The End of the Duopoly: What’s Next for App Growth

TRAILER

Neha
What does data chaos look like?

Lorenzo
The algorithm is drunk.

Daria
Marketing managers are oversegmenting.

Lorenzo
A perfect targeting won't save a bad creative.

Episode

Neha
Hi everyone and welcome to today's episode. We're diving into one of the hottest topics in mobile marketing — channel diversification. Joining me are two experts, Lorenzo Rossi from Replug and Dasha Gordeva from Yango Ads. Welcome both. It's great to have you here.

Daria
Thank you, Neha.

Lorenzo
Thank you, Neha. Thanks a lot for having me.

Neha
Why don't you tell me a little bit about yourselves, your background in mobile marketing and beyond.

Daria
I have been working in Yango Ads for two and a half years now. Previously I've been working as a marketing manager for different companies with web products and mobile products. In Yango Ads, I have been a product marketing manager for Yango Ads Campaigns, which includes all kinds of advertising products for web and app campaigns. This year, I transitioned fully to working as a business development manager for Yango App Campaigns and now growing our mobile clients globally.

Lorenzo
On my side, I can be pretty synthetic. I've been in the mobile industry since the beginning. Since I moved to Berlin, I discovered the world of app marketing. The first mobile apps were about customizing your lock screen on your iPhone. And I was working for this company that was producing this type of applications. And within this, let's say, 12-13 years, I've seen the mobile industry changing a lot from CPI oriented campaigns, then of course, all the fraud problems and moving toward the event optimization, the raw optimizations. And of course, right now we see much more attention on full funnel optimization. I remember, when I started working mobile, ASO was not even a thing. And right now it's of course, a part of the strategy. I think I haven't mentioned, but I am the co-founder of Replug. It's a mobile app marketing agency based in Berlin. And yeah, very happy to be here!

Neha
10, 12 years in the industry, Lorenzo, is not a joke. You must have seen everything. Marketers often face what we call the messy middle of user acquisition. And what does data chaos, I really like that phrase, data chaos, what does data chaos really look like for app marketers?

Daria
If we're talking about the picture of several years, we're moving more and more into this walled gardens and privacy laws and all of this siloed data from each platform and combining all that and trying to connect all the dots — that's for me the data chaos. For marketers that's the first part of it from the overall market and app marketing picture right now in 2025, but from another internal side, I think it's also clear KPIs within the companies and teams.

Because what I've seen is that teams in one company can talk about the same metric with different names or vice versa. They say one thing and they mean completely different metrics by it. And we are already in a very fragmented data world. And it's even amplified with this data chaos within the companies where there isn't one single source of truth or one document where it's clear what we are tracking, and what are our main KPIs for the whole business, not just different teams within the business.

Lorenzo
Data chaos, I agree, it's a very cool word to define it. It's just when teams don't know what's working and teams don't know the return on investment on their dollars. This is data chaos. And we have seen that after iOS 14, of course, data chaos is increasing. It's a real problem because it generates inefficiency and there might be several reasons. One of the reasons is the one that I mentioned that teams use different naming. They look at the same data, but with different names for the events, for this type of things. But one of the biggest problems that at Replug we see is that many companies don't know how to set up the MMP. 

So, the tracking part is often not working. The attribution windows are wrong. And they are right to the point where they say, OK, the MMP is something that we don't need when instead it's super important to have a clear source of truth when it comes to data, especially if you want to adopt channel diversification, you want to expand and scale across different channels. So for me, that is when you basically cannot track and you try to guess what's working and what's not working.

Daria
Okay, so we are adopting this term 'data chaos'. They're saying that probably we don't need MMP at all. So that's their sentiment towards it. When in reality, I feel like that's the only solution. Yes, it will take time. Yes, you need to really have resources and knowledge to set it up. But without it, it's just blind leading the blind. It's just like you have no data to base your whole strategy on.

Neha
Oh, I love that phrase, the blind leading the blind. I feel like that's half the industry. Sometimes we're just like, what's going on? No idea what's happening.

Lorenzo
Sometimes it's convenient for someone to be blind because without clear data, with data chaos, people reach goals. So the problem is always at the top. If they are not the right goals, maybe it's also convenient to be blind and to have data chaos. We personally don't like data chaos, but I mean, you do you as we say.

Neha
Or just throwing it in your lap, like do whatever you want. OK, so that kind of comes down to my next question. What are the most common user acquisition channels today and how should marketers approach their targeting strategies?

Lorenzo
We are working with many clients, of course, that are approaching advertising, but also many clients that want to scale. And we work with all of the channels available to promote a mobile app. Of course, we still see a lot of meta and Google being allocated the majority of the budget in the media mix. I believe that when you start advertising an app, but also when you want to scale Meta, it's still one of the best channels because it offers a lot of possibilities in terms of targeting, in terms of testing different creative concepts, and different messages that you want to deliver. And to be completely honest with you, I hear a lot of, yeah, Meta is not working anymore. The targeting is broken. The algorithm is drunk, these kinds of things.

I know it's not that I believe — we have proofs and we have case studies, we have success stories that say Meta is working. We actually released a targeting checklist a couple of months ago. If you fix certain things with Meta, if you follow the structure and most importantly, if you don't give the possibility to Meta to spend a budget however they want you will get results. This is something that I can tell you with quite certainty. Then, of course next, to Meta is Google which is still too big of a platform to be ignored. Especially if you have an Android app because the reach is huge. If the budget is good, the alogorithm can deliver the targeting. Here, with Google win, we all know it's less precise, you can test less creative content whatsoever. But those are part of the old duopoly that has been dominating the advertising space for the past decade, I think. And there are of course, Apple's Search Ads, TikTok, Snapchat, this kind of other channels. Especially TikTok is very big. Now it's working quite well, that needs to be part of the media mix. One thing that it's important to note is that Apple Search Ads nowadays, in almost 2026, are not being used very well. We have seen many cases at Replug where app developers are relying too much on Apple Search Ads because they can track it better. We have discussed before about data blindness and with Apple Search Ads, even after iOS 14, you are able to track like as if it was before iOS 14. And for this reason, many app developers, they say, OK, let's try to scale up all searches. But Apple Search Ads is scalable just until a certain point because it's a search platform. So if people don't search, they don't find you. It's difficult to scale up all searches for a niche app. That's always something that we like to remind our clients, our leads, our partners.

Neha
I have two new phrases today. Data chaos and algorithm is drunk. That is a phrase I have never thought.

Daria
We have still more time. There will be more phrases for us to take away home with us.

Neha
Okay, I'm making a checklist. My new phrases that I enjoy. Algorithm is drunk.

Lorenzo
We like checklists.

Neha
Dasha, do you want to add in your two cents?

Daria
I really agree with all the channels that are mentioned. However, I think that lately, especially again with all the algorithm updates and new markets and new regions coming into play besides the tier one countries for many of the app developers. There is some interest toward other DSPs as well. And here, actually, I wanted to talk about the best targeting practices. And it's going to sound really controversial. I don't think that there would be a thing that it's going to cure all of your campaigns. 

But we see that oftentimes marketing managers are over-segmenting. So, they're making things overcomplicated on these platforms where the algorithms are as advanced as in the Metas and Googles of this world. So, I understand the sentiment. As a marketing manager myself for many years, you really want to find your niche audience. You really want to find the user personas that you spend so much time working on. But if you are putting way too many constraints on the campaigns, that can actually hurt them. So instead of finding your very niche audience and really pointing the algorithm towards the 10 people that you feel like will have 100% conversion rate, instead, it just limits the data pool for the algorithms to work with and to learn on. And at the end, the campaign just doesn't take off. This is especially true for very advanced platforms with advanced algorithms. 

But just to keep in mind that when you are just starting, maybe you should have one very broad campaign to see how it performs with a very large pool of data and whether it will outperform your other segmented campaigns. Again, there is always the question of balance and there's also the question of the middle ground between targeting and reliable algorithms. It's just that in my experience in the past few years, I've seen a lot of marketing managers leaning over to the over segmenting side, which is why I want to highlight it here that maybe you should think about is trying broader campaigns to see how they work.

Lorenzo
It really depends on the type of app you are promoting. If you have a ride hailing up and you want to scale worldwide and you want to test a DSP, of course, don't segment like it doesn't make sense. But if you have an app that is designed for a specific demographic, let's say a dating app and you want to target just females or men or young people, old people. If you have an app that is designed more for women, like a pregnancy app, this type of thing, I understand you might be interested also in husbands maybe, but you need to segment in a way. True. If you segment too much at the end, you're probably going to make the algorithm drunk, as we were saying before. Drunk on segmentation indeed. As you mentioned correctly, I believe it's a matter of balance. It's a matter of channels, the channels you're using, the app you're promoting, and then you can decide if you want to segment or if you want to go more broad.

Daria
I think that in the industry, what I've seen more of, is how to build the personas, how to really understand who your customers are, and then go into platforms and segments. And there are cases where there should be some restraints, as you mentioned, like pregnancy. Although I think husbands should be considered because, you know, there are two people that are having the baby.

Lorenzo
We should debate about that. We should debate.

Daria
There isn't enough content that is talking about how you should also test how the algorithms work by themselves.

Lorenzo
Indeed. Indeed.

Neha
I found a new phrase that I like, Dasha, cure your campaign. It sounds like a sickness. I love it. I've heard fix your campaign, modify your campaign, adjust your campaign. But for the first time I've heard cure your campaign.

Let's talk a little bit about diversification. We've talked about all these major players in the industry. And I think it's critical that we talk about diversification where we don't have to rely on one channel. So why is diversification extremely important right now and how should marketers adapt their creatives when moving away from Meta?

Daria
Recently meta updated its algorithms and it's called Andromeda update. After this update I've seen the divide in the industry. Some people are saying they've seen way better results and some apps are saying their campaigns just tanked. So, imagine you're having all of your user acquisition strategy depending on this one channel, and then there is an update in algorithms and all your campaigns just stop performing. So, there's very little time for you to quickly come up with a new strategy, to find new channels, to have media plans for them, etc.

Again, it feels like a very common knowledge at this point that at all areas of your life, you shouldn't be putting all your eggs in one basket. But that's very true for marketing and user acquisition as well. Again, especially with this very recent and very colorful example with updates in Meta.

Lorenzo
We haven't seen big changes in Meta's performances, to be honest. Actually, with some of the clients who are spending big budgets over the past weeks, over the past months, the performance has been getting better. Now, I don't know if it's because we still rely a lot on targeting. The Andromeda update, in my opinion, it's just a strategy from Meta to push advertisers to move toward their interest, meaning fully automated with A-plus campaigns, advantage-plus audiences, advantage-plus creatives. And trust me, I don't manage campaigns so much anymore. But every time I open the Meta Ads dashboard, I see that there is a new advantage-plus thing. And right now, my 13 years old niece would be able to set up a campaign on Meta. It's so easy. But you lose control. So you lose all the control on targeting. You lose all the control on other things that, in my opinion, are still very important. As long as we will be able to do this type of targeting. And this is related to the Andromeda update. To go back to your initial question Neha, yeah, channel diversification is crucial and critical also if you want to scale. And also, as Daria was mentioning, it's important not to put all the eggs in one basket because we had last year, do you remember when TikTok was banned from the US?

Daria
Yeah, yeah.

Lorenzo
They stopped all the dating campaigns from one day to the other one overnight. So imagine the dating app that was just running on TikTok. And the day after they wake up and everything is paused. So those things are important. Don't put all your eggs in one basket, of course, but also channel diversification. If you want to scale, it is crucial because you cannot scale an app worldwide just using Meta, or just using Google, or just using a DSP, or a single channel. If you want to really scale and again, I'm thinking about a ride a headingiling app like Uber and Bolt. These type of apps that really want to conquer the markets, they are active everywhere. We are now promoting a dating app in the DAH market and we use everything because it makes sense to test everything when you really want to scale. Of course, you start from Meta, you start from Google, TikTok, Snapchat, but then you test DSPs and we are active with several DSPs. We test channels like Reddit, this type of minor channels, but where you can find your niche audience. And in my experience, what we have seen at Replug is that if you want to adopt channel diversification and make it right, you need three conditions. And the first one is the budget.

Neha
That's like the dreaded phrase of everybody, budget.

Lorenzo
So, if you want to do channel diversification with 10K per month, what do you do? How many channels do you want to run? 10K, 20K, one, two maximum. The second condition is the tracking. And to go back to data chaos, you need to have your tracking in place. The attribution needs to be aligned across the channels and everything needs to work. And the third rule, and then I shut up, but it's the most important. It's the creative strategy that needs to be adapted. One size fits all strategy doesn't work.

Daria
 Yeah, Lorenzo, you mentioned your like checklists. And I think that if listeners take away just one main thing from this episode, it's this checklist of yours.

I completely agree with you. These three things, the budget, the tracking and the creatives, is where most people are making mistakes, not mistakes or maybe they are mistakes.

Lorenzo
Let's be harsh, bad mistakes.

Daria
Let's be harsh, yeah, mistakes. When they are starting a new channel, when they're expanding, because one, they want to see if the platform is quote unquote working with very small budgets. And of course it won't because there's not enough fuel to start the car.

Lorenzo
The algorithm doesn't get enough alcohol to get drunk.

Daria
Exactly. Yeah, there's not enough booze. So they're saying it's bad results. So the experiment didn't work. Then again with the tracking. So, it wasn't set up correctly. They don't want to spend time. It's just an experiment. We don't want to put our resources into setting everything up. Then, you know, we don't see any conversion. It's not working. And the third part again is like, oh, we have creatives from TikTok and we're going to just put it into this DSP because we don't want to, you know, use resources on creating new videos or banners, even though the context and the audience that's going to see these creatives is very different from like TikTok's context and audience. But again, not many resources, it's just an experiment. And again, it doesn't work. So it can be an experiment, but these three pillars are very important to put into place to have the real results of this experiment and not just half-ass it and then say that the platform doesn't work.

Neha
I wanted to ask you a little bit about the biggest mistakes in diversification, but do you want to add anything else? Maybe more booze-related phrases?

Daria
Oh, booze-related phrases. Each algorithm has its own time to get drunk. People have different tolerance for alcohol and some people get drunk after one pint of beer and others can drink 10 cocktails. So it's the same for platforms. And when you are starting a new channel, don't expect the same behavior as you have in others, especially in the big ones. So if your Meta campaigns stopped a slotting period after one day, it doesn't mean that now all DSPs will do the same and vice versa. So please, if you have agency like Replug or you have ad network partners, talk to them about what to expect from each platform. I don't just assume they're all the same.

Yeah, indeed. That's a very good point.

Lorenzo
Every channel has its own characteristic and also the KPIs need to be different. I'll make you an example. And this is something that not so many advertisers talk about because it's not a very common channel. But we have been pushing a lot with OEM placements. And maybe not everyone listening to the podcast is familiar with this. Basically, you can place ads on Android devices like Xiaomi, Samsung, Oppo, this type of thing. The performances are very different from any other channels. And you need to be aware that you will need to wait a bit more time for the cohort to develop to see the differences.

And this is valid for all the other channels. Because, as you mentioned, if you start a channel and you run it for one week and you spend 300 euros and then you say that it's not working. Well, I'm sorry, but this is a stupid mistake — you based your assumption on something that it's not correct. And we need to be harsh. I think I invented a quote and this is for you Neha.

Neha
Thank you. I'm honored!

Lorenzo
I used to say that a very good creative can work very well with a bad targeting, but a perfect targeting won't save a bad creative. And this is valid for all the channels. Yeah, I know. I know you can print it.

Neha
Yes, I'm going to. You've got to copyright that, sir.

Daria
I was just going to say, but make sure to use copyright.

Lorenzo
So yeah, the creative part is very important for all the channels and you need to have it in place because otherwise it doesn't make sense to diversify. Because again, you measure the results with not the right setting.

Neha
So, we've talked about creatives and the do's and don'ts. Let's talk a little bit about tracking and attribution. What role does MMP tracking play in diversification and how do you avoid getting lost in attribution complexity? Because I know people just go haywire when we hear the word attribution.

Daria
I'm going to start from the end of your question, because I don't get the sentiment of getting haywire when we talk about attribution.Marketers actually should be curious about how the different channels are performing and how it all fits into the big picture. So, maybe it's my ideal world, but when I'm starting a new channel, when I'm starting a new campaign, it's in my interest to get everything tracked properly and attributed properly so that I know how the test went, how this fit into my whole strategy and did it help my business at the end. We talked about it in the beginning, how tracking is one of the pillars for not only diversification, but in general for UA and marketing. For me, it's like navigation. So if you are going somewhere, you want to have a map, you want to know where you are and how long till your endpoint. And tracking does exactly that. Tracking, web analytics, MMPs. So going into user acquisition without MMP, I just don't see the reason for it. If you don't know how your budgets are performing and if this brings real business value for you, then it's just playing the slot machines.

Neha
Okay. So we've moved from alcohol to now gambling. I'm just kidding.

Daria
The metaphors are going crazy. We're recording this on a Friday. So maybe that's that.

Neha
Yeah, that's true.

Lorenzo
I like Daria that you're referring to the MMP as your Google maps, basically. It's the navigation for traveling. I mean, I can only agree with you. If you check my LinkedIn, you see how many beefs I had with people that were saying that MMPs are not useful anymore.

Daria
That's crazy for me. That's just crazy. Sorry, I couldn't pass it on.

Lorenzo
Yeah, but it's crazy for me too. I am a person. I like to see other people point of view. I understand if you only spend again, small budget, only on on meta maybe you don't need an MMP maybe you can go with the meta SDK but since we are talking about channel diversification and since we want our clients and our partners and our apps to scale we are gonna spend a lot of money and if you spend a lot of money on different channels on different platforms android ios how the hell are you gonna measure your results if you don't have an MMP in place? There is a solution. You build a tracking system in-house and it's going to take you 10 years and it's going to be fantastic. But still, you need a tracking system and there is no other way. You need to have all the attributions in place, meaning sometimes we see MMPs that are set up with wrong attribution windows that are still tracking 28 days post-click for Google with seven days post view and this type of things. And it's messy. So everything needs to work. The scan setup, it's so important. And people think that, okay, we set up the scan once and then it's done. But they don't realize that we need to keep testing. We need to test different events, different setups, different windows. Now with Scan4 you have so many settings that you can tweak, find conversion course. These type of possibilities, they need to be used and they are important to track your results and to see where you are going. But let me add one last thing. It really depends on what are the goals, because of course, Daria, Lorenzo, Neha are very interested in seeing the real results. But sometimes people like to see better results on Google. Funny example. Last year I was talking to Alit and they told me we have an MMP and we also use Firebase and we use the MMP for tracking meta and we use Firebase for tracking Google because it shows better results.

Neha
Oh, dear God!

Lorenzo
Yes. What the heck, man? Seriously?

Neha
Oh, my God, sir.

Lorenzo
But I understand the marketing manager. They have their yearly goal of achieving fantastic results on Google and then say, okay, let's get the numbers from Firebase because they look better. And by the way, Firebase is not an MMP.

Daria
I think talking about setting correct KPIs and really being honest about what's going on with your campaigns, that's a whole different episode. We should note that for the future. Another controversial one.

Lorenzo
On a Friday night, maybe we can talk about it.

Daria
Yeah, there's going to be even more metaphors about alcohol and slot machines.

Neha
Lorenzo, I have one last question for you. If you had to recommend one overlooked channel for user acquisition in 2025 or well, it's only two months left for 2026, which one would it be and why?

Lorenzo
I mean, do you want the cool answer or you want the real answer?

Neha
Oh, are we going diplomatic versus non diplomatic now?

Daria
I think with how this episode was going, I would prefer the real one.

Lorenzo
The real one, of course. I give you also the cool one. I mean, right now everyone is talking about CTV, connected TV, this type of thing is going to be fantastic and great. It's going to give great results. Probably yes, but it depends on your budgets and your app at the end of the day. Channels that we see that are performing very well recently are the OEM. Again, we have been testing them. If you set your goals and if you understand that they have different KPIs, they can bring your results. But what we are seeing as well is some specific DSPs that are working very well. So the algorithms are less and less drunk. Of course, as everything, you need budget to invest in the learning phase, because you guys know that if you want to test DSP, you will have a learning phase that is probably six times longer than Meta or Google. So, these type of things are important to clarify. So those are my answer. OEM, DSPs. And I mean, I love TikTok as a channel, but that's not a new channel. So if you are at scale and you don't run TikTok, probably you should fire your marketing manager.

Daria
Ouch!

Neha
Okay, so we really went non-diplomatic here.

Lorenzo
I'm not. I'm not diplomatic. I have a lot of enemies in the industry.

Neha
This was an insightful conversation. You know, we spoke about our vices, alcohol and gambling and all these fun phrases. Thank you for tuning in. And don't forget to subscribe for more conversations on mobile growth.

Lorenzo
Thank you. Bye bye.