← AdTechGod Pod
Loading summary
Transcript32 lines
- [00:01]
AdTech God
This is adtech God and this is a commercial message. Finding the right audience shouldn't feel like doom scrolling with Experian. It doesn't. Experian syndicated audiences help you reach holiday shoppers, car buyers and more across over 200 top platforms with over 3200 pre built audiences. There's no more doom scrolling. It's just audience targeting you can trust. Made simple. Learn more@experian.com architecture Again, that's ex P E R I A N.com marketecture.
- [00:48]
Commercial Narrator
When did making plans get this complicated? It's time to streamline with WhatsApp, the secure messaging app that brings the whole group together. Use polls to settle dinner plans, send event invites and pin messages so no one forgets mom 60th and never miss a meme or milestone. All protected with end to end encryption. It's time for WhatsApp message privately with everyone. Learn more at WhatsApp.com welcome to the.
- [01:15]
AdTech God
AdTech Godpod, your window into the world of advertising technology and the people behind it. I'm your host, AdTech God. Welcome to the AdTech God pod where we speak to those driving partnerships in AdTech. Today's guest is David Danziger, the SVP of partnerships at Distillery. David's work at companies like Liveramp, Habu, Merkle, Axiom and more. I've actually engaged with David for years online. We've DMed each other, interacted on X, interacted on LinkedIn, but we've never met. So I'm really looking forward to getting to know him today. David, welcome to the Pod.
- [01:53]
David Danziger
Thanks, Ad Tech God. Great to be here. Excited to meet you. This way. This is fantastic.
- [01:57]
AdTech God
I know it's amazing, but I feel like we've, we've literally messaged each other for three or four years, so it's crazy. We've never actually crossed paths in person, but maybe one day.
- [02:07]
David Danziger
Exactly. I'd love that.
- [02:08]
AdTech God
Or maybe we have David and you just don't know how to just floated by on a cloud. Could be. David, your background's incredible. I think we first started interacting with each other years ago when you were at the trade desk, but since then, now you're leading up partnerships at Distillery. I like to start the podcast with kind of getting to know you and your background and what made you join Distillery and what got you to this point.
- [02:28]
David Danziger
Yeah, fantastic. So I don't know whether to work front to back or back to front on that, but most recently joining Distillery was a. Was an exciting opportunity because I got to know Distillery when going back a little bit when I was running data partnerships for the Trade Desk and Distillery and had pivoted away from being a dsp. And so I got to know some of the Distillery crew as they got into audience targeting and they were very advanced even at that stage. This is going back 2017, 18, very advanced in the use of ML, machine learning, some early stage AI at that point, and we're doing a great job with it. I saw their audiences popping and so got to know Distillery then. So it's very exciting to be part of it now, but going, you know, kind of going way back if we sort of take a journey standpoint. I almost fell into this industry as I feel like a lot of people do. And by this industry I mean intersection of data and marketing. I was coming out of grad school and interviewed with this startup in Austin, Texas that doesn't really exist anymore. They've gone through some acquisition, but they, they were doing neural network optimization for direct mail at the time. And it was through there that I first got into the world of data and direct marketing. From there went to Axiom where, you know, I spent some time on the product side refining certain types of audience products on the B2B side, then B2C. But the, it was really the, the latter period during the time at Axiom where we were migrating some of the data offerings from stuff to be used in offline and email and into online. And that's where I came into, you know, getting Axiom audiences on platforms like Blue. Kai got to understand a lot more about ad tech where I first got to meet Jeff Green who at the time was at Microsoft. And you know, this was pre Trade Desk days and he was just, just starting to build that up. So it was kind of through those later Axiom periods where I got into the, the true ad tech space, the intersection of data and ad tech. And that continued as I went into Merkle and then onto the Trade desk from there. So it's kind of an interesting pathway that I think is probably consistent for a lot of us. Where we start in one area, it kind of moves along a different trajectory and then on to Trade Desk and data partnerships from there. Habu with data Clean rooms. Because I really felt like there was an opportunity for companies to use one another's data in some interesting ways and new in different ways. Habu got bought by Liveramp and so that, that kind of germinated from there. And so here I am today.
- [05:02]
AdTech God
Incredible. Over the span of your career, how have things changed? In particular, you said neural network optimization first of all, what the heck is that? Yeah, let's see, let's start there. What, what does that even mean?
- [05:15]
David Danziger
Yeah, it's funny, neural networks are, they're still around and still kind of the underpinnings of some of the, some of the models that are used today, both for, you know, scoring data sets and evaluating the quality of data and even underpinnings for certain types of language models. And literally in the interview for that one, you know, it was, it was for an internship originally and the guy was like, so what do you know about direct marketing? And I was like, well, I know a little this, little that. And I was talking about things I learned in class. They're like, what do you know about neural networks and back propagation? I was like, I can't even spell that. I don't know what that is. But it's funny with respect to your question of how it's changed, like the hardware and compute power of what we were running on during that era, this was, let's be honest about my age, this was the late 90s. And so we were running on hardware where if we were scoring a million record data set that would run overnight and we'd come back and see what the scores were. And so when I think of the change, you know, essentially you're scoring things in binary fashion to know who to target, who not to target. It would run all night versus today where you know, something like that would run in an instant and you know, you'd be able to get, get things back immediately where you can make determinations very much more for a real time environment. You can do things that you never would have thought of before, but all with the same general underpinning of who am I trying to reach and with what message. We, we can just do it in a much more complex and comprehensive way now and way more, way faster. But the same principles have always applied.
- [06:52]
AdTech God
It's the technological advancements over especially like the last four or five years have been incredible with the power of, you know, machine learning, the improvement and speed of getting back, you know, the results that you're looking for. Like you said, in an instance you type something in and within a few milliseconds the answer and the result is there for you.
- [07:12]
David Danziger
The other thing, I guess I would just also mention, you know, I mentioned the time difference in terms of how quickly, but also that the number of dimensions or vectors that you can feed into models like that for outputs that are meaningful as well. You know, before going back in time, we were certainly heavily constrained by okay, how much data can we put in? What types of data are even available to put in? Wildly different from what it is today in terms of the advancements and sophistication on that which has been fantastic to be a part of.
- [07:42]
AdTech God
And see, I guess back to my question, like with the technological advancements, the power of AI, the decision machine learning that you've seen, like this is obviously creating a situation, an opportunity for companies such as yourself and others in market to really kind of churn out these custom audiences and these targeted segments that you're looking for a lot faster. Are you finding that the benefits of all this technological advancement is actually creating highly targeted audiences that are maybe more niche? Are you finding that it's still kind of a wider audience is performing preferred from advertisers? Are these custom segments more diverse than they used to be? You know, think prior would be something like a automotive enthusiast would mean someone who was interested in a car. Today with the amount of computing power, I would think you could find an automotive enthusiast that also shops at Vons that is also interested in Warby Parker glasses. And so we want to highlight the AR feature in our vehicle which then gives them, you know, improved whatever. Are you finding that these audience segments are becoming more targeted and refined now or do you still feel like the standardized 100 audience segments are still preferred?
- [08:55]
Financial Services Narrator
So you're about to make a trade based on a friend's text, but which you do you listen to is it, we could buy a house in Tulum, get optioning those options, we could lose everything. Or let's do a little research, get your head in the trade and make the investment decision that's right for you. Learn more@finra.org TradeSmart It's a great question.
- [09:25]
David Danziger
The answer is they are much more refined, but there are still pockets where the general ones are preferred. I'll use as an example. I mean you talk to people in the industry constantly and so you know it with, with as linear moves, linear TV moves into connected television, there's still a healthy dose of connected television advertising that is bought and sold on, you know, basic age and gender demographics. And that's partly a reflection of supply and demand available things like that. And that's, that's going to change. But to, you know, really more specifically to your question, broadly speaking, the, the models have gotten much more specific in the types of custom data that can be deployed. Whether audience driven or, or not are, are much more sophisticated. And your example of automotive enthusiasts is perfect example where you know, automotive enthusiasts used to be in and of itself a useful segment. Now you can understand much more, much more closely who's in market for an automobile at a any given moment. What types are they most involved with by, at a brand level, at a model level, what you can apply geography, you can apply when they're most likely to be in, in proximity to certain things. So all these different dimensions and then the frequency with which they can be updated so that you recognize, you know, at a certain point, ad tech God has, has already purchased his Mercedes. I'm, I'm making stuff up. I don't know if you're a Mercedes buyer or not, but he's already purchased his Mercedes and therefore he should be out of the cycle. And so, you know, if I look, for example, at what distillery is doing, we're were updating approximately 21,000 models and related segments on a daily basis. And that would have been unthinkable, a, to have 21,000 segments because of the specificity, and B, to do it on that frequency. And those are the kind of changes that really drive a difference in terms of what, you know, the effectiveness of a certain brand's marketing and how they're doing it.
- [11:20]
AdTech God
How do you think things have changed in way, you know, selecting these audience segments for marketers and for buyers today prior, you would, you know, select a particular type of audience segment and you would basically click a box, say, this is who I want to target. You'd still kind of have to eyeball, from what I recall is you'd have to eyeball what audience you want and you'd have to select that segment and move forward. It always felt like a little bit archaic in a way that with all this power, all this commuting power, that you're still. The final step, and probably the most important step is prone to human error. And so picking the wrong one or picking an inaccurate segment could really negatively impact. How do you see things have changed today with, you know, AI and better models in place?
- [12:05]
David Danziger
It's a good question. And the answer is it is changing. And there's. But there's tons of room for improvement as I even, you know, you described it perfectly sort of in the challenge, in the question. You know, when I, when I look back, I started at the trade desk in 2013, and at the time, the way you would find the segment you were looking for, if you're, you know, if you're a media buyer and you're going to run a campaign for coffee drinkers, you would type in, you know, coffee drinkers into a search bar. In the, in the Data management platform section and, you know, 37 segments related to coffee drinkers from 28 companies would show up and you try to make a decision as an individual based on how much did it cost. And you know, did this salesperson come in and do a presentation in our company last week? So I think it's probably a good segment, but there was, you know, there's no real intelligence or guiding methodology to help you select. And a lot of platforms have done better. Like, I know in Trade Desk, we introduced relevance scoring a few years ago, that was very strong and other companies have taken steps as well. But to your point, this is still an area where we as an industry could do a lot better and we're super optimistic. Actually, actually within Distillery, I know we are about the idea of companies introducing more and more automated intelligent decisioning capabilities into their platforms so that they can say, look for your KPI and what you're trying to do. This is the right audience that you should be selecting for coffee drinkers or whatever based on the, the under underlying science that we've got running it. So the, you know, if I put all that together, there's ways that platforms I think can do this a lot better and many are starting to try to with, you know, driving more algorithmically selected data and enabling that as an option for different customers. We've seen, you know, we've seen Amazon start to steer that direction. Yahoo, Google DB360, the Trade Desk buyant, lots of different ones. And I'm leaving out plenty of platforms too, so. But there's room to make it much, much easier for automated intelligent decisioning. And from a, from a Distillery standpoint, we generally feel like, hey, the more, the more platforms do that so that, you know, we're competing on data science versus what, what's the lowest possible price or which salesperson was in the door last. The more we get to do that, the better it's going to be for the industry. Not only, I mean, not just for Distillery, not just for the platforms, but for the advertisers, for the agencies. We think that's a win all the way across the board.
- [14:43]
AdTech God
What do you think in from where you sit in the ecosystem at Distillery, running partnerships, obviously the growth of AI and machine learning and the importance that's playing a role in our industry overall, because it's obviously disrupting things in a way. In some cases very good. In some cases it's creating challenges for some of the ad tech companies. What do you feel is a positive change that's happening in our industry that you've witnessed over the last 12 to 18 months.
- [15:12]
David Danziger
One of the biggest ones and it, it feels almost trite to say it just because it gets so much attention now is agentic AI and sort of the, the realization not only of the agentic features of it but also the realization of what we've been able to do with large language models too and the communicating back and forth among companies where somebody who's trying to figure out the right means of targeting can, can go at it just with a, A, a chat style, chat GPT style interaction either with a data company. Most agencies are introducing these capabilities as well where instead of going back and forth between data company and media buyer of here's what I'm trying to do. Oh, here's what I recommend. Okay, what if we refine it this way? Oh well, I think you should do this. Here's the model that we do. How do I get to that data? So many of those things literally can just be done now through chat functionality where we've got, you know, not to make it and that distillery has its own tool for things like that. So the long and the short of it is it's AI that doesn't, you know, it doesn't replace the human aspects of it but it certainly makes it way more efficient so that someone can get on to doing the higher value activities that would have been taken up with emails back and forth to arrive at. Here's what I'm trying to do, here's how I get it done, here's how it's going to be sourced, here's how it gets to point A, from point A to point B. Here's where it activates. All of that is much more effective now even than it was 12 to 18 months ago. And we're going to, I'm confident there, we're going to continue to see improvement on it. It might not always keep up with the, I'll call it the bluster from, you know, from our stages at different events but it's, but that being said, you know, for all the chatter about it I, it does really feel like it's already making a difference on different fronts.
- [17:06]
AdTech God
It does feel like it's materializing. I think the overuse of the term AI, I think is a negative to the industry because people's expectations are a little too high too soon. That being said, like I've had interviews with people who are using it, like you said, like front end campaign decisioning, audience selection. I've had interviews with people who are Running this to optimize operations, to remove the menial tasks, the tasks that you know is a process that you've already done day over day for the last five years that a machine can do on your behalf. So you can focus on other things like client care and customer service and improved inventory sources. Like all of it across the board seems to be implemented. But when we go on stage and we talk big picture, I think that's nice to hear. And some people are very good at it. They go on stage and they're like, this is the new world. That's great. The problem is we've heard that before. I don't really want a new world. I want to know how it impacts me and my role and my buying capabilities or my selling capabilities. Day to day. It's great. I want to hear about the new world because I'm sure it's coming at some point. But today, how does it impact me? And it's sometimes it's hard to see what positive impact it plays on our day to day.
- [18:22]
David Danziger
Today I think you hit a big theme on where it really can make a difference. When you said, you know, so that I can focus on. I think you use the term client care and client, you know, client service because one of the other feels like opportunities that gets lost in, in these discussions is the time freed up for that client service, client care. Because a lot of these solutions, they don't, they don't replace the need for human interaction at all. They, if, if done properly, they complement it. But they're still the part where somebody has to understand precisely. And I, this is where I think AI still misses a lot. Clearly is, is in the nuance specifically of what a client is trying to accomplish or the nuance of how something's being measured. Ultimately that at this stage is still only conveyed in human to human interaction. And, and if so, if, if, if, and so if AI automates and simplifies to your point, the more mundane parts of it that are, you know, that have to get done but can be done much more easily or in a fraction of the time so that that much more time and resource can be devoted towards understanding and deeply understanding what it is the customer is trying to accomplish and understanding the best ways to do it, then that that should lead to good outcomes. It can't simply be a cost cutting mechanism, but rather a method to improve what it is that we're trying to do from a client to client partner interaction.
- [19:55]
AdTech God
Yeah, I mean I think, I think even platform usage. Right. If you think about the Future of where you type in a particular prompt with the category or in the type of inventory you're seeking or you're KPIs or what you're looking for for your campaign. None of that is possible without having a good understanding of what your client is looking for. And so even though I could see a world where some of the menial tasks are being replaced by AI, I think the human element will always remain the customer service aspect. I would hate to chat with a bot every day, especially if I'm having issues like that, to me would irritate me. I already get irritated with Zendesk tickets. You tell me now that I'm chatting with a machine, when I'm upset and frustrated that something's not working, I would go nuts. So I do feel like the client success, customer care aspect, the relationship aspect, the partnership aspects will always be there. And then on the operational side, hopefully the goal is how do we operationalize our product or our solution better and faster and more efficiently, rather than how do we completely remove the human element and make it all technologically powered. And I just don't see that happening in the near term. But I do see, unfortunately, companies talking about that being the future. And I just simply don't. I don't believe it. Or at least I don't want to.
- [21:21]
David Danziger
I think I don't want to. I don't believe it. And I don't even. I've got to think over time it will, what you and I are talking about will prove to be, you know, not that I, not that I assume we're such spectacular prognosticators, but what, but what I assume will be found is that it changes the nature of the types of people that you need in different roles because a lot of the certain things can be automated. It puts a different, different emphasis in terms of the skill sets of who you want in client service roles, what they're expected to do, the nature of the day to day that goes with it too, that, that will probably change, but it does, it certainly doesn't obviate the need for that kind of client service for exactly the reason that you said tech God, in the sense that, you know, in certain circumstances the last thing someone wants to do is interact with the machine on that basis. And I, I don't think it's just that you and I are Luddites in saying that either, because I don't think one of us are.
- [22:19]
AdTech God
No, no, there, there's nothing more irritating for me than something that happens and I email and they say we will get back to you in 48 hours. Your ticket will get pick. Even that alone irritates me. It's like, well, if I'm spending money for a solution and I'm spending money with you every month and I'm paying whether it's subscription fee or usage fee, and it's not $5 a month, like it's substantial spend. Like I want someone to pick up the phone and call me and say, hey, just don't worry, relax, we're all over it. I'm going through the process internally. Whether then that stage requires automation and AI and whatever to find the solution, that's fine. But I think from client to, you know, vendor, I would much rather have the human element in the middle. And that's just. Maybe I'm just old, not sure.
- [23:04]
David Danziger
No, I don't think so. I think actually it's been interesting to watch what companies succeed and how they succeed in our industry too. And it feels like certainly some companies have done very well that end up with a reputation for it. Yeah, but I can't get my account team to call me back or whatever. But I feel like those tend to be the exceptions rather than the rules in terms of how individuals end up wanting to do business. And so they choose companies where, and certainly they choose teams where they feel comfortable that, you know, when it comes down to it, their, their team of humans with whom they work at the other company are going to be ones that are, you know, feel like they're in the trenches with them and eager to help them succeed too. And so the ones that have totally, you know, kind of leaned into an account service model have tended to do.
- [23:51]
AdTech God
Pretty well, especially like our industry. Like you look on Even something like LinkedIn, which is professional network, but then you see the, the selfies and the photos and the group photos of people. Like, I know some people are like, oh, cringe. Like this, this is annoying. But at the same time, like, it's kind of nice. It's kind of nice to see that you've got two different supply side platform people hanging out with two different publishers and then a data solution and market and they're all taking a photo together and half of them compete with. Yet there's a relationship that's built and an understanding that's built that, hey, we all work together in some capacity, some are better than others. But it's the human element and that partnership piece that plays such an important role with it. And I don't know if that'll ever go away.
- [24:32]
David Danziger
I hope not. And I agree with you. I think it's probably something that when you and I were first messaging, it was on similar topics too, in the sense that there is an inherent camaraderie within the business, partly because it is a, a business of interrelationships as far as how the different companies work together. And so it. To your point, it's not unusual to forge these personal relationships even among companies that compete with one another, partly because, you know, you never know when an individual within those companies moves from point A to point B. And you want to be able to work more closely with them or develop and have a relationship with them because you like them and you trust them, but also just because it's a, it's. It makes it a more fun way to do business if it's, you know, if you get the chance to interact with people that you respect and that you trust and where your relationship goes beyond just the transactional of, oh, we have to get this contract done or we've got to get this, you know, this campaign live. It certainly helps if there's, you know, already a personal relationship there that runs a little bit deeper.
- [25:38]
AdTech God
David, when you, when you look at the span of your career working in ad tech, working in the advertising industry, just to close things out, like, why are you still here? Why haven't you shifted your, Your skill set can obviously be transferred to other industries pretty easily, I think. But why, why ad tech and why advertising in general?
- [25:57]
David Danziger
Yeah, some of it is what we just covered in the sense of, like, they're, you know, I've got relationships that I feel like they're, they're people that I enjoy spending time with. They're people that I learn from, from. And so that's always good on the personal front. But the other part of it is, is that learning from part the, the world in our industry that is made up of advertising technology and data, you know, people like, say, oh, it's changing so fast. And it's true that there's always changes, but they're in, in the sense that there's always things to learn about what we're doing. And, you know, while you and I joked about it beforehand, we may not be curing diseases, advertising and making people aware of what, you know, what products are available. And to the degree that it supports content out there, all of that is still important. It's not going anywhere. And so, you know, between being able to keep learning, between it being with people that I enjoy and doing something that I legitimately think matters for, you know, funding content for overall, you know, where things go and how people learn about things. That all matters. So you put all that together, it's, it's an interesting spot to be in. I don't, there's, there's never a day where I wake up and think, well, I've learned all there is to learn here. There's nothing more to do. See, that's, that's, I just don't see that happening anytime soon or ever. So put all that together. Advertising, technology, data, it's a good, good intersection to be in.
- [27:31]
AdTech God
David, I wanted to thank you for being here today. Thank you to Distillery for having you here. Love the chat. I really appreciate your time and glad we finally got to do this.
- [27:39]
David Danziger
Indeed. Thank you, ad tech God. Great to be here. Thanks a lot.
- [27:42]
AdTech God
Thank you.