Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Incrementality & App Programmatic - Remerge Deep Dive | Pan Katsukis
35 Plays4 months ago
SUBSCRIBE TO LEARN FROM PAID MARKETING EXPERTS 🔔
The Efficient Spend Podcast helps start-ups turn media spend into revenue. Learn how the world's top marketers manage their media mix to drive growth!
In this episode of the Efficient Spend Podcast, Pan Katsukis, CEO of Remerge, dives deep into the intricacies of incrementality testing and its critical role in modern marketing. Pan shares insights on optimizing media mixes, avoiding cannibalized spend, and navigating the evolving challenges of privacy regulations in the ad tech space.
About the Host: Paul is a paid marketing leader with 7+ years of experience optimizing marketing spend at venture-backed startups. He's driven over $100 million in revenue through paid media and is passionate about helping startups deploy marketing dollars to drive growth.
About the Guest: Pan Katsukis is the CEO and co-founder of Remerge, with over a decade of experience leading programmatic advertising and app re-engagement initiatives. He is passionate about leveraging data-driven strategies and incrementality testing to help global brands optimize their ad spend and drive scalable growth.
VISIT OUR WEBSITE: https://www.efficientspend.com/
CONNECT WITH PAUL: https://www.linkedin.com/in/paulkovalski/
CONNECT WITH PAN: https://www.linkedin.com/in/katsukis/
EPISODE LINKS:
https://www.remerge.io/incrementality
https://www.remerge.io/methods
https://developer.apple.com/app-store/user-privacy-and-data-use/
https://www.google.com/ads/data-driven-attribution/
https://developers.google.com/ads-data-hub
https://support.google.com/analytics/answer/1008015
https://www.adjust.com/resources/insights/ios14-idfa-privacy-report/
Recommended
Transcript
Incrementality as a Goal in DSPs
00:00:00
Speaker
but I think like that's a goal because you have like then always this bucket that drives like a lot of incrementality, a lot of incremental results. yeah Those users at that time, and this is I think like the most valuable one you just want to use for your algorithms, then across the board to have an understanding like, okay, why is it that these customers do like so much additional incremental value compared to others?
00:00:23
Speaker
um So this is something where things are working really well and you can have like, you you can dig deeper to understand it and then use those insights and learnings for your models. So yeah, then that's why I think like as a DSP, it it definitely helps in running a lot of incrementality tests to just like have this transparency and then use it also to to to fuel it and into your ah model compared to just like looking at attribution models with all the challenges that attribution models have.
Remerge's Market Focus on Incremental Conversions
00:00:57
Speaker
One of the unique things about Remerge in the market is your focus on optimizing towards incremental conversions and your incorporation of incrementality testing kind of into into your ethos. That's something that a lot of marketers struggle with. And you know we're we're running Remerge right now at at itself. We actually did a MMM recently at one of our first models.
00:01:23
Speaker
and saw pretty good results with with Remerge. When you speak to performance marketers about this topic, what are some things that you know you kind of communicate to talk about how to optimize towards incremental conversions and how to kind of protect against cannibalized spend, which is you know a really big challenge in our industry today?
00:01:44
Speaker
Exactly. Like I agree. I mean, being in our position, obviously we can do what we can control. So we can't control the spend you have maybe on other different channels. But that's the point. Like I think that's where incrementality comes into play because we can get transparency in terms of yeah what you're what what else you're doing independently of what you're doing, but we can just like see yeah what is really incremental, what we do on top of everything else.
Limitations of Attribution Models and Transparency
00:02:08
Speaker
right So I think that's ah that has always been a key question for a lot of marketers, specifically when when you're talking also more about re-engagement, re-targeting, because there are already users. You're already in this kind of hole of what the user is doing. Are they coming back organically? Are they doing a purchase organically? And even if they do a purchase, like it's, it's what they're purchasing. Are you able like to help them to spend more or or less? So there are a lot of questions, but you actually can't um ah find out with normal attribution because normal attribution is like comparably simple. Yeah. Like rule-based is something, did something happen with this channel? Yes or no. And if yes, then okay. Almost everything is like 100% attributed to the channel most of the time.
00:02:51
Speaker
So like what we as a specific channel can do is just like, look, and that's why incrementality is I think like a good approach because you just the run an experiment where you can see, okay, what would have happened actually with when, if you're not running with WeMerch versus you're running with WeMerch. So, and you can, you can look at that at the results, how much of an impact ah is there like exactly and, and, and how much is there like not. And customers then can use this information to then calculate Is it actually
Adjusting Attribution for Better Media Mix Management
00:03:21
Speaker
worth it? They can even like, I know a lot of customers, what they do is then they adjust the attribution logic more accordingly. So to balance it a little bit better, specifically when you do that approach where across similar channels where you can adjust the attribution just settings, then yeah, if you run incrementality tests, for example, and then see like, Oh, but maybe you have waited this channel a little bit more than the other channel.
00:03:45
Speaker
But that's exactly where can ofbi cannibalization comes into play. yeah Which channel cannibalizes ah which one, right? So with incrementality could get a better feeling and then try to fine tune the attribution settings so that you don't need to run like consistently incrementality experiments, but then like always do it occasionally, maybe once a quarter or so once a month and yeah have an understanding like what's going on actually in your media mix there. up now So this is maybe, yeah, just like as a summary, our approach and and where we can support the customers to get more transparency on the bias? I think measuring incrementality testing is relatively straightforward in that there's multiple approaches to do that and then you can get results from that. However, optimizing and bidding towards those conversions is kind of ah another conversation. So you you know when you think about the applicability of the results of an incrementality test
00:04:43
Speaker
Some of the outcomes, like you mentioned, are calibrating your attribution model. Another outcome of that might be calibrating your media mix model. And then, of course, informing budget changes that you want to make, whether in a specific channel or across your mix. However, the the part that's missing that I think is really interesting is how do you leverage that data to feed back into the optimization process?
Leveraging Incrementality in Programmatic Advertising
00:05:10
Speaker
Meta has started to look at some things around this where they recently you know released an update that you can optimize towards incremental conversions via their publisher lift tests. Is there the ability to do that in the programmatic space and is that something that you're looking at?
00:05:26
Speaker
That's one of the goals, right? Specifically because you can and understand, you can control the results of incrementality tests on a broader scale, yeah where you're just like, the challenge I think like for, for, for getting there is to run a lot of incrementality tests. Yeah. So, and because that's maybe not as often like, uh, uh, running just normally ah towards attribution.
00:05:48
Speaker
This is where like where when you want to optimize in a really good way becomes a bit tricky. So that's why like maybe you need a size of meta where you have like a lot of data, a lot of understanding um so that you understand like similar patterns that works across like the ah customers maybe across different spends across different times. So across just like different signals.
00:06:08
Speaker
And, but that's exactly the goal because like not even, so what you try to achieve also is not only to understand which uses at the best time and what message is is the best right now, but also to cut off the ones that actually provide like a negative incrementality. yeah So actually everybody who would have converted organically and you're not contributing to anything extra, you actually don't want to then touch them, right? I think like that's the the interesting aspect of that. and Yeah, but but I think like that's a goal because you have like then always this bucket that drives like a lot of incrementality, a lot of incremental results, yeah those users at that time. And this is I think like the most valuable one you just like want to use for your algorithms then across the board, right? To have an understanding like, okay, why is it that these customers do like so much additional incremental value compared to others?
00:07:00
Speaker
Um, so this is something where things are working really well and you can have like, you you can dig deeper to understand it and then use those insights and learnings for your models. So yeah, then that's why I think like as a DSP, it it definitely helps in running a lot of, uh, incrementality tests to just like have this transparency and then use it also to, to, to fuel it in, uh, into your ah model compared to just like looking at, um, attribution models with all the challenges that attribution
Remerge's Advantage in Low CPM and New User Discovery
00:07:27
Speaker
models have.
00:07:27
Speaker
you know as we As we think about programmatic as as a channel or a channel category in relation to other channels that marketers are running when they're scaling their their media mix, it offers some some unique advantages. I mentioned before that we had some positive results from the test that we ran with Remerge. One of my hypotheses for why this was is that there is the ability with remerge to get really low CPM kind of high impression traffic. And perhaps that leads to identifying a lot of net new users that would not have seen our ads otherwise. So that's just one hypothesis hypotheses ah that I have for that. But
00:08:12
Speaker
you know As you think about programmatic within the context of a larger media mix, how do you think marketers should think about kind of the the balance there? When we're thinking about you know we're running our paid search campaigns, we're running TV, we're doing some some app programmatic. what What place do you think that app programmatic sits within that in that space?
Programmatic's Role in the Media Mix
00:08:33
Speaker
Like if you imagine the programmatic inventory, it's like all the apps that are being used by by people outside of like the the classic ones like Facebook and X and so on. um And like, for example, like all the games yeah where you have the opportunity to show like our new sites or whatever, there you you can get the user there maybe also very likely much more cheaper than than on the the same user on Facebook.
00:09:02
Speaker
And I think like that's our Instagram. And and that's ah also something like depending, especially when you're looking at performance where where this can make the difference. yeah yeah Because Facebook instagram might already be too expensive to exactly reach that ah that user. So because like you wouldn't hit maybe you your goals. So that's what where programmatic can always like fit in to just like be able to expose those users with your message and with your method messaging to to target them. yeah yeah You just like tap into all the other whole app ecosystem app store. There are the apps that are out there where you can have the opportunity to just like be able to bid on specific users. So I think this is yeah just like some source where you don't want to miss out tapping into. And yeah that that's, I think like ah where DSPs can help in leveraging it.
App-Specific DSP Challenges and Optimization
00:09:50
Speaker
And then within the DSP space, obviously there's, there's a lot of, you know, competitors that, that you come across. What is unique about what Remerge is doing?
00:09:59
Speaker
Yeah, so I think like what what what we do maybe differently is, is I think like that we are focused on app specific right at first so we don't touch like and and any web and we have done it from day one. So I think like everything the whole technicalities.
00:10:15
Speaker
Um, the whole integrations that we have, the goals we drive for our customers are really, I think like optimized for the whole funnel inside apps. yeah And I think like also for apps themselves that they are very engaging for users that you can just like do more profit with that. I think like as a, as a brand is where we have always tried to just like optimize the funnel, the flow but for that. Yeah. and and all other, like what we discussed, the incrementality systems, also like how we deal with less signals now on iOS. yeah this is ah This is something which is which has always been like a key challenge for us and and for the industry, ah where we have invested like a lot of R&D resources specifically to make that achievable, right, to to to make it successful.
00:11:00
Speaker
And I think you have but winner when I said that we come from the retargeting re-engagement space and that we work with so many broad different types of customers, helps us also in understanding like the behavior a little bit better on apps, on the phone, and and what works from the past 10 years, what we have been doing, which we can also use for our models um to just like yeah leverage that ah a little bit better.
00:11:26
Speaker
um So I think like the the but um that's not like a specific feature, I would say, that that makes a, that differs us from a lot, but I think like ah specifically to just having done that for so long across the world, ah across so many different types of customers that so many users let me know is I think something yeah that helps us in driving good performance.
Evolution of Incrementality Testing
00:11:51
Speaker
Yeah. And it's much a ah philosophy as well. You know, I think the idea of optimizing towards incremental conversions has become much more top of mind for for folks. But even now, you know, I get on calls with with reps from from certain channels.
00:12:06
Speaker
that might say something like, you know we need to increase our budget here based on this CPA. um And it's you know a specific campaign, which I know that is not incremental. And so like the way marketers make decisions by leveraging different measurement methodologies and results can really shape a media mix and you know can can be really challenging if you're using the wrong data. And even within you know leveraging things like incrementality testing, if your test design is incorrect and you are not clear on the assumptions that you're making, that can have wide-reaching effects.
00:12:45
Speaker
Yeah, I think so too. Like um when I think back when we did, it's also, I think like six, seven years ago when we started our first incrementality tests. And I think that also evolved like how to do it and to be able to tap into everything. I think like even with with with iOS, where we have been running like huge scalable campaigns for big brands. where you don't have an ID, where you just want to see, okay, let let's just tap into the inventory. Let's just see what's happening there. Because like ah when iOS came out with the ATT infrastructure, like nobody, you had something such a cheap inventory because nobody was buying it because it was not really measurable. So we were just running incrementality campaigns and and would just like see, okay, let's just buy the inventory. Let's show all the message because the user behind
00:13:31
Speaker
But behind that, it's still a valuable IOS user. So, and then just like trying to to understand that. And I think like being able to navigate the industry and and the the whole supply with with get with her um a good understanding of what the impact is. and It's definitely helpful to just like optimize and get to get better, get faster and drive value. you um So.
Adapting to Apple's Privacy Changes
00:13:55
Speaker
Yeah. And for me, it's always very exciting because I think it's always very negative or perceived negative when Apple and Google are changing the rules. For me, also as a founder, entrepreneur, it's more like, oh yeah, it's the next level. So we have to just rethink again and maybe you can't just stay still and not do anything more. You always have to rethink the challenge, what to do and adopt and hopefully make like this as a competitive advantage. So yeah, this mentality is I think also very important to just like be able to survive in the industry. So because I That's what I told also the team like 10 years. If you look back at the logos we had 10 years ago in the industry, it means already something that your logo still exists and is not like being eaten by other players or that you got bankrupt or whatever. So it's not so easy.
00:14:50
Speaker
Sure, sure. I mean, I remember when iOS 14.5 was released and the changes that we saw, you know from an attributed perspective, and then you know we're seeing our cost per installs and and all of our metrics go up. And we didn't adjust our budgets down dramatically. And a lot of it was paying attention to what the the market was doing in the supply and and it was this very interesting scenario where if you had conviction in some of these different areas and you could deal without seeing the measurement and signal, you were actually taking advantage of a supply and demand kind of issue, right? So, yeah.
00:15:32
Speaker
I mean, i like also just to point out because that was always like one of our core hypotheses. I mean, when when this all happens and I think also going forward, when you have like less IDs and signals, what's all going to happen because the users behind it, they still, I mean, they don't change so much. They are still the same ones, but it's just like, okay, how do you.
00:15:49
Speaker
Effectively just like target them and and make it be able to leverage it. Yeah, yeah but but that's exactly the point. but Somehow it needs to balance it out, right? That you can buy because like it it gets so much cheaper because nobody feels comfortable and tapping into this pool of light. Yeah.
00:16:06
Speaker
no idea environment. But if you are the first one to to jump into the pool and swim through it and then see like, maybe's maybe there is value. you Yeah, nobody else that you have like the whole lake for yourself. And um yeah, you can get get something out of it. Even if it's maybe a bit more expensive to to just like explore.
00:16:24
Speaker
Um, but yeah, because it's like so much cheaper, it might also be good. Yeah. And this is like where you learn faster and, and you can just like leverage that better. So I think, uh, yeah, being on the forefront and, and just like be open to, to check, to change and and try to test the experiment.
00:16:39
Speaker
And always ah exactly like you said, like always double check, is it incremental? Does it make sense with the impact on my end? I think it's it's the right approach and and you need to do more like this compared to the past where you just like had one attribution provider and then everything is good and easy and you can go. 100%, I couldn't agree more.
Apple's Attribution Framework and Remerge Strategy
00:17:00
Speaker
And, you know, we've talked a lot about the historical changes, not to undermine the changes that are coming in the measurement ecosystem. Obviously, I want to talk a little bit about what you guys are are doing with Google, but you mentioned Apple first. So I am curious, you know, Apple recently announced a new updated attribution framework at attribution kit. There's some interesting changes there, perhaps on the retargeting side as well.
00:17:26
Speaker
What has Remerge's approach been thus far to that? We are always very open and have implemented and then try to optimize towards a scan in the past, like specifically also scan for quite a lot. And I think like even when you look right now at the single their singular has like a dashboard where they see, okay, which partner has the highest scan for impressions and support. I think we are always number two behind Reddit and I don't know why Reddit is like somehow ahead of us, but yeah. And so, so we try to just like, um, um, support it as good as we can. And also the new attribution framework they're rolling out. I think it's not changing for, for like practically and not really much. Uh, it just has like a, I think a broader view of what can be done.
00:18:11
Speaker
Yeah, but our hypothesis is also like, I think Apple tries also to just like give a little bit better tools for navigating to, to the marketers and they want to improve. ah So we also hope that by supporting it, by, by, by also learning how, how measurement is happening there, that we can get better and and at lead it at least like also more data points ah on top of anything else. So I think, yeah, we we have, we have a team just like focusing on that. We also have like a good context, I think, ah with the Apple scan team. So that's, ah I think also something but that can help us on the retargeting point. Yeah, we have been also looking about it. It's a sort to just make it clear. Apple is not going to create a framework. I think we're targeting is getting better. So.
00:19:00
Speaker
Scan and then the attribution framework is mainly for measurement and that's also where the retargeting or like the Re-engagement aspect as a metric will come into place. So ah you can't select I think IDs or audiences where you can specifically target but when you just like run campaigns and there is an attribution You can see if that user or that audience had the app already installed before. yeah So that's like technically a re-engagement and this is what you then get the transparency there. So in the past you saw all like okay new installs independently if they had the app before. So that's ah but what what is changing on the Apple side.
00:19:40
Speaker
so Maybe that's why I said like, it's not like crazy big. And that's also not to expect that they will have something like what Android or with the Google is preparing with a sandbox to have an alternative targeting framework um that is more privacy friendly.
00:19:55
Speaker
And maybe before we get into Google, what's what's your kind of opinion on the differences in approach between Apple and Google on on this as relates to to privacy and also kind of offering you know advertisers the ability to run effective
Comparing Privacy Approaches: Apple vs. Google
00:20:12
Speaker
campaigns?
00:20:12
Speaker
Yeah, I think everybody knows Apple is not very cooperative with the industry and like achieving their strategic goals. ah And um I think there is like for Apple, it's like clear. And even if you look specifically in Europe, you have a lot of advertisements that I saw even during the Olympics, if you watched ah TV there, it's just like, okay, they, they just go all in on the privacy front and everything else is like not privacy. You're getting like spied on the whole time by third parties and so on. So this is like key message. And I think that became clear. So their strategic goal is okay. Just like tried on the consumer side. to be as aggressive as possible and block the advertisement practices, how it has been done right now. yeah and and that not I mean, they don't care if there is no alternative framework for for for for everybody that maybe benefits from advertising, because also Google Apple doesn't have any share in that, so it doesn't hurt them. And they just go forward. There is no intertrust case that is risking them, which is like all true for Google. And Google is 100% the opposite direction where it's like, Okay, complete consent from the industry. If there is no consent, they feel very ah challenging to go forward. So they're very cautious. And you know, like also even on Chrome, it takes like four years. ah But even if you look at Chrome, they could just say, well, Chrome has nothing specifically as a as a platform to do with advertisement in the first place. And all other browsers have already
00:21:43
Speaker
like cookies switched off is only chrome but even for that one to just like do that switch it takes them years because they have the requirement um to just like have an alternative framework where advertisement companies can still thrive because like otherwise the hypo hypothesis there everything goes or the budgets go to google and advertisement that that's like a tricky case yeah so two completely opposite
Privacy vs. Personalization in Advertising
00:22:08
Speaker
approaches. Yeah. But everybody like both say they want to be like really privacy first, privacy centric. And yeah, like if you dig deeper, neither Apple is like obviously hard or hardcore privacy centric nor Google obviously right right now. But yeah, that's how it is.
00:22:23
Speaker
Yeah, there's there's this very you know weird interplay between the the need and desire for more privacy from a consumer perspective, and then also the importance of personalization and context, not only from ah you know an advertising perspective, but you know can consumer apps, you know the everyday mobile usage of your phone, um chat GPT. If I'm able to you know enable more data sharing, less privacy with some of these tools, my experience becomes fundamentally better. And I know that that's that's kind of a balance that that you're striking with some of the work that you're you're doing with with Google as well. um what what What do you think about that balance?
00:23:10
Speaker
No, yeah, i've I agree.
Remerge's Data Philosophy and GDPR Alignment
00:23:11
Speaker
Like I think I said, like from, um from day one, also when we, when we started with re-merge and our philosophy on that. So like we have never, and and and even, even right now, like when we talk, when we work with different customers or so, we don't just like mix some and enhance an ID, uh, like, like maybe others have been doing. Yeah. Where you have shows, okay, you have an a user and then you just like put everything on top and this like kind of, you have a better understanding whereby.
00:23:37
Speaker
when you could just combine different campaigns and, and, and, and faces and you're creating this user profile for like, and, and our approach was actually always because we work with the like huge enterprises. We want to keep the data separate. No IDs are just like being mixed mixed and matched, but yeah, we keep the data in silo. And, and obviously then, well, since 2018, when, um, um,
00:24:02
Speaker
when GDPR really was enrolled and forced, then you anyway had like to do to do that. And even then, you you know yeah have the categories being data processor, being a data controller. And I think 80, 90% of platforms out there are data controllers. So once the advertiser works with them, they control their data. So it's like theirs kind of, right? And they can use it for for for dealing with that.
00:24:25
Speaker
We have always been like a data processor because this was also like our philosophy where just like, okay, we are an extension, we are a tool where it's the, it's the customer's data. And like, like we do everything just like on the customer side and, and that's that. Yeah. And I think like going forward and yeah also the ID.
00:24:42
Speaker
Uh, for us, it's more like a technical instance to just like be able to match, to, to match IDs. Yeah. So to match users like technically, okay, this is how between supply and demand, you can, you can do a match. Um, but we don't necessarily need to, because like, even, even with that, most of the time we work with audiences, we work with segments, we work with, uh, um.
00:25:02
Speaker
Yeah, like a group of users that have like similar characteristics where we then say, this is what we want to bid for those audiences. This is the goal altogether, what they want to achieve. And, and it's like a couple of thousands users or something or hundreds of users. And that's already good enough, like for most of the cases. Yeah. But technically you needed to do an i idea to do that matching. Now with a Google approach and I think like the key.
00:25:24
Speaker
High philosophy of that is just, okay, just don't do targeting based on a specific ID, but just do it on they're based on audiences. Yeah. So you always have like a kind group of users and you don't know specifically which one it is, but those group of users, yeah, they, you can still do targeting. You can still do.
00:25:44
Speaker
Measurement and and so on it is just like design that you don't conscious reverse engineer to i exactly understand Okay, that was this one specific use that I'm still able to create like my profile around it Maybe I get I'm able to to get some address and home address and work address and this and that right so I That this approach is not able to be done anymore. So I think like the approach definitely makes sense. It's obviously like technically not as easy because you have to change. And I mean, because we have like one also R and&D team just focusing on sandbox and you need to just like rebuild the complete stack in order to be able to do so. Yeah. So it's not like. um
00:26:24
Speaker
ah Simply, oh, you change the one ID with one audience ID and then you can just like left and right match it. Obviously, technically it's a bit more complicated like than that, but I think high level philosophy wise, that's the approach. And I think it's a good approach, but yeah, like doing it while the whole system, the whole machinery of multi-billion dollar industry and digital advertising is is ah used to it. It's like super tough. ah the But still, I think like it's good that Google tries mainly to get everybody together um and then just like invest heavily also to support everybody there to do testing, to ah connect each other, to drive things forward. and and then And it's like, I think also their philosophy to to let the industry
00:27:06
Speaker
move that forward. So they try to like give a general guidance, but how the details look and look and what needs to be done, um it's figured out by each party with each other. So we also try to drive together with our partners, with the attribution providers, with the SSPs, um yeah the the the interfaces that it works.
00:27:26
Speaker
the APIs that also this more privacy friendly approach on sandbox is going to work. Yeah. But it's like a heavy investment for everybody. And the problem is it's like short-term it's not incremental revenue. It's just like, okay, you just want to like long-term somehow to survive and in in that environment. Right. But yeah.
Collaboration with Google on Android Sandbox
00:27:46
Speaker
How long have you been working with Google on this and and is that unique differentiator to Remerge compared to you know other DSPs and the effort and investment that you've put into this?
00:27:58
Speaker
Yeah, I would, I would say so. Like I think if you would talk with Google directly, Google says that we are the biggest contributor on the Android sandbox ah specifically. So usually in the media, you hear more like the Chrome because it's more imminent with the third party cookies. So you have like other players that have been contributing there, but they also don't touch actually Android so much because it's not as as much of a, of a business for them anyway. So so for us, obviously, because we're app centric, Android is the key element and, um, I think when it comes about like the methods, the the comments and and and and the whole documentation, we are the biggest contributor. And we do that now, I think like since almost two years where we have like one team completely, I think it's like six developers just do sandbox that have every two weeks they have a call with Google. um We talk like regularly with all the
00:28:48
Speaker
partners in the industry and try to move them forward to do testing, to do experiments, how it works, to have the whole technical setup up and running. And yeah, like I think why why we believe it's also strategically ah important because in in the last one and a half years, like we are shaping exactly how it should look like. So we know the use cases for all different advertisers, what we have, and like the integration with our partners. So for us being involved and sitting having a seat at the table and say like, oh, you have forgot, you forgot this, you have to add this so that it works is important to for us. And I think like for all our customers, and I think like that's good also going forward. If you work with a DSP that just like starts doing it and implementing it once it's all being done.
00:29:30
Speaker
Then it could be that things are missing. Then it's like difficult to, to adopt. And that's like what we try to avoid and and just like, yeah, really have a good sense of how it's evolving and then like have a good transition that the performance campaigns are working also. in the future What do you know about, you know, when this investment is going to start paying off, meaning, you know, when these tools are going to start getting published and and implemented? Yeah. So what we, I mean, obviously there is a timeline why it's always like a bit tricky, but for us, endbo and I specifically on Chrome, because they went like, they say they will get deprecated. So it's, it's happening exactly then. And now, yeah, it's there like in a difficult situation because also of the antitrust things that where people just like saying, whoa, you can't do that.
00:30:13
Speaker
um And on Android, I think it' ah it has been a bit more clear and and ah not as like crazy with with saying like, oh yeah, the idea is going to get deprecated by Q3 2025 or something like that. So then they have never announced like a deprecation.
00:30:28
Speaker
Uh, date, yeah they even haven't, they've never actually said they're going to deprecate. There's always what they, what Google's as underlying when, when, when we talk about it. And so they, they want to be like on on that side or safe side, ah but but we' are talking more about like practicality. So I think like they said by end of this year, early next year, all APIs will be rolled out. And I think like also then it's, it's getting a bit more tricky to change like core elements of the approach.
00:30:53
Speaker
And once the APIs are being rolled out on all Android systems, and I think it will also be fact ah backwards compatible. So you have then access to all the Android devices or like a huge ah part of the Android devices out there where you can then run campaigns on sandbox. here So that's, you will have a live environment which is scalable where you can test it. and And the good thing is like still the whole ID infrastructure will remain and So you will have at least like a year where you can already start testing and ah can can compare and already just like optimize the model. how How would be the results look like? How would the results look like when you would run it on ID? How would it like one when you, when you run on sandbox? And then you can see already my like will sandbox have already. some advantages even compared to the i idea. One advantage is obviously you can't opt out of the sample. Everybody is addressable. It's like all the same rules for everybody. So like if, but because even right now you have like maybe some Android users that don't share an ID, yeah you can still set it off, like go and switch it off in your settings. um ah So now the
00:31:57
Speaker
they they get a bit more, better addressable. So that's, I think like some interesting aspect how it can get lab. Plus you will also get like additional signals with some of the APIs. Google will also provide like with the topics, APIs, there will be new signals. um So I think that it will be exciting and that all will happen next year. Next year may be like still very technical, but like at least like the first half of the year, the second half of the year, I think you can really start comparing and seeing some good business cases and scalable campaigns when we have like a better infrastructure, when the SSPs have that implemented, when attribution is solved. If technically everything works, I think that would be exciting. And then if that works, yeah, then we'll see what Google says next, if they're really going to change um
00:32:40
Speaker
their approach on how, on IDs. I feel right now, like on online, they are also saying, well, okay, they are not deprecating third party cookies completely. So there will be a consent and the user decides like similar what we have with ATT. So like, if you would ask me, that's maybe something that could come in also on Android, where you will have like at at the beginning or if an app is open, whatever, where there will be constant ask, like, you want to share your ideas or no. If not, you will have like still the sandbox.
00:33:07
Speaker
um But at least like it's likely maybe at least at the beginning that we will have like, and you will have IDs or some IDs to target. And for the others, they might opt out like you have on Apple, but this is hypothetical. So we don't know exactly, but I think something like that will definitely happen.
00:33:26
Speaker
Right. And I know that Remerge produces a ton of great content on this stuff on your website and blog. That's definitely something I'll include in the in the episode links that that folks can check out to to read more and and learn more about these changes. Speaking of hypotheticals with the last couple of minutes of the podcast, love to get some predictions from you on on what the future might look
Generative AI in Creative Processes
00:33:49
Speaker
like.
00:33:49
Speaker
One of the interesting things, I think there wasn't an attributed author on this, but there was an article on your blog blog about kind of some future changes and future productions to the the industry and some interesting thoughts about how generative AI and programmatic may work together to kind of create a more automated creative experience.
00:34:13
Speaker
um How do you see that kind of going, the the collaboration or the integration of generative AI into you know the the work that you do and the advertising that you do on your platform? No, exactly. like i think yeah When you talk about generative AI, I think yeah creative is what everybody's thinking about. do You have a lot of ah work and um i mean we have an own creative um department and they are also experimenting and like working on it. and What we see directly and what we have done
00:34:43
Speaker
I think the the first thing was specifically when looking at video creators and to just like doing them more dynamic and also localizing. Yeah. So I think like when you have global customers, specifically games where it's, it might be like very difficult to get like dialects of voices and and things like that, or even texts.
00:35:04
Speaker
text to get localized, um, different processes, process steps can just like be, be faster with generative AI. And then at least like the, the, the, the people who are proving that who are just like double checking, it's mainly like that part. Yeah. But you can, you can have. creative, set up, tested, experimented much faster than in the past, that specifically because I think everybody's aware when when you create a video, when you have like different assets and you want to iterate, you want to like see, okay, if if you change like that color, if you change that button.
00:35:36
Speaker
um So they usually usually have like manual people and that is also what we get requested so many times. They would like to do that. They would like to experiment, but they don't have the resources internally. So yeah, this ah is, I think like when you look at the different processes and what you can do and that can just like scale things up. um So we we see it already happening. It's already quite advanced, I would say. And yeah, and that helps. I think in performance marketing and just like, because like a lot is about experimentation and ah getting the message right. And and ah yeah, a generative, I think like, definitely helps there.
00:36:15
Speaker
Right. you know I lead a lot of our creative production efforts at CELF and obviously we're always experimenting and want to you know create as much of a personalized ad experience as possible and do a lot of persona testing. and you know As you were talking about this, I was i was just thinking about the the fact that you know we we may develop personas and say, okay, we're going to we know that you know our customer base tends to fall within this age range, this gender, dealing with this issue, we'll create this this thing. um But you you know imagine if it could be personalized to the the user level or to some very specific audience level and ads were automatically created based on the specific things that that person is dealing with, specifically what they look like. right There's some really interesting yeah thought experiments there. Yeah, exactly. Like I agree. So, and, um, um, I think, like, and I don't know if you have already like ah some, some, some documentation or like some, some materials case studies or so to see like, um, uh, the results there, but I said, like, I think it's a definitely when you think about scaling and, and what to do, and and maybe also finding like some interesting insights, what works and why, uh, Like, yeah, so worse engineering it, it's a, it's also already like very helpful and that can help you in all other channels, yeah, in TV and what you're doing, which then closes the doom. So yeah, I think that's, that's good. And then in programmatic, it's maybe comparably not so expensive also to test it out and, and get all those learnings that you can use for your broader strategic campaigns. Yeah.
00:37:41
Speaker
A hundred percent. Yeah. I've, I've always believed in the ethos that you test small and scale, meaning that you're not going to, you know, use messaging and and testing on a $200,000 TV ad before testing it in a bunch of different places to to make sure that, that it works.
Remerge's Mission and Future Focus Areas
00:37:56
Speaker
Last question. So the the first time that we connected, we were kind of talking more about just like, you know, mission and in life and and lifestyle and things like that. And you mentioned that you love what you do. A lot of folks at Remerge love what they do. You just celebrate your 10 years.
00:38:12
Speaker
And you're not stopping. you You don't want to retire anytime soon. You're like, I'm in this, right? You know, as you think about re-emerge over the next ah three to five years, where do you see it going at a high level? Obviously, a lot of the work that's being done on the on the privacy side is going to continue. But what are some things that you're kind of super excited about? And what does that roadmap look like for the foreseeable future?
00:38:32
Speaker
As I said, like I mean, for for for us, because it's just like constantly changing the industry, it always keeps ah us busy and it's very exciting. It's like always a new challenge. And not only like I think from the outside factors, yeah are what Apple is doing, what Google is doing, legislation, competition and things like that. ah it's It's what's happened, but also internally, right? Where we try to grow, where we try to mature, where we have so many people that just like also get get better, get get bigger. And it's ah it's super exciting to just like see that, that this whole organism somehow just like gets, it grows up. Yeah. It's it's it's really exciting. And um and and and for for for us, like we see but we still keep
00:39:15
Speaker
very focused so to our mission and what we want to achieve, right? So we stay focused on app marketers, on app developers to just like drive value with up products, with their tools they're offering to the audience to help them navigate in the marketing ecosystem as good as we can to drive like, yeah, scientifically very proven good value that the money they spent is really well invested and drives value to them. And so yeah, but with that in mind, we just like, um try to get better and better and deliver good results for everyone. um And because it's like constantly changing and you have always to adopt and there will be new things, new ideas. So we try to still be innovative and be creative and um yeah, just like do something good for the industry. So that's why I think we are also everybody's motivated. Everybody has fun. We try to keep it all in a and a good fun environment as you might ah know or heard already and yeah so and I think like that um that's a good good different ingredients to yeah be happy in life
00:40:19
Speaker
ah A hundred percent. Yeah. Yeah. I'll put the 10th anniversary post in the in comments as well. yeah There's some interesting costumes to to check out that post. Yeah, we are known we are known for our like yeah theme parties. Like every anniversary we have a specific ah theme and in Berlin where we mostly do that, it's it's already quite known in the industry and everybody comes and it's like yeah a lot of fun. So amazing stuff. Thank you so much for being on the show.
00:40:49
Speaker
Well, thank you and good luck with everything and speak to you soon.