Session: Finding Signal In The Noise: What 2025 Gaming Trends Reveal About Measurement - C. Agaoglu.

[music] [music] So actually there's another session right now by adjust uh finding a signal in the noise. what 2025 gaming trends reveal about measurements.
Everyone, thank you for being here. I'm John Su. I'm leading growth uh across Europe at adjust. And today I'm really excited to talk about uh finding signal in the noise. That's phrase uh that has been used a lot but in gaming uh in general we can say it's not just a metaphor but it's our daily reality. Um and today teams they have more data than ever but data is fragmented. um

gaming companies working with many networks which means more more dashboards more reports um they are dealing with on a daily basis and the problem is not about tracking the performance actually uh but it's all about finding the right signals to predict growth uh and also uh that will give you uh an idea when you are scaling up which signals they will get louder so today uh I will going to unpack this topic Okay. And let's get started. Uh, it's

so let's look at our gaming report um in 2025 and we will talk about the confidence gap. So measurement has quietly become uh more complex than ever. Right nowadays we are working with this aggregated uh privacy centric data, user device level data and modeled outcome which is creating multi- layer uh uh data structure. Um and that often reduces the confidence instead of increasing it. And when we look at install and session growth uh in 2024, we have seen 4% year-over-year growth in installs. But when it comes to sessions, it slightly went down. Which means acquiring new users, more users, that doesn't necessarily mean more engagements. That also tells us a story about the campaign types, right? Nowadays, we are shifting towards value based campaigns. Which means that when we are releasing that new report around

March, stay tuned please. So we might seeing a different picture and let's look at the number of uh partners uh gaming companies working with uh I mean ad networks and marketing channels in 2023 median partners per app was 5.7 and in 2024 that increased to 6.2 too. Uh which exactly highlights that uh the companies they are diversifying their marketing mix and when we look at hyper casual hybrid casual titles they're actually uh hitting over 10 12 different partners right. Uh what does that mean? Uh that means more touch points more overlaps and that's the main reason why we started having deeper conversation about attribution settings reattribution settings how we can customize it how multi-touch attribution come into play.

uh and we can evaluate the campaign performance in a better way. Right? What matters the most? Analyzing the marketing mix and then creating u creating a strategy marketing strategy to to give the right credit for the specific marketing channels. Uh that's what we are focusing on nowadays and as I mentioned growth is uh moving away from volume to valuebased campaigns. So the the most important thing is efficiency per user. growth is just an efficiency game right now and we all hear about CPIs are increasing. I wanted to put in context um in 2024 we have seen 17% year-over-year CPI inflation uh in North America and in the US uh and that means yeah costs are increasing but at the same time when the winning gaming companies they are not just focusing on higher volume of users they can beat higher because they are generating more value out of per user

right uh so that's shift shifting our strategy uh right now because it's a very competitive market. We are trying to find the better way to extract more value per user and everything we have discussed right now that's moving towards to this new growth equation margin to bit. So basically margin to bit that refers to confidence predictive LCV multiplied by uh payback payback speed. So why that matters because that simply decides who can scale and who cannot and LCV confidence that's how sure we are about future value and how accurately we can predict that value and payback speed how fast we can recover UA spent right uh let me give you an example like average median LCV by day 30 it's around 63 uh 63 cents and if we are spending three $2

or more to acquire a user which will require uh a very uh ex exceptionally high early monetization right that's why UA teams they are emphasizing on retention and early spend patterns uh to improve that payback speed and let's talk about uh how we can build that uh confidence right how we can uh create that early signals which will predict LTV because not All early metrics they are equal right uh there are some metrics which will be still valuable which will be still important for you to keep track of them but uh that doesn't mean they will also share this insights for you to predict the future value. So what we can talk about here is the first signal I want to talk about here is early monetization pet entry. So high-end TV players don't necessarily spend early uh but they

recognize value early which means um uh this higher TV players maybe they are going to spend but maybe not in the early stages because it's just a matter of timing right if you can understand and measure the intent behind it you are just one step ahead of the game and retention shape not only just retention rate uh but retention shape is becoming important because if you have a stable retention floor which means the users are coming back to your app regularly. They understand the game dynamics and the most importantly it is giving your monetization system enough time to work. Uh let's look at this examples like there are two games a very similar day one uh day one retention rate 32% and 38%. And the first game that stabilizes at 8 to uh 10% by day seven and the second one that falls to 2% uh which they will uh you will see completely different results right that's why we are focusing on this uh healthy stable progress and

the third signal is progression velocity. So early progression uh is how players experience value you know not how long players play but also how far they get and how fast they get because LTV is coming from that habit formation right uh and long signal sessions that can be misleading maybe just one session 40 minutes length but then you are losing that user and the other one you are seeing this healthy growth over three days and they will come back to your app because they understand the game dynamics. and probably they will also engage with your monetization system. That's why it's important. And finally, the last signal is ad tolerance signal. This is not about how much ad revenue you are generating uh for the early first two days or uh within the first 24 hours, right? This is more of like understanding if these users are going to tolerate that, right? So we are uh measuring here how much players accept what is the tolerance rate and

this is the proxy for our long-term monetization strategy. So that doesn't mean that we are not going to work with this valuable important metrics. Of course, we are going to keep this existing dashboards, right? CPI, CTR, IPM, day one, day seven retention, ARPO, ROS, those are still very valuable matrix. But what we are introducing here is adding this uh early signal layer, expanding that inep events u in inapp event set and that will give you more signals in terms of that specific intent. For instance, monetization intent. You can set up a specific event for overview reach level X by day two that can give you good indicators if these users are going to be high LTV users or not or second session same day or rewarded at accepted that will show you at tolerance signal.

So if you're asking me since I'm making things a little bit more complex with this additional set of inapp events right if we are going to track this activities manually in 2026 of course not we will automate this marketing flows and we will help you to find out uh that specific insights uh regarding uh predictions uh with pulse alert notifications. So that's the feature that's a growth solution. It's uh uh we we we have introduced it has been uh so long it's in the market but we continue improving the product. First I will give you a quick information about what we have built and how we are improving this. So basically this alert notification uh that helps you surface any specific abnormalities uh unexpected behaviors in terms of ED spans install inapp events CTR IPM you decide or you can also customize this alert logic if you if you are looking for any app behavior signals any red flags any

performance metrics you can also to do so and then we will be sending to you either email notification or select or teams notification this is more of like having access to write information uh instantly and you can start act on it. So there are a couple of examples here. Uh we have introduced this group logic. So instead of creating like two different logic now we can combine them together. The first example is more of like campaign performance optimization. If CPI stays below uh target for three days scaling spend. If CPI rises above targets you can either pause it or maybe you can just reduce the spend right and the second one is uh for creative fatigue because creative layer that's the only layer which is 100% in human hands right now right let's admit that optimization at networks they have very advanced algorithm so we are not doing much but in terms of creative set if we are receiving right signals at the right time we can understand better like how

often we should change the creative set uh how many creative sets we should create on a weekly basis or on a monthly basis. So um you can create an alert logic for if performance drops after x impressions reducing the budget. If new creatives outperform baseline you can boost the spend. So pass alert notification this is uh more of like creating that specific threshold and alert logics and getting notified instantly but that will require analyzing historical data right we are also making everything easier for you instead of uh having this like manual analysis and now you can use this growth copilot which is our AI assistant so this is like plugandplay AI layer built into datascape it's all uh integrated with aggregated data so you can question or prompt anything you want. It might be a specific metric you are looking you might be uncovering um different patterns or maybe you are running going

to run an incrementality test with historical data and you are just asking the question what type of campaigns or specific patterns you can start running an incrementality test. So sky is the limit here and for the teams they might go further. We also offer adjust MCP and that allows you uh adjust measurement data to be securely and seamlessly integrated into any agentbased uh architecture. So uh with this growth copilot not only for identifying and uncovering that specific signals or metrics or patterns but you can also create reports and now growth copilot is also integrated to pulse. So what we are doing here is let's say there is a specific uh executive summary for the last week's data right you are sending it to upper management and you would like to prompt this every week you can just define the prompt once uh this is

basically what you want to know and then you will choose when and how often this should be run and copilot uh will prepare the analysis in the background and this readytouse report will be delivered on time. Uh this is just an example for this weekly executive summary but we talked about I want to go back to this creative fatic topic because this is a complex topic because every company is analyzing this in a different way it might be just a simple uh analysis in terms of impressions if it's going down by 20% if you're seeing a decline uh that says that you need to update the creatives it might be CTR it might be CPI you decide but of course this is not going to be a static threshold right let's Say you are working with this alert logics, you are creating that specific alert logic based on this uh historical data. But when you are creating new campaigns, running new campaigns and with this marketing mix, the static data will change and you can also automate this with the scheduled

report. uh you can run you can decide of course first how you are analyzing this and then you can have the scheduled prompt and you can run this every week every two weeks and that will help you to update the threshold on the alert notification side. So this is such an exciting topic and that's the future of measurement. So our plan is connecting all this growth solutions with growth copilot. Imagine incrementality uh integrated with growth copilot whenever you are running a new campaign uh with a new ad network or in the new go and growth copilots will be informed and we will generate that incrementality test automatically without you going to incrementality reports and creating it manually. So it's all about leveraging AI uh for operational purposes. So if you're interested in we will be happy to have uh to continue having this conversation. We have a boot in this floor so please let me know. Thank you very much. [applause]

Uh I think we have a little bit of time for questions. Does anyone have any questions?
Thank you Canu for this lovely speech. I would love to know I'm focused on UI creatives and now with AI we do tons of them and I was wondering what type of reliable signal would you look at in order to say this is a winner ad. So it all depends like how you are monetizing it. uh first uh you know there will be like different signals but on our side the most important thing is collecting all this uh specific like revenue activities doesn't matter if it's like AP driven game or ad revenue driven game of course like roast will be very important here right uh but at the same time if you're focusing on like targeted ROS campaigns so there will be different set of KPIs if you're just focusing on volume like CPIs if it's a eddriven game and you're trying to just you know recover UA spent like in the

next 7 days or 14 days there will be a different set of KPIs like what I can suggest here is sometimes like we are tracking those activities like metrics separately but at the same time actually you can just merge them together you know so if you are giving credit for like different like uh maybe Ross retention and everything so we can just create that combined metric together and you can give enough credit for two separate activities It is that might be a better way to evaluate like what is the winner creative here right
uh yeah we have one final quick question
I can do um couple of [clears throat] quick questions one AB um firstly uh on the optimization piece does it have right API capability so is it if you applying these rules based optimizations
hm M
can it apply it to the relevant channel

whether it's meta snap
yeah of course of course alert notification it's just notifying you in an email or select notification way but that's also integrated with campaign automation so if you would like to apply the same alert logic to change bid budget or campaign status so you can also create a campaign automation rule
okay and tied to that um in terms of your correlation essentially
we are working with all this uh major ad networks we are integrated with them so you don't have to go change and budget. One thing I can highlight here is if you're able to change bit budgets or campaign status on at network dashboard, you can do this things on adjust dashboard as well. But if network is not allowing you to do any changes on subsource level, then we are also not allowed to make those changes. Basically, it's just an API connection. We are sending this information to add network side. So you don't have to open like 10 different tabs to update it. So you can just do it on the adjust dashboard. And the last quick question is correl the correlation analysis. So the additional signals are going to be

different for different advertisers. The thing the intermittent uh parts of the journey that link to a purchase that correlated to a purchase the proxy event. Do you does the analysis also come up with those suggestions or should the advertiser define that and then they just use your tool to set that up?
Yeah, even for the same genres or very similar apps um there will be different dynamics because we don't know like what is their marketing mix, how they are setting up their campaigns. Sometimes they are just focusing on CPM or CPI based campaigns and the other ones they are focusing on day seven day 21 targeted ROS campaigns. Uh so they completely like operate uh differently. I think the most important thing is just analyzing like historical data to understand better and also your objectives right like not everyone have the same financial security in terms of recovering UA spent uh that also uh quite important.
Brilliant. Can we give Kantu another round of applause, please?
Thank you very much. [applause and cheering]