applying spatially aware augmentation of video to product placement or virtual product placement. And we are currently doing that as think of us as like an AdSense for virtual product placement where the brand buys the campaign from us, the creator comes, puts in the object and gets paid. So it's a two sided network. And that's the business we launched with as of Q4. We've now expanded to say now we're software and you don't have to buy the ad from us, you could just use our software and the advertiser can use our software to deal with the creator or the creator can sell to the advertiser. It's just like DoubleClick was a network and then they became DFP for publishers and DFP for advertisers. We've done that, we've already established deals for that and we've opened that up to be able to do it inside of ctv. So we're playing these multiple roles as being the fundamental technology for video augmentation. Started out with paid. But now we're going to allow for like organic use, we're going to allow for free use of it. We're just going to let a thousand flowers bloom.
- [22:59]
Alan Chappelle
Touching on kind of the impact of big tech because that's one of our focuses here at the Monopoly report. And you know what you guys are building, you know, I see, you know, Meta certainly has their AI enhanced ads offering. I'm pretty sure YouTube's got one as well. Do you see what you're building as something that is complementary to those types of offerings or competitive? And like how do you see your relationships with, with those companies evolving over time?
- [23:25]
Omar Tawakul
It's very complimentary because our attack on this is very different than any of what they're doing. A lot of what they're doing is using AI techniques to create ads that are interruptions. So we don't do that at all. And what we're replacing is the manual industry of, you know, a brand finds a creator, they ship them some goods, they sign a contract and they take four to six weeks of this process of negotiating what they're going to say. That's what we're replacing by saying, hey, we can automate that into something far more scalable that looks like an ad campaign. And as we build this as an API, we're also going to all the platforms and saying you can embed this API, it's still you who would be, you know, Google or YouTube or Facebook or TikTok. It's still you selling the ad. This technology is just about the photorealistic augmentation in it. What they're spending most of their money is, is not that it's generating ads from scratch.
- [24:27]
Alan Chappelle
So you may not have a horse in this race, but I'm just curious because you're so thoughtful on this stuff, how do you see the LLM space playing out in terms of privacy or. Well, just. I'm going to leave it open ended.
- [24:38]
Omar Tawakul
Yeah, I mean, I think the most interesting thing that happened just happened a couple weeks ago, which is we've all been looking at the space saying there are billions being spent training these LLMs, so stop trying to compete with these guys. The real differentiation layer is to build on top of them. And what you do there is you build some sort of agentic workflow that has your own unique interaction data that isn't found in the LLMs that goes deeper in some sort of workflow. And I fundamentally believe that is the strategy of 2025. So I really believe that. And we are kind of parallel in Some way we don't compete with LLMs. We take our own interaction data and figure out how to augment real world videos, not generated videos or human generated ones. So it's a very different technology. I think the thing that shocked me a week ago was this Chinese company, I believe it was called Deepset, that trained an LLM with $6 million in open source capabilities. And it is damn good. And it really challenges this idea of hey, the game was over, you have to spend billions to be able to do something here. So I don't have a firm conclusion there. I'm just surprised and I think it's notable. Having said that, other than this piece of evidence that I just showed, I do think that you're better off at the agentic layer and doing something that's differentiated there.
- [26:05]
Alan Chappelle
So when the US created the atomic bomb, there were scientists who were predicting that the Russians would not have one for 50 to 100 years. And I forgot what it took them three years, four years. So, and there's sort of an analogy at play here. You just wonder if somebody is going to figure out a way to completely upend the model and all that money spent on LLMs, you know, not to say that went to waste, but it may not nearly be the differentiator that folks think it is.
- [26:34]
Omar Tawakul
Yeah, yeah. Having said that, I do think that, you know, ChatGPT already has a brand advantage that was similar to the one that Google had in the early days. Like they have a bunch of people paying them to do it. I have shifted my search behavior to ChatGPT. I am not even really searching. I'm communicating with it back and forth, turn after turn to really fine tune what I'm looking for. So I think their position, even if people find cheaper ways to do it, they will also find cheaper ways to do it. And having that many consumers and enterprises, I think they're relatively secure. And what's crazy about the market on this kind of stuff is sometimes you can overhype the financial aspect of it. And so you have a correction. But that doesn't change the fact that the fundamental thing is still valuable. Or if you remember the Amazon Bomb and Amazon.com speak that they had in 2001, they were right. We overhyped a bunch of stuff, but they were wrong in that Amazon was fundamentally on a long term path that was good. I think the same thing that's going on here, there's too much hype on LLMs. Some of it we're going to prove are going to be investments that will not pay off. There will be a bubble bursting. But there are people who have developed fundamental tech that are, are just going to prove that regardless of how you price it, it is fundamentally going to impact everything.
- [27:57]
Alan Chappelle
So that makes sense. But we haven't really found a business model yet. Like, they're making some money in subscription. There's some folks like perplexity or playing around with ads. I mean, you seem, you're shaking your head to say, Alan, you're focusing on the wrong thing. Like, what, what am I missing here?
- [28:16]
Omar Tawakul
They're making money on. It's just like Amazon. They're making money. They're just investing so heavy into the future that, yeah, they're burning crazy amounts and maybe the market will force them to burn less amounts, but people are paying them a lot of money for subscriptions. The fact that they're saying, you know, hey, we're losing money off of that. Yeah, of course they're losing money. The market is telling them, go spend, figure out AGI. And yeah, they're playing hype on top of it. And you're using that hype to, to get better stock prices and better valuation and do this unfortunate flip from, you know, what they were for, for profit. I'm not so excited about that. I think that was very unfortunate. But the market's letting them do it, so of course they're going to, they're going to spend that money until we tell them they can't. But even after we tell them we can't, they still have people paying them real money today and that's just going to increase.
- [29:12]
Alan Chappelle
Okay, fair. So I would love to. Before we wrap it up, I'd love to get your take. It's now January 10th. We've been waiting for at least a couple of weeks for the decision in Google's ad tech trial. And like, I'd be curious for your thoughts on how you see all of this shaking out.
- [29:29]
Omar Tawakul
I really hope that we allow TikTok to continue in the U.S. it makes no sense that we wouldn't. The level of hypocrisy that we have, you know, a bunch of US companies that control the data for the planet and then all of a sudden China's the demon because they're doing something we could do. It's a level of hypocrisy I literally don't understand. Like, I don't understand how a person with an intellect can really believe somehow that what we're doing is holy or what they're doing is evil. Now, that doesn't mean we don't put guardrails because, you know, they control their country. Like China doesn't allow Facebook to do a bunch of stuff in there. I don't mind us putting some guardrails, but I just don't enjoy the hypocrisy behind it, which is somehow we're good and they're evil. I just don't buy that. And the rest of the world doesn't buy it, by the way.
- [30:17]
Alan Chappelle
Well, and the other thing is there's only so many times that the DOJ gets to play the national security card without revenue revealing any of the background. So maybe there's something, maybe there's some kind of smoking gun there, but I, I don't see it. And there's really very little direct evidence. And if you're going to ban TikTok, then you should probably think about whether or not you're going to allow, well, Meta or, you know, any of the other social platforms, because all of those issues are, are the same essentially. It's just, it's just who gets to do it.
- [30:48]
Omar Tawakul
Right, right. And then what I wish they would do is figure out how to make sure our, our kids can't spend infinite time on TikTok. Like I remember there was this video that, that, that played about a year ago where they showed the version in China had more restrictions for younger kids on what they could do in terms of time spent and what content they saw. I'm like, sure, yeah, let's do that.
- [31:16]
Alan Chappelle
Yeah, couldn't agree more. I've got a five year old at home who already knows how to use the phone and so good fun. We're going to wrap it up there. But I just want to say, Omar, thank you so much for coming on and chatting with me.
- [31:27]
Omar Tawakul
Always very thoughtful, always love speaking to you, Alan. Thank you.
- [31:32]
Alan Chappelle
That was a great conversation. It's always fun to talk with Omar. We've got a bunch of other fantastic guests coming up on the Monopoly Report podcast over the next few weeks. Next week we've got Julia Shulman and we're going to talk about the CTV space and some of the impacts of Google's recent flip flop around device fingerprinting. So look forward to that. Please subscribe to the monopoly report@monopolyreportpod.com or on Spotify, Apple, YouTube or wherever you listen to your podcasts.