Ignacio Ortiz, X3M: “AdTech used to be free — and that was a mistake.”

TRAILER

Ignacio
Honestly, I think now is probably the best time to be in the business.

Neha
I generally feel like getting stupider day by day.

Ignacio
At some point, we go back to the fact that ad tech is free. We think this is a strategic mistake.

Neha
It's heartwarming to know that everybody's facing the same problem.

Ignacio
We're so far away from needing very precise predictions at a very precise user level.

Neha
I love that you went old-fashioned, you know, brain and spreadsheets.

EPISODE

Neha
Hi, everyone, and welcome to Points of Growth, a podcast from Yango Ads, where we explore the fast-moving world of ad tech. I'm your host, Neha Dawar. In each episode, we bring together voices from across the industry to talk about advertising, app monetization, and the trends shaping how brands grow in today's attention economy.

Our guest today is Ignacio Ortiz, an entrepreneur who began his career in capital markets and finance before diving into the wonderful industry of ad tech. He's the co-founder of Xtreme, a fast-growing mobile mediation platform, and Lumit.ai, a division within Xtreme that allows publishers to tap into user-aware, dynamic AdMon setups. Ignacio, welcome to Points of Growth.

Ignacio
Thank you so much for having me.

Neha
So we come to the fun part. So always an introduction about yourself. Can you tell us what sparked your journey into ad tech and the whole digital experience?

Ignacio
It's interesting. You know, you mentioned my background and I started in finance, which led me to marketing, interestingly, which was completely unexpected. And that's how I started learning about, let's say, the ad tech ecosystem, the gaming ecosystem at the time. I was with Etermax, a large Latin American studio. And on the side, I've always been investing as an angel investor. I've always been paying attention to any business opportunity that I could find, kind of understanding it, trying to figure out what the pain points were, what potential solutions were. And, you know, the more I worked within anything related to monetization at the time, it was more user acquisition, but alongside my current co-founder and then chief revenue officer, I kept looking at ad tech in general and feeling that it was increasingly a pain point for studios and for the mobile ecosystem. And eventually that led us to start the mediation.

Neha
It's a very colorful journey. Lawyer, finance, and now sort of in the whole ad tech industry. Coming back to what you said, you know, pain points. That was actually going to be my next question. What pain points in the ad ecosystem were you trying to solve when you created Xtreme?

Ignacio
Let's talk about gaming for a minute, even though we really do address mobile app developers who monetize with ads in general. But within the gaming ecosystem, it's extremely interesting to see how despite owning the supply, so the user base, they really don't have much power in how ads are sold today. And we always found that to be both unexpected and interesting, and of course, a pain point to the studios. There's many ways to explain it, but at some point we go back to the fact that ad tech is free. And we think this is a strategic mistake on behalf of studios back in the day, because as a result of being free, of course, innovation has gone elsewhere.

There has been very little, I would argue, supply-side innovation over the past few years. You can see that in how the industry has evolved. I mean, when I sit down with studios, something I really enjoy doing is start talking about anything but monetization. So I ask them, you know, how do you manage notifications? How do you think about pricing and discounts on IAPs? How do you think about user progression and so on? And everything they discuss is extremely nuanced and dynamic and resulting from the user behavior. And then I say, okay, fine. Now let's talk about monetization. What exactly are you doing? And it's funny because the answer is not much.

They tend to give pretty much every user the same experience with the same placements, formats, mediations, and so on. And we just don't think that makes sense. But in their defense, I think it just is a reflection of the lack of innovation that we've seen in the industry. So, that is kind of the starting point for everything we've done. And then off of that, I think you see a couple of other points worth mentioning, one of which increasingly common to hear complaints about is that quality. And you can decide whether you want to have long ads or short ads or high engagement ads, so to speak, how many clicks you need to leave. But you need to be able to choose right. And studios really struggle to control that experience. So that's a pain point. And then maybe the final thing I'll say, and I'll leave it at that because we can go on forever. But yeah, I mean, look, optimization windows. You may have between two rewarded video placements. You might have a few minutes. You can run an auction in a few seconds and then there's a ton of time that you could probably use to try and get a better price. We argue that you should be able to like and I think studios would like to be able to and you know if you're unable to that's fine but if the market is willing to buy that ad at a higher price it would be great to be able to sell it.

So I typically point to these three issues when I think about pain points but I'm sure there's more. You can go to a conference and ask around. You'll find everyone has a different one to bring up.

Neha
I think of the most common ones it and we've heard this not just from the gaming industry and the ad tech industry and it's just about this concept of personalization and that's been a topic that's come up again and again time and time again, I mean it's heartwarming to know that everybody's facing the same problem but also like what are we kind of doing about it.

So, do you have any thoughts on the state of mediation in 2025? Like, is something working? Is something not working? What's broken? What needs to be fixed? What needs to be eliminated completely? Fire away. It's your time to shine!

Ignacio
Yes, I'm putting the boxing gloves on. No, look, there are things that work. Let's be honest. If you don't have a large operations team, if you don't have a lot of know-how, honestly, I think now is probably the best time to be in the business. I mean, it's fairly easy to set up and then you get, it's fairly easy to get the networks in and then you can make it work and bidding makes it all very simple. And to some extent, you can get most of the way without being extremely sophisticated. Waterfalls were very complex. We know that. And this has really made it easier.

So I'll give mediations that because they, I would argue, imposed bidding on the ecosystem. And it is true. It is easier to manage and simpler to set up. I would argue everything else needs to be fixed. Like I mentioned earlier, optimization windows. I think, look, if bidding was outstanding as it was positioned to be when it was launched, it wouldn't need to be forced upon the industry, right? Waterfall would have died away because there would have not been a better price within the waterfall.

And yet you look back maybe two years ago, everyone was talking about trying to run some networks on Postbit because they were able to sell their inventory there. And I mean, you really should. And I say this full heartedly. You really should comply with the policies of whatever mediation you choose to work with. So, you know, don't go get into trouble because you will get into trouble. But if you can, like the fact that some studios were able to sell their inventory into Postbit is telling you that there was demand, right?

There was a higher price that you could have gotten had you been able to optimize with a second optimization window. So, I do think that it would be great to get the option to do that. Same with that quality, which I mentioned before. And then there's not enough, I think, or not enough public, at least, conversation around this. But there's a lot of nervousness within many non-mediated networks. I mean, those that don't own mediation partners, and they feel that they compete at a disadvantage.

Most of the inventory that they can acquire comes from mediation platforms which are also owned by their competition so they struggle when they understand that they are to some extent platform reliant on companies which compete with them and that is a risk a fundamental risk as far as they see it so that is definitely something which is not from my perspective because I don't own a network and i work for studios but i when i look at the ecosystem in general that's a conversation that i often see come up people are concerned with the risk that they're running and these have become very large companies.

And I'll say two more quick things. It's very hard to launch remediation as I very well know. Bidding was successful in walling off demand, right? So if you launch a mediation now, you need to get certain networks and you need to get certain demand partners in and it's become very hard. Several of the large established players have said that they only work with a few approved providers and if you're not one of them, it's basically dead in the water and it's very hard to get them to integrate. I think that's a problem because it just makes it very hard to compete in mediation. It would be great to have more competition.

Neha
So, I'm a firm believer that if there is no competition within the industry, there's no way that you're going to push yourselves to do better, whatever it may be like service or price plans or whatever it is, just a general sense of things.

Coming back to your point about waterfall, how do you see the evolution of waterfall versus in-app real-time bidding playing out in your view? Is there something new to come after real-time bidding?

Ignacio
The answer is always going to be yes. Whether I know what's coming or not is a different question, but we've seen so many things change.

Neha
And when is it going to come? That's another question.

Ignacio
Exactly. And everyone's trying, though. I think bidding has definitely won at this stage. Waterfall plays a part. We, of course, continue to support it. But, you know, fewer and fewer players play in the space and the bulk of the inventory is sold into bidding. Most of the studios at this point are also working mostly with bidding. So I think bidding at this point has already won.

So when I talk about optimization windows, if you think about it, to some extent I am talking about like the waterfall approach, right? Being able to have successive attempts at getting a higher price so i know for a fact that studios would if they could do anything to increase prices so i'm sure that something along those lines at some point will show up again i do think that challenge today has more to do with the policies of existing mediation players and not necessarily with the technical ability to do this or the willingness on behalf of studios what i do think definitely comes next is i'll go back to user level setups i don't understand how studios have gone this far without having AdMon catch up with everything else they do. So even if you are the best mediation player in the space, it's unlikely that you will be the very best mediation for every single geo format placement user, like session, impression, whatever it may be. It just is very hard to imagine.

So building in some intelligence around, okay, so this is the specific scenario, right? This is a new user on iOS in this market. He's going to watch a rewarded video. What do I do? And you need to walk through all of that and then come to a conclusion and create a setup specific for that. We are there. We can do it. It's a matter of learning. And then the challenge here is that since no one's really doing it, at least that I know in a very like nuanced, granular user level, whatever way, like the technology is there, but it takes time for every single studio to figure out exactly what the right setup is, right? This is not just plug and play.

Neha
Do you think that this whole concept of learning about your users, I have my own opinion, but I wanted to hear from you. Is it a one-time thing or is it an ongoing process? Because the world is changing so fast. Attention spans are dropping. People are getting hooked onto one thing versus another. So is this something that you foresee that if I do it once and for all, I'm okay? Or you turn around and say, no, you need to continuously learn about your audience, your users.

Ignacio
I think you need to continuously learn, but not just because it makes sense. I think you need to continuously learn because you can. I'm still fascinated by what we've done as a society over the past 24 months, going from no chat GPT to basically relying on it for pretty much anything we do. That's where technology is right now, right? We have a completely doctorate level assistant in our cell phones and we consult them every single minute. That's the type of technology that we have at our disposal. Thinking that it's going to be a hard-coded solution for what setup is best for our user base makes no sense because we will lose, right? We're talking about the competitive advantage of studios here.

Because at the end of the day, everything that we're discussing is designed to improve LTV, not or not whatever, not CPM. Definitely this is an LTV like word. How does the studio make the most out of its users? And that then feeds into the way loop, which feeds into scale. You can go ahead and hard code anything you think makes sense when you look at it and study it. It will not last long and you will be losing out.

Neha
I've been having this conversation with a lot of people just coming on the topic of ChatGPT and I try not to use it as much as I should. You know, I know a lot of people rely on it a lot and like not just ChatGPT, just anything. Let's take Maps as an example. 10-15 years ago, if you asked me, Neha, can you drive from point A to point B? And at that time, Google Maps was unheard of. I would be able to tell you five different ways to get from point A to point B. But now it's like, wait, how do I get there again? What are the directions? How do I put this in a nice way saying I generally feel like getting stupider day by day.

Ignacio
To make it worse, you even get anxiety. If I'm going home from the office, so it's definitely a trip I know, but I'm not using ways to know if the traffic is going to be worse on one side or the other that's already a terrible experience so yes I agree we've gotten completely relied on these things but I will challenge you on one thing though because I've had this conversation recently and while I agree and there's been a couple of studies recently published about how let's call it chat GPT but of course the whole thing is affecting how we think and how we rely on it and at some point you'll even get like self-esteem, right? Can you do it on your own or can you not? But I would challenge that on the flip side, when I see what we've been doing internally within the company, and I'll give you a concrete example.

So we're pushing everyone to use as much tools as they can to figure out what the right fit for us is, right? And then of course, there's productivity issues and we're fast growing. So we need to take as much advantage of this as we can. But in my particular case, I'm a lawyer. I have absolutely no technical experience, but I started using these lovable and base 44 and all these, let's say, vibe coding websites, tools, the amount of internal tools and marketing tools that we have built, not just by me, but by people with absolutely no coding experience since then is outstanding.

And these are incredibly useful. And as the one of the persons that is benefiting from building them out, I'll be very respectful of people that design product because I'm nowhere near this. but it's given me insight into how to think about the product and how you build these things out and the user experience. So I feel like I've actually learned and become much more creative as a result of using AI than the opposite.

Neha
I've seen a lot of conflicting information. People are just like, I feel like I'm getting stupider, but some people are like, oh my God, it's just opened up a whole new perspective for me. There are things that I had no idea that I could do with it.

Ignacio
That's it for me.

Neha
It's fascinating. It's just about finding that fine line of balance where it's just like, yeah, don't make me sillier, don't make me stupider, but also please help.

Ignacio
But that's the other thing. As you know, I moved to Vietnam two weeks ago, right? So now I walk around the city, which I'm getting to know. And so it's fascinating. And one of the things I've started doing, which is really fun to try. So I ask chat GPT questions about Vietnam in deep research format, but I tell it to give it back to me in podcast form. And then I can listen to the reply. But I will ask it like, whatever, what's the history behind this neighborhood, right? So it's a podcast built for me. Really fun.

Neha
That is fascinating. Why has that never crossed my mind before?

Ignacio
It's really good to try.

Neha
Ironic. We're talking about a podcast while on a podcast.

So do you actually believe in linear digital customer journey driven by ads or is it more about loops and detours and all these kind of micro decisions now?

Ignacio
As much as you can figure out loops and detours that make sense, you're definitely sure. It's tough, right? I mean, let's be completely honest here. We can talk about user level things but the truth is given that we're doing it we're giving everyone the same experience jumping to the user level is a long shot so at the beginning just setting up some very basic cohorts already makes sense and you need to make sure that the incremental gain that you earn from the different loops actually pays off i'm not trying to tell people that you need a billion whatever setups probably not But there's definitely better than just one.

Neha
Okay, so what kind of role do paid touchpoints like ads play versus owned ones, websites or apps in shaping sort of a holistic user journey and driving monetization?

Ignacio
By own ones, do you mean cross promotion types? So when do I occupy the space with my own inventory as opposed to third party demand?

Neha
Yeah, that is what I was trying to ask.

Ignacio
That's a great question, which I've actually spent a lot of time thinking about in the past. It takes me back to my marketing days. But if you want to do cross promotion successfully, and I the worst person to ask because I don't know that we were successful, but I'll tell you what we tried and did. And you know, I'll tell you what happened. Essentially, the way we thought about it and the way we think about it is you want to figure out what it makes sense to cross promote. So in our view, that made sense mostly when we thought the user was about to leave. So we built a churning model, right?

So we were trying to predict when that user was likely to disengage in the near future. And at that point, we would start trying to convince that user to play a different game. In our case, the challenge there was and I got into trouble with my current co founder back then who was not my co-founder, because what we did was we started using at the time brace essentially to launch notifications into the app pitching our other games. And what they were saying, which was correct was, look, that's just another interstitial ad. The user doesn't know. So if there's more placements for ads in the game, then I should also have a shot at buying them. So what we ended up doing was we'd have to buy our own inventory, which was tricky. And it, which is why I say, I don't know that

I'm the right person because we were never really able to do cross promotion at scale with this model. So maybe it works and we just didn't fine tune it well enough. But I do think it makes sense. Again, we're optimizing for LTV. If the highest LTV is one where we're just showing ads, then maybe cross promotion is not right because you're cross promoting something that doesn't make sense. It also has to do with figuring out what you should show who and how.

Neha
And there's so many variables. You've mentioned twice about predicting what the user is probably going to do next, et cetera. And I have a question. Did you use AI for that? Or was that something that you went and studied and made a prediction based off your brain?

Ignacio
Definitely back then, no. But honestly, some of the predictions are so obvious that if you were to try and guess, my right-hand person right now within Xtreme is our former head of data at Etermax. He's extremely talented. And it's always frustrating because I keep coming to him with these crazy ideas about things that we can do with AI. And he's like, no, that doesn't work. I figured out that the people that most understand AI are the ones that make sure that they're honest about the limits and what you may and may not do easily. Right. Or when it even makes sense to use it. Sometimes he's like, I can get to the same result without needing to train a model. It's so simple. Let's just figure out these two or three variables. If that works, don't call it AI. That's all simplified. So sometimes it is really just a much more simple approach and you can still get to a decent result. Of course, that was six years ago. Today, I think there's probably tools that make it much easier.

Neha
So a lot of brain was used to not, not a lot of brain and spreadsheets.

Ignacio
Yeah, past behavior and so on. But what I'm trying to say is we're so far away from needing very precise predictions at a very precise user level and so on that today you can kind of come together with two or three variables and I'm sure you can do better than what you're actually doing right now with no prediction whatsoever. So don't get drowned in figuring out what the perfect one is because you can do part of the job fairly easily.

Neha
Interesting. I love that you went old fashioned, you know, brain and spreadsheets.

Ignacio
That was six years ago, to be clear. Today, not so much. Today we would be pushing for AI, I would guess.

Neha
I still use my brain and spreadsheets. I'm one of those until I really, really need help. Then I'm just like, OK, maybe I will allow myself to get assistance from artificial intelligence.

Ignacio
I'll give you an example which we found recently, which was extremely interesting to us. So we sat down in January of this year and we said, okay, so brainstorming session, what can we do to grow as a company, right? Aside from what we've built, like what should our next big step be? What should we be thinking about? And basically everyone that we talked to outside the company was pushing us, especially back then to build a network. That is a big no for us. We philosophically feel, and even being strategic and selfish, let's say, I think that there's much more to win by remaining true to our mission of supporting the supply, right? So we don't want to build a demand source ever that's written into our DNA.

But we said, okay, so brainstorming session, if we wanted to build a network, what would we have done? What should we do? And the answer is, we should probably train a model to figure out who the VIP users are. And we should probably train on the data we have access to right and then we said okay fine we're not going to build a network but is that model worth anything to anyone regardless and the answer is yes so we did it and the first iteration of that model can very easily figure out who VIP users are within the first session of a new user before he even sees his first ad so we're not just simply tagging those users and sending them back to our clients saying, okay, it used to take you seven days to figure out who was a likely buyer. Here you go. This 10% of your new users that we've tagged, they are four times as likely to convert to paying users. And these are the exact numbers that we tend to find on average. So 10% Forex is likely. So it's incredibly powerful, but it was the first version of the model we trained. It wasn't really as, it's going to get better over time, right? But that was launched within, we had the conversation and I think within a month we were able to launch the model.

Neha
Okay, so coming back to these VIP customers, how do you define them? Like, is it their usage history, their intention signals? Is that what you call them? Basically, what categorizes someone as a VIP user?

Ignacio
That's the interesting part. The studios are best positioned to identify a VIP user according to everything you just mentioned, right? Because they have all the access to the user behavior and they figure out how that user, what he's doing or she's doing and what that means. We tend to focus only on AdMon data. So what do we predict the CPM for that user to be? Who do we predict will be buying it? How do we predict networks in general will behave? And then when we run enough requests, we can figure out how far off we were and we can read into the difference between our prediction and the reality and that is what feeds us.

Neha
One last question — you've done all these things and you mentioned that you moved to vietnam recently so both extreme and lumen are registered in argentina but you've moved to vietnam two weeks ago do you see a significant difference between these two markets?

Ignacio
Argentina is an amazing country when it comes to tech talent. It's an amazing country when it comes to building products, especially in technology. We're fortunate as Argentines to have had, I don't know, I guess a couple of companies early on which were very successful and which invested into the ecosystem. And as a result, there's been a lot of tech in the country for a long time. It also has to do with the fact that we have our economic cycles, which tend to be a little bit more volatile than others. And as a result, we are export focused, right? Working within Argentina for, you know, exporting tech is a great way to kind of stabilize that volatility. So there's not many large gaming studios in Argentina, unfortunately, but there is a lot of tech. So we were fortunate to find extremely talented people that were willing to bet on us and work with us and grow the company. If you become extremely westernized, you forget that more than half the world's population lives on the other side of the world. So, you know, it makes sense to come. Vietnam is similar to the extent that they are great at building products and exporting tech. They are different to the extent that they have a ton of developers building, especially within the gaming ecosystem, at least, which is what I've seen so far. They've built some amazing teams. So, you know, we target the same user base. So to some extent, the product itself doesn't need to change as much. What does need to change and which is part of the reason why I'm here is, to be honest, it's incredibly interesting to me. And then, you know, the food, the coffee, there's so many things which I can talk about Vietnam, which have nothing to do with work, but which are extremely exciting.

Neha
Did you ever visit Vietnam before like saying okay I'm moving there or you said okay gaming industry let's go?

Ignacio
Yeah, it's funny — the quick answer is no. I realized I woke up on a Saturday morning probably in like May and I said I think I need to we need to move to Vietnam and so my wife she was somewhat concerned she'd never come here before either and I was leaving for MAU soon. So she said, okay, fine, but you need to go to Vietnam. You need to tell me that it's okay. And you need to find a house. So I actually came here a month ago for literally like three or four days. I came to Hanoi and Ho Chi Minh. She ended up choosing Ho Chi Minh.

Neha
Thank you so much, Ignacio, for joining us and sharing your journey from finance to ad tech and for unpacking sort of the complexities and intricacies of ad mediation and platforms. Thank you so much.

Ignacio
It's been really fun. I really enjoyed it. Thanks so much for having me on.

Neha
If you've enjoyed this conversation, make sure to subscribe to Points of Growth wherever you are. We've got more insightful episodes coming soon. Until next time, thanks for tuning in.