AppNexus Founder Brian O'Kelley - adtech, AI, advertising and more AI

Brian O'Kelley is the co-founder and former CEO of AppNexus, which built core infrastructure for programmatic advertising and was acquired by AT&T for $1.6 billion in 2018. He's now the co-founder and CEO of Scope3, a company working to reduce the carbon footprint in digital advertising. In this episode of World of DaaS, Brian and Auren discuss: The zero-sum game of auction paradigmsAI tools revolutionizing software developmentBuilding personal brands in techClimate tech's shifting busine...

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:02Auren Hoffman?

  Hello, fellow data nerds. Welcome to World of DAS. I'm your host, Orin Hoffman, CEO of Incubate and GP of Flex Capital. Discover more episodes, get weekly data as a service news, original content, articles on data and more at worldofdas.com. That's worldofdas, D-A-A-S, dot com. Hello, fellow data nerds. My guest today is Brian O'Kelly. Brian is the co-founder and former CEO of AppNexus, which built core infrastructure for programmatic advertising. It was acquired by AT&T for $1.6 billion in 2018. He's now the co-founder and CEO of Scope3, a company working to reduce the carbon footprint in digital advertising. Brian, welcome to World of DAS.

• 0:45Speaker 01

  Thanks, Orin. Thrilled to be here as a data nerd.

• 0:48Auren Hoffman?

  Yes, you're definitely a data nerd, for sure. And you and I have been friends for a long time, so I'm really excited to dive into all these things. Now, what do you think is the thing people most misunderstand about how digital advertising works today?

• 1:02Speaker 01

  I think that there's a lot of folks who are younger than us who don't realize that programmatic and this idea of auctioning ads was totally arbitrary and that it's just one way that you can match buyers and sellers and that there's tons of different ways to think about this. And so anytime there's a dominant paradigm, there's a skeptical part of me that's like, somebody came up with that. And that doesn't have to be this way. In this case, it definitely was somebody who came up with it. And I love thinking about different ways to solve the same problem with different tools, algorithms, ideas, concepts.

• 1:38Auren Hoffman?

  What would be an alternative universe? The classic would be just how people would sell it on a site, just like people sold it in magazines and stuff?

• 1:46Speaker 01

  I look at Amazon and how they do sponsored links. I want to go sell my book. I go give them $1,000 and I'm expecting to sell more books. And if I don't, then I don't give them more money. I want to tell them who my audience is. I want to tell them the people I think will buy my books. I'm data nerds. The idea is that I want to contribute as much information to help them find me my customers. And then at some point, the algorithm will pick it up and find more people like that.

• 2:13Auren Hoffman?

  Yep.

• 2:14Speaker 01

  The problem with programmatic as a paradigm is auction causes buyers and sellers to fight. It's a zero sum game. Who pays more? Who pays less? And it creates these weird dynamics where I don't want to share information with you because you'll charge me more. Or I don't want to share more because you'll just arm my inventory. And if you read ad tech press, it's always about buyers versus sellers and publishers are getting screwed. This whole thing is about the economics of the transaction, not about making the pie bigger. And so to me, any paradigm that's zero sum is missing the point.

• 2:46Auren Hoffman?

  In some cases, you have people on all sides of the ecosystem. So Google would be an example of that. And I know that you testified as an expert witness in the DOJ case against Google. If you had to say the strongest case for their behavior, anti-competitive behavior, what would be the case that you would make?

• 3:04Speaker 01

  I think that the challenge of any market leader is that there's a tension between what makes you the most money as the market leader and then what's best for the ecosystem. And so I'll use what's now called GAM, Google Ad Manager, but was previously DoubleClick for publishers. The whole point of that tool was to help publishers monetize their inventory. And if you look at that product and you say, what are the things that were best for publishers? Well, if you look at the tools that Meta has or that Spotify, just tools that publishers build for themselves are dramatically better than the tools that Google is building for publishers. And so something's broken. Why are publishers using Google's tools if they're not the best possible tools? Why is optimization not built in?

• 3:49Speaker 01

  Why doesn't it have the ability to ingest a Cappy, like a conversion API? Why is there not more order management? I mean, Google shut down their order management product. Google actually made the product worse over time. And so this is when you start wondering if there's self-interest. And of course, Google wants to sell as much as possible through its ad network, through its programmatic platform. I mean, the margin on programmatic or their ad network is dramatically more than the pennies they get for the ad server. And so I think that's when you start wondering what structurally is either preventing a competitor from building a better platform or keeping Google from investing their talent to make that platform more effective. And so that, to me, is the smoking gun.

• 4:31Auren Hoffman?

  And why do the publishers stay on that platform if they are in some ways gonna have it worse off? Is it because they get better access to inventory from Google?

• 4:43Speaker 01

  News Corp was an investor in AppNexus, and so they had significant reasons to want to migrate. And they said that when they talked to Google about migrating, Google pointed out that 50 or 60% of their programmatic revenue came from Google and that they would be putting that at risk if they switched. And what News Corp said was, we can't take that risk. Even though we know you have a better product, we know that we're investors, it's just too risky. Because what if we lost that Google money? We would be completely screwed.

• 5:13Auren Hoffman?

  Got it. Okay, that makes sense. Now, ad tech has been a great place to invest. You and I have done a lot of investments in the ad tech world. I think investors have made more money in ad tech in the last 20 years than any other category except for SaaS. Why does it get such a bad investment rap?

• 5:29Speaker 01

  Well, the problem is that ad tech is very difficult to understand from the outside. And I think the SaaS mantra of CAC and LTV, and there's these really simple-like things like companies should have a burn ratio of XYZ. It tried to break companies down into very simple machines. You hire a growth person, and you do this, and you should have 52 employees at this revenue level. What you and I both know is that there's so much more Complexity to building businesses than just hitting certain multiples. And ad tech almost exposes the fact that there's agencies that work with brands and that agencies need to be managed in one way and then brands do other things and there's global markets.

• 6:13Speaker 01

  And being an ad tech company means you're in one of the most crazy, global, complicated, fast-changing markets in the world. And to me, if you can survive in ad tech, you can do anything. I think it's a lot like Wall Street. There's more complexity on Wall Street than any place in the world. And it's scary to investors because they don't get the market. And therefore, they're trusting entrepreneurs to build fantastic businesses and to do things that make sense. And they all get confused about net margin versus gross. Meaning, I can grow a business to a billion dollars really fast, but I'm only taking pennies on the dollar. Or I'm running some kind of arbitrage scheme that hits the limits of the market. Or it's just like anything else. It's complexity is scary.

• 6:57Speaker 01

  And I do think investors are waking up to the idea that building a buy the book SaaS business has caused massive expansion of SaaS apps that are going to get made obsolete by AI in many cases, or get washed out by a down market. Ad tech's not like that. And I think there's a lot of value still to be created in the ad tech universe.

• 7:20Auren Hoffman?

  I think there's two things that do make it hard in ad tech. One is that just there is a lot of smart people with a lot of competition. Other people are constantly trying to like, if you find something, they're going to get smart and they're going to try to find it as well that's out there. And then the second thing is that just the whole ecosystem changes every five years. So I think that's one of the reasons I love investing in it because you can take market share really quickly. But in some ways, it's better to be an angel investor in ad tech than a series B investor because you see these companies that are 50 million and then they grow to 100 and then they go back down to 50 like really fast. And so you almost want to be the investor where you're investing at like 1 million in revenue.

• 8:03Speaker 01

  Sort of. I think it goes to the problem is there's reasons that that happens, often because you're really good at selling and you have a great execution and then you're in a niche in the category. You've found a niche, but the niche is only 100 and it's so juicy that everyone comes after you. But then you have the breakouts. I think this is the same in any industry where AppNexus, if you invested in our Series B, you would have made, I don't know, 50 times your money or something like that. It would have been an extraordinary investment. And I think we see that with Trade Desk, which barely raised money. You've got Criteo. You do have these breakouts where they found a category, were able to defend it and then crush the competition. And so to me,

• 8:45Speaker 01

  that's a combination of great entrepreneurs and then really savvy strategy to understand both the size of the niche and the structure of the organization to take advantage of it. So I don't know. I think there's some great investments out there, but I think in general, it's not just that you should invest early. It's that companies aren't going to be massive, that you should be planning for companies that get rolled up and not necessarily get to IPO.

• 9:10Auren Hoffman?

  Yeah. In antic, there's so many companies that get acquired for between 100 and 400 million dollars. And there's so many of those companies where actually in SaaS, it's like a little bit weird. So either you have the multi-billion or it's a zero in a way. You mentioned Trade Desk. I love Trade Desk. I don't totally understand why their stock is down over 50% this year. Can you break it down for us a little bit?

• 9:35Speaker 01

  Yeah. So the question is actually, why was Trade Desk worth so much?

• 9:39Auren Hoffman?

  Yeah. Okay.

• 9:40Speaker 01

  Their stock went from 50 to 135. And what happened was that they just hit their numbers again and again and again and again and again. And at some point, I think investors were just like, wow, there's no end to this particular rainbow. I'm just going to keep riding it. So I think actually, If you look at this stock historically, there was a bubble that burst, but it didn't change the fundamental nature.

• 10:05Auren Hoffman?

  Yeah, it's still a 26 billion-ish market cap company, which is incredible.

• 10:10Speaker 01

  If it had never gone to 135, we would be like, oh, my gosh, Jeff Green's a friggin genius.

• 10:15Auren Hoffman?

  Yeah, which he is, by the way.

• 10:16Speaker 01

  Yeah, totally.

• 10:17Auren Hoffman?

  Yeah.

• 10:17Speaker 01

  I'm just saying, the issue for them is that once you seem vulnerable and you seem human, then the market is trying to ask a lot of hard questions. And so I think that's the downside is that you look at the stock coming down and people are asking, why is it down? And then they want to go, well, wait a second, what happened with your TV platform? And then what happened with this?

• 10:37Auren Hoffman?

  Yeah, it's really about the bubble, not about the pop.

• 10:40Speaker 01

  Yeah. And so if they stayed on the plan, it's an interesting lesson for all of us, which is there's something to be said for humility and just being properly valued versus trying to justify a crazy multiple.

• 10:54Auren Hoffman?

  Yeah, yeah, because you can get in this weird treadmill that goes in there. Now, the monster ad tech company is Applevin. And it's weird because like, a lot of people in ad tech don't even know Applevin. At least as of today, their market's worth $93 billion. What did they do right?

• 11:12Speaker 01

  I think performance is a different market. It's not ad tech in the same way that we think about it. And they're a performance company. And performance is such an obvious category. Mobile gaming, which is gamers game, just selling games to gamers and using that ecosystem. It's just a massive market. Gaming is a big market and gamers want to gaming ads and virtual currencies and all these things. It's a different world. And I'm looking at TikTok shop and a lot of the live commerce stuff as being a similar new category where I don't get it. I wish I did. And I don't think the ad tech companies of today are well positioned for that for a lot of the same structural reasons that we don't get.

• 11:57Speaker 01

  Apple of it because it's going to operate in a totally different context. Basically, it's not a walled garden per se, but these are platforms that are almost self-contained. And we, as older people who've been doing this for four or five decades, it feels like, we have to break our brains and think about a world where the surface is no longer web-ish. It's not web-like in any way. Mobile games are not like the web at all, and neither is live commerce. You have to break your brain to understand that concepts don't correlate. That makes any sense?

• 12:30Auren Hoffman?

  Now, you recently joined the board of LiveRamp, which is where I used to work, and it's a public company. And I know you, you don't seem like the kind of guy who'd love being on a public company board. There's a lot of process and stuff like that. Like, what made you want to join a public company board?

• 12:48Speaker 01

  First off, AppNexus was almost public. We had filed our S1, we had through comments. I guess part of my question for myself is, could I stand being a public company CEO? Would I get fired after one day or just get fed up? And so, how do you know what it's like to be a public company CEO without being one? So, I really admire Scott Howe because he's been through everything as the CEO of LiveRamp. And I feel like that board, which has John Patel and, but has Omar Tawakal, and it has just these amazing people, it's a great way to learn in a place where it's totally different from my everyday experience. It's a big company, global company, but I get the tech.

• 13:31Speaker 01

  I don't think there's that many people who can understand a company like LiveRamp. So, I can actually be valuable.

• 13:38Auren Hoffman?

  You're one of the very few people who probably could really be valuable there.

• 13:41Speaker 01

  So, that's my goal. Could I be helpful? And I don't know if I help all the time, but at least I can understand some of the pieces. I love the team. I think the LiveRamp leadership team and most of the employees I've met are just awesome people. It's a heavily regulated environment that's changing a lot, and so I've learned from that. But I'm always trying to learn, and I felt like it was a new experience for me. And yeah, I would love to keep finding new places where I could push my limits. Following in your footsteps is a great way to do that.

• 14:12Auren Hoffman?

  Because like I'm still a shareholder and obviously we've had like a not a great stock performance over the last five or 10 years. It's basically been flat. The stock's been flat since I left. It hasn't moved at all. But as a fiduciary for the board and the shareholders, how does one in a board meeting? I mean, obviously it's a tough company. There's lots of things going on. But how does one have these conversations internally? It's not like it's up and to the right. It's a slog. So how does one have these conversations internally?

• 14:41Speaker 01

  Well, what's interesting is the company has grown dramatically. If you look at the revenue and the scale of business over that time, I mean, they just keep growing and building this company through super chaotic, crazy market times. So if you think about the strategic value of being the data pipes for a trillion dollar industry, I guess the conversation is partially looking like, well, how do we go take advantage of and grow the foundation that we've built? The headwinds have been cookies and privacy regulation and not having consistent national privacy regulation and not having really a final answer to cookies has made it really difficult to see the value

• 15:23Speaker 01

  of what I think is a extremely ethical organization built on trying to be privacy compliant in every single place versus a market where if you just throw some crappy fingerprinting, you actually get more spend because you're not restricted to what is, I would consider ethical. So it's almost like a negative ethics arb that feels like we're dealing with. This is not just for LiveRamp. This is broadly as an industry. I don't think we're properly valuing data. I don't think we're properly valuing data pipes. I don't think we understand the incredible value of connecting people to media. And I look at Wall Street where data is the most important asset.

• 16:06Speaker 01

  It's like data plus quant equals profit. I don't think that's the case here. So there's a structural disconnect I do think will resolve itself. And the answer to your question is, a lot of board meetings, we look at ourselves like, we're doing all the things.

• 16:22Auren Hoffman?

  It's just because somehow, over the last decade, they just haven't been able to tell the story? It's a marketing problem in a way, or is that the thing you think these things are?

• 16:32Speaker 01

  I think it's the Warren Buffett quote. When the tide goes out, you'll find out who's wearing shorts. And we've been in the longest, craziest bull market ever. We have 17 years where even through COVID, we've generally seen markets go up and to the right. You've seen digital advertising grow dramatically again and again and again. And I think we're due for a correction, and I mean that in all senses. I think it wouldn't be a bad thing. I don't want crazy global market crash, but I don't think it's a bad thing, actually, to get a little bit of, let's pull back, let's see what's really here, and let's rebuild on a better foundation. Look, you can always tell the story better. The numbers can always be better.

• 17:16Speaker 01

  There's always things you could do, but I would just say, if you're to grade performance and you're to grade market conditions, really feels like ad tech, and this is the VC side of it, it's the public market side of it, with the exception of Trade Desk and to some extent AppLovin. I don't think people are fully understanding the opportunity broadly in the advertising sector.

• 17:38Auren Hoffman?

  Got it, okay. And you don't think they understand it just because it's so complicated? What's the reason that you think people understand? Because obviously, investors are smart, they're looking for underpriced things and stuff.

• 17:50Speaker 01

  Yeah, well, again, I can only speculate, but my sense is the continued questions about cookies are really challenging. To me, cookies are the tide. And if the search trial means that Google has to spin Chrome, will an independent Chrome have the same perspective on cookies that Google does? Because most of Google's cookie decisions are based on CMA and other regulators. An independent Chrome might look a lot more like Firefox. And if that's the case, then we would probably not have cookies on by default in the world's biggest browsers.

• 18:25Speaker 01

  In an AI future, maybe that doesn't matter because everything's just happening in the AI chat anyway. So to me, what's the bet? The bet is we will survive and I'm going for survival. And then I think the second bet is, is there going to be value in being the data pipes that connect marketers to all the major platforms and publishers? I mean, that seems pretty darn valuable. That's my take.

• 18:48Auren Hoffman?

  Yeah, for sure. Okay, I love it. Now, the implications for AI for creating ad content is clear. Just anything or content related with AI is going to be great. Where else do you think AI is going to play in the ad tech ecosystem?

• 19:04Speaker 01

  What we've been doing at Scope 3 is basically imagining the thing that I think AI is so good at, which is taking a prompt and then using that to understand things. I can say, hey, tell me about the whole web and it will go out and search and do things. I think things like content classification are perfect out of the box use.

• 19:27Auren Hoffman?

  Yeah, that's a perfect thing.

• 19:29Speaker 01

  So we just said, well, why don't we just do this in the most efficient, sustainable way? Why don't we just ask agencies to tell us what their brand is trying to do and then run that on the internet? And it's extraordinary. It's incredibly fun. If you're an agency person, all of a sudden you're a prompt engineer. And if you're a brand, now all of these complex things you're trying to balance across your whole portfolio of needs, you're actually putting into an agent to do it. So what actually we built was less the agent part, but it's more all the things around it. How do you calibrate that agent to make sure it's making good decisions? How do you get transparency to understand what it's doing? How do you give feedback to the other side, to the publishers and content producers?

• 20:11Speaker 01

  So it's almost like the wrapper around the agent. And to me, that's how we have to start thinking. It's like, how do we create the conditions for AI to be incredibly successful? And then how do we help humans get really good at instructing agents? So I think that's a paradigm shift. I can see the same thing happening in the broader media decisioning world. Planning is obvious and every agency is doing it. There's workflow things, but I guess I'm more skeptical about workflow than others because, yeah, okay, fine, I can have an agent read my email, look for someone asking for a meeting and then send back my calendar. But the thing is deciding who to meet and how to prioritize that meeting. I don't know.

• 20:54Auren Hoffman?

  Yeah, it's very hard.

• 20:55Speaker 01

  That's my most valuable resource in the world is my time. So why would I outsource the thing that's most important? I don't know. So I just keep hitting that wall of what are the things I do today? I want those slacks in my brain. I don't want AI responding to my slacks because I'm trying to learn from what my team is doing right now. So I don't know. I'm just skeptical about that for some reason.

• 21:18Auren Hoffman?

  Today, the way a lot of people are just interacting with AI tools is through some sort of chat interface, through some sort of quote unquote prompt. And it does seem like there's a very specific type of personality that does that well. And today, that is the dominant AI UI is the prompt. And they think that that only works for a very small, under 10% of the population who's good at that thing. So either more people are going to have to get good at that thing, or we're going to have other UIs that come in of how we should interact with these AIs. Where do you think we're going to go?

• 21:59Speaker 01

  I always use my wife as my favorite power user of all tools. And I mean, power user as not the most power user.

• 22:06Auren Hoffman?

  She's not a software developer like you?

• 22:08Speaker 01

  No, she's a psychologist. It's like even worse.

• 22:13Speaker 01

  What she does is she writes a document in Word or whatever, and then she copies and pastes it into ChatGPT and then has it clean it up and fix the formatting and all these things.

• 22:22Auren Hoffman?

  Yeah. And then you have to go back into Word.

• 22:24Speaker 01

  Back to Word.

• 22:25Auren Hoffman?

  Yeah, which is so terrible.

• 22:27Speaker 01

  So frustrating. And so I was like trying to tell her to use Google Docs and use Gemini built in.

• 22:33Auren Hoffman?

  Even within Gemini, it doesn't really work that well. It doesn't show the edits within it. Well, show it in the box and then you have to click yes, and then it changes everything. I wanted to change like 90%, but there's still a couple of things I want, and it doesn't allow you to edit. It's still not great. Yeah.

• 22:46Speaker 01

  Right. And that's my point is, what is it, side by side? Now we have a diff while you're redlining.

• 22:52Auren Hoffman?

  It should be like an editor and then you could accept or not accept the edits that your friend editor does.

• 22:58Speaker 01

  Right. And then she's like, well, can you export this to Word? But it only took the first 500 words for some reason. And so she's so frustrated with this thing. And it's like, it's so powerful, but what I like using it for is I was working on an investor deck. I was like, what do you think about this investor deck? Could you help me explain this? What was your takeaway from this? When I write a summary, and I'd say it's C minus, but there's a couple of parts of it where throwing it to the AI is like man on street interview. It's like, what would some random person think of this?

• 23:30Auren Hoffman?

  Yeah, and you could even say like, hey, assume this personality and then take a look at it or something like that. And you get many different types of takes.

• 23:37Speaker 01

  But it's just catching logical things.

• 23:39Auren Hoffman?

  Yeah.

• 23:40Speaker 01

  It's like, hey, you didn't connect these thoughts or hey, this slide was confusing. Fine, like showing it to my wife and being like, what do you think of my investor deck? It's not like I'm showing it to you, but I don't want that necessarily. So a lot of my ideas or things, I'll just like, hey, tear this apart and sometimes it won't have a good idea. So I think what we really need is almost like that. I do like the copilot or some kind of thing that's not creepy, but more like pair programming of somebody who's looking over your shoulder being like, hey, by the way, maybe more like spell check where it underlines the word or something subtle where it's not just so intrusive and all or nothing.

• 24:17Auren Hoffman?

  I'm still shocked that the AI tools within Word or Excel or Google Sheets or Google Docs are still so bad. You think the spreadsheet stuff would work so well and it does kind of work, but the UI is so bad that it sometimes just doesn't save you any time at all to use it.

• 24:35Speaker 01

  No, I won't even try. We tried Slack AI at Scope three and we tried Notion AI. Then we tried, there's a tool that was AI of AIs. I feel like everyone's trying to just plug it in, but I will say enterprise search is still broken and AI is not fixing enterprise search. And you're like, oh, let's put an MCP server on top of all of our tools. Dude, I don't have time for this. Can you just figure it out for me? And this is where Google.

• 24:59Auren Hoffman?

  Even the summary, they're summarizing my email. The summary is so bad, but if I put my email, if I actually paste my email into ChatGPT or perplexity or to whatever, the summary is quite good. So it's like, you just didn't want to use the computer. How are you deciding why you're doing this summary? It's very, very odd.

• 25:18Speaker 01

  Yeah, I turned it off on my phone, all the Apple and

• 25:21Auren Hoffman?

  And the phone is so bad. The iPhone is horrible.

• 25:24Speaker 01

  It's solved. I'm with you. If you look at BlackBerry 20 years ago, the dominant device, everybody had a BlackBerry.

• 25:32Auren Hoffman?

  Yeah, I had one. I was amazing.

• 25:34Speaker 01

  They missed the paradigm. They missed the shift. And yet, it took a long time for iPhone to be as effective for typing as a BlackBerry. The new generation isn't necessarily better than the old generation. It's better enough in certain areas that people switch. I didn't switch because of the typing. I switched because of the camera. And

• 25:55Auren Hoffman?

  Yes.

• 25:56Speaker 01

  I think that's where we are.

• 25:57Auren Hoffman?

  And the GPS, yeah, and all those things, the apps.

• 26:00Speaker 01

  Exactly. I think we're in a mode where the thing that computers are good at is actually not solved by AI.

• 26:07Auren Hoffman?

  I didn't get an iPhone until 2011. I kept on my BlackBerry and then I finally switched in 2011 to the iPhone. And yes, the iPhone had a lot of things are way better, but these things are sticky too. That BlackBerry was still good enough for in many cases.

• 26:23Speaker 01

  Totally. That little scroll wheel. I mean, it was great. I loved it. And so that's my point is I'm just worried that we're in that moment where AI does some things. It's the camera, it's these pieces, but most of the stuff I do, I don't want it or need it. And yet, those few use cases are so powerful, it's going to win. It's going to be what we do. So yeah, I think the broader point is for everything we do, we should ask the question, is this one of those things where it is uniquely capable? And then maybe we don't have to use it for all the other things.

• 26:53Auren Hoffman?

  No, I would say that most software developers who are at least startups are already using these AI tools to make them a significantly better software developer. So that has definitely moved pretty quickly.

• 27:07Speaker 01

  For sure. I wrote our AI sustainability platform using AI. And I was writing production quality code so much faster. I'm not a professional developer, but I am good enough that AI help, I can ship software.

• 27:22Auren Hoffman?

  What did you use? What was the tool that you used?

• 27:25Speaker 01

  I was using Claude.

• 27:26Auren Hoffman?

  Okay.

• 27:26Speaker 01

  Claude did a great job.

• 27:27Auren Hoffman?

  So you use Claude on the side and not as part of your IDE?

• 27:31Speaker 01

  No, I couldn't figure out how to make the whole thing and have GoPilot or whatever. It was fine, but Claude was give me a function that does X, Y, Z. And it was 90% there. So even though I was copying and pasting, it was a dramatically better experience. And that's kind of the same thing we're talking about.

• 27:47Auren Hoffman?

  Cursor is essentially an IDE meets Claude, essentially.

• 27:51Speaker 01

  Yeah, exactly. And so I think today that's what I would do. But broadly speaking, like that's the stuff where AI is amazing and I'm all for it. So that's my point broadly for where we are. If you think about ad tech, the difference between being able to build something and not, it used to be building a real time ad server was basically impossible. You had to be a hardcore computer scientist who could write high frequency trading algorithms. And now you could do it. It means that all this complexity and all this spread of things, it's just going to get worse. The barrier to entry of it's hard? No. So I don't know. I think that's both terrifying and exciting as an investor and as a founder. The gap between I can imagine it and I can bring it to market is going down.

• 28:37Auren Hoffman?

  Yeah, at least bring it to market with something that's good enough. Maybe it has some scaling problems or some other things or whatever, but it's good enough.

• 28:44Speaker 01

  Yeah. So the assets that matter, like distribution, Scope 3 has a global team that is able to sell through agencies and brands. We've got amazing brand relationships. We've got holding company deals with all the majors. We've built an incredible marketing engine. I've been building my own brand. Thank you for having me, by the way. This is a big thing. And so we have the ability to take those ideas to market so fast. The products we just announced, these weren't like years of development. These were really six months from idea to ship. But the ability to drop that into our distribution channels to market, to tell the story becomes paramount. And if I have another great idea and I can get it to market in six months on top of my framework,

• 29:29Speaker 01

  that's why all these acquisitions happen. Cause I bought a company in France and November and shipped a product built around it in March. And I could take it globally. I could bring all this AI expertise. I could plug it into my existing channels. I think that's how we have to think a lot is the building part is not the challenge. It's how do you plug that into all the other assets and other unique or proprietary things we can do that nobody else can do. Nobody else can do becomes a lower bar.

• 30:02Auren Hoffman?

  I noticed you mentioned building brand. I noticed you are more active in building your brand now. You're posting these great, very deep ad tech insightful things on LinkedIn and really going deep. What made you want to invest in that?

• 30:16Speaker 01

  There's a guy named Joe Zappa, who is a marketing consultant to the industry. And he gave me a hard time in the end of last year. He was like, you have this incredible following. You've got this great brand. You have a blog that you never write on. What are you doing? He's like, you literally have the asset that everybody else in this industry is killing for, and yet you don't use it. Well, I post when I feel like it. And he basically said to me, dude, you're a professional. You'll post when I tell you to, and you will fricking go out there and do your damn job. And that is to build this brand. And he's like, I promise you, if you make a plan to go out there and post and build your brand, it will pay off big time. And I said, no, I said, absolutely no.

• 31:01Speaker 01

  My brand is not professional CEO. My brand is crazy ad tech dude who wears shorts to board meetings. He's like, that's fine. Because nobody knows that. He's like, people want to hear what you have to say.

• 31:13Auren Hoffman?

  Yeah. And also you have a very deep knowledge, in some ways, very technical knowledge. It's a niche. Certain people who appreciate that niche find it super interesting.

• 31:23Speaker 01

  Yeah. And he was right. So I went back, I went out for Christmas. I was like, you know what? Damn it, he's right. And I just started posting twice a week.

• 31:31Auren Hoffman?

  Almost on your calendar. So you're forcing yourself to write something interesting.

• 31:35Speaker 01

  Yeah. Monday and Wednesday, I've got a writing block on my calendar.

• 31:38Auren Hoffman?

  Okay.

• 31:38Speaker 01

  I try to spend 30 minutes and I'll write a post. And these days I'm maybe a little bit more active because we just shipped this product. And sometimes I'll weigh in on an issue, but I don't know. I just feel like, here's the stat. Last year, just idly messing around on LinkedIn. I had a million views for the whole year. In Q1, making this a priority, I did 1.7 million views in the quarter.

• 32:02Auren Hoffman?

  So obviously that's a big difference. Yep.

• 32:04Speaker 01

  And I touched, I think, I think 200,000 or 300,000 people read my posts. So you're talking about an industry that's, what, a million people? And so that's pretty good reach, especially if you're launching a new product. This is something that people want to hear. And then I noticed also that it helps because I go on a podcast and people know my name and they're like, oh, Brian's talking to Orin. That's a can't miss. I went to ramp up this year. I talked about it on LinkedIn. The room was literally packed and there were people outside who couldn't get in because I promoted it on social. That's to me, the essence of brand building. And so I guess I realized, can't take it for granted. Even if you're Brian O'Kelly, most people have no idea who that is and the rest of the people just don't care.

• 32:47Speaker 01

  So you kind of have to be out there talking and sharing your ideas.

• 32:51Auren Hoffman?

  It's a lot of work. It's a job. It's a huge job to go do this.

• 32:56Speaker 01

  And I also comment and I try to participate. I try to amplify people. Like I'm trying to be a good social connector. What I haven't done is tried to go on other platforms.

• 33:05Auren Hoffman?

  LinkedIn is your main one.

• 33:07Speaker 01

  Yeah, but I don't have time to go to X and go other places.

• 33:10Auren Hoffman?

  In some ways, that's maybe a better strategy. Just go all in, all in.

• 33:14Speaker 01

  Yeah, exactly. And so it's easy. It takes me, what, four or five hours a week, probably? So I've basically taken a chunk of my job and said, this is my job. But I'll tell you what's cool about it is yesterday I was on the phone outside my house in my little alley over here. And this woman walked by and said, hey, you're Scope 3, right? And I was like, yeah. She's like, oh yeah, I saw you on LinkedIn. And I was like, cool.

• 33:36Auren Hoffman?

  That's cool.

• 33:37Speaker 01

  Total brand building. I did it.

• 33:39Auren Hoffman?

  Wow. Now even the Brooklyn hipsters are following you.

• 33:43Speaker 01

  She didn't know my name, I don't think, but she was like, Scobee. So I was like, that's awesome.

• 33:46Auren Hoffman?

  Now, you did Right Media, then you did AppNexus. Most people, they're just not going to go back into the grind again after that. Doing startups is hard. It's a lot of work. It's a lot of uncertainty. You have to do a lot of annoying things when you're in the early days and stuff like that. And you are one of those few people who chose to do it yet again. And you just become a venture capitalist or private equity person, you know, whatever those people normally do, or sit on a bunch of boards, etc. But you chose to go into the super grind. Why do you think that it is for you? And is there a way to predict that in other people?

• 34:26Speaker 01

  I think there's people who are doing this because they want to make money. There's people who are doing it because they have an idea. They want to be famous. There's a lot of motivations and it's okay to have ambition in those motivations. I actually like doing it. I thought about being a VC and I've done enough angel investing. I don't really find money to be a very interesting objective function. I don't get why you need a billion dollars or $2 billion or 10. And that's much more about ego, I think, at some point than it is anything else.

• 34:58Auren Hoffman?

  A very big portion of what drives people. There's an ego component to it, for sure.

• 35:03Speaker 01

  Yeah. And I decided that my ambition was to be ad tech famous and no more.

• 35:07Auren Hoffman?

  Great stuff.

• 35:08Speaker 01

  And so I've already achieved that goal. And so then, once in a while, someone will bump into me on the street and be like, you're that AppNexus guy. I'm like, yeah. But that means I'm not bound by trying to build my own personal fame platform. I have more than enough money. So I actually get to do what I like. And what I actually like is getting a bunch of smart people together, choosing a really interesting mission, and then going absolutely crazy trying to figure out ways to bring an idea into the world to life. And it's addicting when you can say like, well, I had this dumb ass auction idea. And all of a sudden, it's on every computer and every publisher and every... Dollar in the world is going, I love that. That to me is like, if you could be an artist of any sort,

• 35:51Speaker 01

  wouldn't you want to be a startup artist where you have the canvas of hundreds of brilliant minds? And this industries of people who are excited to try new things. To me, that's the most fun, scalable way to create together, collaborate with people, learn. I learn more every day than most people because I'm just throwing myself into these new environments.

• 36:13Auren Hoffman?

  You've worked with some really talented people over the years. I've met a lot of the people that you've worked with who've gone on to do other incredible things. Do you have any type of rubric for finding talent or noticing talent that maybe other people don't notice? Or maybe developing that talent?

• 36:32Speaker 01

  I'm looking for sparkle and it's hard to define, but it's like, I love it when people have non-conventional ideas or non-consensus ideas about things or are willing to have hard conversations, to debate ideas, whether or not they're my idea or your idea, that intellectual curiosity and systems thinking.

• 36:52Auren Hoffman?

  And you think that makes a good employee and a good founder type person?

• 36:57Speaker 01

  That's part of it. Ambition plus intellectual curiosity and flexibility, plus systems thinking, plus willingness to work incredibly hard. And then a deep sense of ethic or moray or principle, something that's deeper than just, I want to get rich. To me is the starting point. But if you think about it like an athlete, like there's people out there who athletically could play any sport. It's really a question of which sport do you love? But those folks, I want the brain version of that. I want the people who are so, you just talk to them and you're like, whoa. I just learned, like I spent 10 minutes with you and I feel like I'm full. That feeling to me is great. And then how you mentor them, in my experience is you just throw them into stuff. And you are a sounding board and you're an advisor.

• 37:42Speaker 01

  It's exactly like angel investing in the sense that they're not all going to work out. They're going to need a lot of coaching, networking, opportunities. We can throw them a whole bunch of stuff. I've made a lot of bad angel investments, but I've also made some pretty great ones. I feel the same with employees. We can have an interview and I'll be like, wow.

• 37:59Auren Hoffman?

  Yeah, every once in a while, this person becomes a hundred X person.

• 38:02Speaker 01

  Totally. So I just keep thinking fertilizer. How do I fertilize this? What can I do to help this person grow and develop and stay out of their way?

• 38:11Auren Hoffman?

  And is it just giving them more and more challenging things to do essentially? And is that the simple way of doing it?

• 38:18Speaker 01

  I think they find their own, you know, it's like, I don't find them the thing to do. Errol Lido at Nexus, and one day he said, hey, we're going to open source all of our header bidding code. And I'm like, no, we're not. That's our core IP. We invented that. That's one of the few things I've actually done in this world that I invented. He's like, yeah, we're going to give it away. I've already created the prebid.org website and just letting you know, I'm like, timeout. Like you're so far ahead of me here, dude. Wait, what? And obviously we did it and it was brilliant. And now he's the CTO of Cedar. That's a multi-billion dollar company. And he's gone on to much greater things than I ever will. But I just stayed out of his way. I can't think of much I ever said to him that was helpful to him. I just kept being like, okay, you're so smart. Here's five things you might want to throw into your brain.

• 39:02Auren Hoffman?

  And also maybe don't layer that person as you grow or something. Let them grow as the company grows.

• 39:08Speaker 01

  Yeah, but it's weeds, it's swirl, it's politics, it's drama, it's bias. It's all the stuff that keeps people from loving their job. It's how do I give you the tools and give you the resources and get rid of the bullshit?

• 39:21Auren Hoffman?

  The bureaucracy.

• 39:23Speaker 01

  Totally. That's the last thing we need. And so I think that's the hardest thing for folks that work with me is, and I think a lot of folks are learning this lesson, which is none of that bullshit adds value. I'm not saying that process doesn't add value. I'm just saying fundamentally, knowing what matters and figuring out how to get as many resources going in that direction as possible is really the secret to a lot of management. It's just, there's the goal. Go get a goal. I'll be over here. Or more likely, I'll be running there as fast as I freaking can, but giddy up. I think that's the best way of just doing it. Don't overthink it. Just go get the ball, kick the ball, make a goal, celebrate for a minute and let's do it again.

• 40:02Auren Hoffman?

  You're in the climate world and there's definitely been like a vibe shift around climate over the last year or so. How do you navigate that? And how do you suggest other people navigate it?

• 40:13Speaker 01

  I think that the virtue signaling part of climate is not happening right now. So there are a lot of people bragging about how green they were or a lot of this conversation. It's funny because the environmentalists hated it. They were like, this is all greenwashing bullshit.

• 40:30Auren Hoffman?

  There's so much greenwashing.

• 40:32Speaker 01

  Yeah. And so now the environmentalists, as always, have gotten their way. And now everyone's like, fine, we're not going to talk about green. And now the environmentalists are like, but wait, go back to the green thing.

• 40:43Auren Hoffman?

  Because now they might not get funded, right?

• 40:45Speaker 01

  And so I think that's the problem is we had a great speaker at our conference a few weeks ago, Ayana Johnson, and she said, look, if you're operating on hope, the problem is you have hopeful days and not so hopeful days. She's like, that's just not going to get it done. She's like, execution is sexy. Just show up, do your job, fight for the things that matter, and don't worry about whether you're feeling optimistic or not. Just do the work. That's kind of how I feel. We've consistently been able to paint a picture for the industry that there's a huge amount of waste. That waste is financial and that waste is environmental. And if we fix the waste problem, we'll fix both. And so I don't care if you come for the cost savings or the carbon savings.

• 41:29Speaker 01

  Just do the damn thing. I'm actually shocked. Our Q1 was way better than I thought. Coming into Q1, I thought we were going to miss our numbers. I was like, this is going to be the worst quarter ever because of Trump, because of all this pushback. CSRD got pushed back in Europe. Everyone's anti-climate. Not the case. We closed massive deals. We had a ton of lean in. All the work we've done for three years to seed the market has made it very exciting and possible.

• 41:56Auren Hoffman?

  You could position as save money. That's a slightly different positioning or something.

• 42:01Speaker 01

  Yeah. So the reason this new set of solutions is exciting for us is we can bundle more things with the sustainability data. And what clients are saying is we've always wanted to do it. It's just been hard to justify. But if you're just giving us something that we get huge benefits and it's also super green, we're in. And we're so proud and glad to work with you.

• 42:22Auren Hoffman?

  Yeah, people want to save money. So that's simple. And if it like saves money and does some good, well, that's great. I'm happy to do that.

• 42:28Speaker 01

  Yeah. And people are like, oh, you're pivoting. I'm like, no, I'm bundling. And the nice thing about a bundle is I can do all the things and my competitors can't. And so we're going to win. And we're also going to do a ton of good.

• 42:40Auren Hoffman?

  Yeah, a couple of personal questions. Is there a problem outside of ad tech that you would love to tackle someday?

• 42:46Speaker 01

  I was really excited about real world supply chains. And then I found out what real world supply chains are like. And I was like, whoa, that's a lot harder than I thought.

• 42:55Auren Hoffman?

  Right. We had talked about this a few years ago, I remember. Yeah.

• 42:58Speaker 01

  Yeah. I think the problem is that most of these problems are so complicated in reality that it's hard. But look, if I could be the guy who got to play with the user experience for AI, wouldn't it be super fun to be like, hmm, I want to go figure out how we're going to make Excel or Word or my phone, or there's something really cool. The only problem is I have no skill set for that. I would not do a good job. I would like to be the kind of person that could imagine the next paradigm. I think I'm just past my prime. I think I'm too deep in my own little world and have so much pre-existing knowledge that I'm not sure I'm going to see the next paradigm. I think the best I can do, and maybe you too,

• 43:39Speaker 01

  but I know for me, my job is to get as much of my brain imparted onto the next generation. So they make different mistakes than I made and that we do as much of that fertilizing. So when they do invent it, maybe it's someone who was like, I was on that podcast with you and Orin and like it changed my life. That'd be cool. I get some credit for that. But at this point, I think it's footprints in the sands of time more than it's I'm going to go do the next thing.

• 44:05Auren Hoffman?

  Let me push back on that a little bit. Historically, as people age, they are less quick on things. Their calculation speed goes down, their memory recall goes down. So these are all the downsides, but they have more money, they have connections, they've got more ability to get things done. And maybe somehow evens out, maybe one side wins, one side other wins. But with all these technology things coming on, like they're all helping everybody calculate better, they're helping everyone recall and do memory better, helping people process stuff better. So maybe somebody like where you and I would almost be over the hill. Once you hit 50 or something like, oh, my God, you guys are so old. Now, maybe

• 44:48Auren Hoffman?

  50 is the new 35. Maybe you're still not inventing anything like massive at 35 or something, because maybe you only do that when you're 25. But you may have the ability, as long as you still have the energy to go do it.

• 45:01Speaker 01

  No, no, we're gonna have long and productive lives. I just think it's about how do we take those connections and knowledge and systems and make the next generation of innovators better. You know how Sergey and Larry had to bring in Eric Schmidt to be the CEO when Google was scaling up because they were just kids? I would love to be that for somebody. I would love someone to call me up and be like, hey, the next big thing, can you come be the mentor?

• 45:28Auren Hoffman?

  Adult in the room.

• 45:29Speaker 01

  Yeah. That's where I am.

• 45:31Auren Hoffman?

  But Eric wore like a suit. You and I, we dress up still like kids.

• 45:36Speaker 01

  I don't know. I think the new suit is wearing a T-shirt and shorts.

• 45:39Auren Hoffman?

  Shorts.

• 45:40Speaker 01

  But that's just me.

• 45:41Auren Hoffman?

  Okay, that might be right.

• 45:42Speaker 01

  But that's the point is like, I'm never gonna be the suit. But I do think, and maybe I'm just nuts, but there's a point where I can get great satisfaction from helping other people achieve what they're trying to achieve. I do that in a little way as an advisor, as an investor, as a CEO, as a dad. I mean, I'm just increasingly like, look, I don't need any more of the thing. I have everything I've ever wanted or dreamed. It's not about me now. It's about how can I just take this incredible energy I have and this creativity, this demonic need to go make things happen and do as much good and maybe influence and help as many people as possible with the few decades we have left to mess around with AI or whatever we're gonna do.

• 46:23Auren Hoffman?

  I love that. All right, two more question we ask all of our guests. What is the conspiracy theory that you believe?

• 46:28Speaker 01

  Oh, we're definitely in a simulation.

• 46:30Auren Hoffman?

  Okay, so you believe the simulation one?

• 46:33Speaker 01

  100%. There's no other explanation. There's no other way to understand this universe, but it's just a really good simulation.

• 46:39Auren Hoffman?

  Got it. So essentially, that means you believe in a deity, you're a creationist.

• 46:44Speaker 01

  I'm basically a Marvel creationist. I just wanna know how to get to the next metaverse.

• 46:49Auren Hoffman?

  Okay, got it. I like it. All right, last question we ask all of our guests. What conventional wisdom or advice do you think is generally bad advice?

• 46:56Speaker 01

  Oh, follow your dreams. Never do that.

• 46:59Auren Hoffman?

  This is the most common thing people say on World of DAS. I've asked this question now to over 100 guests. The most common one is to follow the dreams is always the bad advice one.

• 47:09Speaker 01

  Oh, my gosh.

• 47:09Auren Hoffman?

  Why do you think it's so bad advice?

• 47:11Speaker 01

  I believe in strategy. Don't follow your dreams. Go fucking manifest your dreams. Go figure out how to make your dreams real or figure out if they're even plausible. Write a business plan. Pitch me your dreams. Let's figure out what resources you need to get there. I just think it's so abstract. There's no destiny. This simulation has rules and those rules are very clear, which is you've got to work your ass off. You've got to figure out what you're going to do. Get lucky here and there. Figure out the simulation. That's my advice.

• 47:38Auren Hoffman?

  Yeah, I love this. Okay. Thank you, Brian O'Kelly, for joining us on World of DAS. I follow you at B O Kelly on X, but also Brian O'Kelly on LinkedIn, where you're very huge content. Well, I love the content there. I definitely encourage our listeners to engage with you. This has been a ton of fun.

• 47:54Speaker 01

  Awesome. Thanks, Orin. Love being on your cast, pod, whatever you call it.

• 48:02Auren Hoffman?

  Thanks for listening. If you haven't already done so, please subscribe to World of DAS wherever you consume your podcasts, YouTube, Apple, Spotify and more. And please help us get discovered by leaving a review and check out worldofdas.com. That's worldofdas, D A A S dot com. And of course, connect with me on Twitter, at Auren. That's A U R E N. Would absolutely love to hear from you.