State of the Brutally Honest UA Q2 by Matej Lancaric

And we have one quote from the actual survey.

These guys are doing 7 million a year across nine channels, but still Facebook and Google doing 80% of the work, which is not really what you want to see.

Sometimes I see uploading going anywhere from 60 to 80% of the budget.

Again, not ideal scenario.

But

you should try to diversify and try to have very even budget split.

So if one UHNL dies, then you have the replacement.

For example,

I just had to post Facebook on one game because on Android, it was absolutely terrible.

It's 4 a.m.

and we're rolling the dice.

Mate drops knowledge made of gold and ice.

Felix with ads making those coins rise.

Jackup designs, worlds chasing the sky.

We're the two and a half gamers, the midnight crew.

Talking UA adverts and game design, too.

Matei, Steele, Shaku, bringing the insight.

We're rocking those vibes till the early daylight.

But KUA, master, eyes on the prize.

Tracking data through the cyberspace skies.

Felix stacks colors like a wizard in disguise.

Jackets crafting realms, left us to the highs.

Two and a half gamers talking smack.

Slow hockey sick, got your back.

Ads are beautiful, they like the way.

Click it fast, don't delay.

Uh-huh.

Uh-huh.

Uh-huh.

Uh-huh.

Uh-huh.

Hello, hello.

Welcome to another Seoul episode.

Today I'm gonna talk about the state of the brutally honest UA.

That's how I named my,

I would say, recent article about

the UA.

I ran um survey on the 20,000 subscribers that I have on my brutally honest newsletter, and

I have some interesting findings.

Let's call it that way.

But before we start, this episode is actually brought to you by our dear friends

Vibe, the most intuitive streaming TV ad platform, which is designed to maximize ROAS for apps and games of all sizes.

It's actually that simple, it's totally self-served, and yes, it actually drives real business results.

Join 5,000 brands that are already using it.

It's maybe it's yours your turn.

You can check it out at the vibe.co.

It's

go there and start your CTV campaigns now.

Let's get back to the episodes.

I'm talking about the state of UA,

and I actually have a presentation for you.

So, yeah, you can see it right here.

Everybody that's listening to this on Spotify, go to YouTube or Spotify video.

So, I call it how to win in 2025.

I have numbers from a few hundred UA managers from that survey that I mentioned at the very beginning.

It's interesting stuff.

I

have this on my substack as paid article, so you should go there and check all the stats that I have there.

This is going to be a teaser.

Let's go.

How to win.

This is me, Matte Lanceric, 11 years in the gaming industry.

Been around, I mean, we're not getting any younger.

Work with around 52 games, soft launched, global launched, a lot of stuff, small, mid-sized, big games, doesn't really matter.

I also have a creative team of 10 motion designers.

People always tell me I should advertise this a little bit more.

So, here we go.

If you need creatives, let me know.

I do also run a playable ads team as well, and have playablemaker.com, which is the playable ads exporter from video.

So, you should definitely check that out as well.

Okay, let's go to the actual agenda.

So, I'm going to talk about the CPI ROS benchmarks that I gathered from the previous work

channels, what's working, what's not working in the actual creative.

We have some benchmarks,

how to test different creatives, how to use AI, and what skills you and your UA team should have.

So, buckle up, it's gonna be a lot of fun.

It's also

created by AI, as a lot of different

images in this video as well.

UA actually has changed no shit sharelog, right?

In the survey, there's one

company that says, or said that one person managing 40 games.

I don't think that's possible that much,

but it definitely changed the whole UA.

When I joined the gaming industry in 2020, 2013, actually,

we were running hundreds of campaigns on a daily basis.

These days, it's way more chilled.

I mean, it's 80% of this is automated.

You don't have to run the micro-targeting or whatever we called it back in the days.

You run a few big campaigns on all the channels, and that's it.

So you have then creative testing and the creative production.

But it's not,

again, like such a big volume of the campaigns that we had to run in the past.

And these 40 games.

at the same time

it's quite a lot so i hope people just don't tell the truth all the time.

Anyway, so what's actually working right now?

And for me, and I've been saying this for some time, the uploading, and

we got some interesting chatter on LinkedIn about our uploading episode.

I still believe Aplavin is the number one channel on iOS, although

Facebook and TikTok are really working well, especially on the AEM, which is aggregated even measurement on Facebook and iOS.

Then TikTok is also very strong with their own ADC, which is Advent-dedicated campaign.

I think I don't really follow all of these

acronyms anymore, but it's the same thing as TikTok and Facebook.

And now Google came up with ICM, which is whatever

conversion measurement.

But it's all the same thing.

It's modeled conversion on iOS and kind of works, especially on Facebook and TikTok.

So these two.

But then uploading, it's kind of moving and then pushing everybody from day seven to day 28 optimization, which makes a lot of sense.

And if you listen to the episode that we

recorded, Alexis, he also explained why

it's a good move and how you can get incremental

kind of incrementally better quality users, which again makes sense.

More room for data collection and also kind of bidding higher, getting more quality users.

And also, what works now for me, besides all the UI channels, and it's the AI-generated creatives, which we talked about a million times on the creative trends episode.

And I'm always talking about.

But besides all of this iOS stuff, Google is still the top channel on Android, and there is a lot of movement on the rewarded UA channels as well.

Right?

So, who is the new kid on the block?

Yes, you guessed it, right?

Is the rewarded UA kind of all the channels?

We were talking about this a few years ago, and we were talking to Misplay because Misplay are friends of the podcast as well.

They were the only player.

Now, it's like millions of them, and you wake up, it's either a new

model on the AI field or a new rewarded UA network, which is fine.

But eventually, it's

gonna get

a little bit

different on the rewarded UA play,

where it's gonna get consolidated, obviously.

Anyway, but let's go back to the actual UA channels.

So it is working, but

there's always a but

it's kind of limited scale, depends on what kind of channels you are working with.

This was the case where you kind of

had to combine a lot of smaller rewarded UI channels to kind of reach a certain scale.

These days, you have kind of media, you have tier ads, you have misplay, obviously,

and Adjo and all the some others, which are kind of getting bigger and bigger.

So the scale becomes less of a problem.

But there is fraud, which is a huge problem and we're actually talking about the fraud problem with with zeno from tear ads so if you haven't checked that episode you should definitely go there and uh and check out like what you need to do in order to kind of limit the fraud it's few

very easy store not easy well for me everything is easy which is uh i'm the ua person but it's like server-to-server integration and a lot of other stuff that Zeno mentioned that you should definitely check out.

Also, iOS

was quite a big problem.

Now it's changing, thank God.

So we can run a lot of iOS UA on the rewarded channels as well.

But then the biggest problem was

players get rewards and they churn immediately, which then fucks up your LTV predictions.

And I've seen a lot of companies struggling with this because of all the prediction models and

seeing how much they can earn on first seven days and the ROAS is such a great

you can see such a great number on the ROAS side for all these UA channels and campaigns on the reward the UA but then it flattens out so I've yeah I've seen this in some of the companies I worked with before.

When I joined and tried to do the UA review, I saw

a lot of money wasted on this because then the LTV plateaued after day 30, let's say, and then the LTV prediction didn't really work.

Easy solution are targets for rewarded UA channels.

But then also coming back to the episode with Zeno,

he mentioned one interesting thing, which is the long-term player progression.

And which basically says, and comes back to the point that I mentioned before,

the rewarded UA channel needs to have paid capabilities to keep the player engaged even after day seven, ideally after day 30,

ideally after 60, 90 days.

Benjamin, as a UI channel, is actually doing quite an interesting job, so you should definitely check it out.

Because then you can also have multiple different offers.

So you start with playtime or a few step offers, then you get time-limited offers, and then you try to still kind of work with the player within the campaign.

So,

this is the kind of way to go in the future of the old rewarded UA channels we're gonna see in the next couple of weeks, months this year.

Definitely, rewarded UA is not going anywhere.

So,

in terms of the UA budget and the mix, these are the survey data that they gathered.

So obviously what I mentioned, it's very similar to what others are doing.

So on the Android, Google ads, uploading, and rewarded UA.

Interestingly, uploading for me doesn't really work that well on Android.

If I need to compare it with, let's say, TikTok, which always sucked on Android, how with the Smart Plus campaigns,

going really well.

Anyway, iOS, uploading, as I mentioned, Facebook, TikTok, and Apple search ads.

I mean, good luck with Apple search ads.

But again, we have one quote from the actual survey.

These guys are doing 7 million a year across nine channels, but still Facebook and Google doing 80% of the work, which is not really what you want to see.

Sometimes I see uploading going anywhere from 60 to 80% of the budget.

Again, not ideal scenario.

scenario but

you should try to diversify and try to have very even budget split so if one UHN dies then you have the replacement for example I just have I just had to post Facebook on one one game because on Android it was absolutely terrible and I see that across different games at the moment But since I ran Applevin, TikTok, Moloco, and some rewarded UHNs, it was fine because I had

redistributed the budget to other

channels and

zero problems.

Okay, so some benchmarks.

These are

some amateur mobile gaming benchmarks that I've seen before.

You can check it out.

I'm not going to talk about all of these, but we have strategy as well.

which are some interesting numbers and mid-core gaming UA benchmarks.

All of these are based on the real data that I was able to

gather, but also we have a PVX

Lambda, which is pvxpartners.com/slash lambda.

So if you want to get real benchmarks and

real kind of insights, you plug yourself in with an MMP and you can get some benchmarks and some data points.

easily and i mean these are our dear friends from pvx we've seen the numbers we recorded episode with DJ.

And ROS Benchmarks, one of the most popular videos.

So you go there, check out the benchmarks, or you plug yourself in to PVX and get the benchmarks by yourself.

Quite a lot of games and numbers they have.

So definitely useful.

Creative trends.

And we are talking about the creative trends on monthly basis.

So I've kind of gathered the interesting stuff that is working.

And you might have seen it anyway anyway on a monthly basis.

But

we have content,

creator generated content, humor.

We've seen the hyper casual gates almost

in every game, which is kind of going for Forex.

Labels working quite well when you were trying to diversify outside of Facebook and Google.

AI videos all over the place with altered gameplay and misleading ads, obviously still kind of dominating a few categories.

But

as we've talked about ugly ads, and we talked about this in the Bending Spoon

episode, that works crazy well because people don't really know

they're watching ad, but they are.

So, very native-driven creatives work well.

Near that experience and the Save Me, Save the King with Freezing Families.

All of this we are checking on a monthly basis with different games.

And surprise, surprise, there are always these

concepts that are performing well and they're not going anywhere.

But then, I mean, talking about this since January, I think, maybe even the end of 2024.

Creatives and the volume of creatives is very, very, very important.

And we have some data from App Slayer.

Higher the spend, more creatives you need.

And

some of the companies are spending millions of seven millions per quarter, which is quite a lot of money.

They are producing more than 2,000 creatives, which is also like resizes, localization, but still you need a lot of creatives and you need to do it on a weekly basis and quite a lot of them.

Even if you do this, I've seen all of the problems or not all of the the problems, one big problem all the time.

And I've been having conversations with so many companies, and it's always the same question: okay, so how do we beat the top-performing creative?

We have one creative which is performing well.

I mean, it was the creative winner and still is.

We can't beat this creative at all.

So, even the upslare data kind of showed that 2% of the creatives still generate like more than 50% of spend, which sucks.

Sucks balls, really.

The only how you can actually combat this is with the loads of creatives.

The more creatives you have, more shoots on goal, higher chance of succeeding, which means refreshing, refreshing, refreshing the creatives every week, multiple times a week.

It's easier said than done, I know, but with the AI,

with so many vendors out there,

and your own creative team, you have suddenly have so many different creative concepts that

you can go and upload them into

your campaigns.

And it doesn't really need to be only a new concept.

You can, and you should iterate

and you can iterate easily, right?

You can refresh hooks, you can

you can iterate on the length, you can iterate on the speed, you can go about pace of the creative, and just, you know, test new angles, right?

So when you're talking about the hooks, and I'm pretty sure you know what I'm gonna, what I'm gonna show you now,

our good friends from Century Games, which are the kings of the hooks on the one left side, it's Frozen City, and I've seen and we've seen this creative million times, also with White Out Survival.

Also,

also with the Kingshot.

How cool is this hook?

It's really cool, and the best thing is it's actually made by AI, I would say.

Yep.

And it really goes directly into the gameplay right away.

And it suits so well.

They, again, as I said, like they are the masters of the

of the of hooks for sure.

But we have we have more.

So we

I was kind of doing some of the research for for different hooks, and there was this slingshot hook that suddenly I saw so many different games actually using this.

So

Angry Birds, Evoni Royal Match, and it's not only Survivor,

even Survival, maybe, but not White House Survival, but so many other games they're still using this.

And you can see it across Social Casino, Forex, Match Free, doesn't matter.

And we're going to see this more often.

So if you find a good hook, just try to think about how you can actually transform this and translate it to your game.

It doesn't really, it's not rocket science, I would say.

It should be quite easy.

And then, if you think about the iteration of your creatives, hook, that's the first step, right?

Then

you can use the real kind of live scene or sequence from whatever videos you can find, or you can build it with AI these days, right?

And it's working so well.

Even the real-life footage, as I said, AI tools, you can use AI UGC kind of stuff.

You can use UGC.

So you can slap a person on top of the video that you have.

You can go and create a selfie AI video with Google VO3.

You can kind of use the same concept and start with a street interview.

You have different elements, sounds, captions, music from the TikTok and use obviously the video on TikTok.

Or you kind of have one video, you cut it into pieces, or you have three different videos, you stitch it together and suddenly you have a 90 second video.

Doesn't matter.

The long form videos work really well on Facebook, work really well on Google, work really well on Apple.

I mean, well, it's kind of 60 seconds or 59 seconds, but still, suddenly from one for two videos, you can have one long video, easy as that.

I know App Quantum, for example, they test or use the one-minute long video, and it's kind of compounds from two 30-second videos.

The second 30-second video, it's the myth of the creative, which is the best performing video.

And the first 30 seconds is kind of like a hook.

So they're kind of testing that.

And suddenly, you have a combination, and then you can have so many different variations with this best performing video, and with the first 30 seconds as a different video.

So, just think about it: how you can easily iterate, just have multiple combinations from different videos.

For example, I have this kind of

two videos.

Yes,

35-second second video

with one

sound and one

interesting music,

slightly different voiceover, and just the same scenes from AI.

Surprise.

I think the first one,

day seven, was 37%, the second one, which is basically the same video, just different pieces and slightly different voiceover and music, 58% day 7 ROAS.

Quite a big difference.

So, with an easy iteration, you can get so much, so much better numbers.

Then,

what I mentioned earlier, the save me and

saving whoever character from the game, it kind of always ends with a fail.

And I'm pretty sure you've seen this video, which I even we are kind of talking about it on each of the trends that

the saving the king with the royal match.

I mean,

they have, I guess, like 10, 10,000 different iterations.

Again, this is one-minute-long video, and they are using five-minute-long video for a long time.

And

again,

then it goes to the best-performing video here.

It's almost the same kind of practices that Quantum guys are doing.

I think Aralmage is doing as well.

So, again, can you do it better?

Of course, you can.

Of course, you can.

And it kind of goes to

the next kind of point, what is working well, which is the

near-death experience.

It's always the snake or something

which is very, very much trying to kill you, and you need to defend yourself.

And the anxiety and like the near-death experience is something that evokes really strong emotions.

And it works so well with all the kind of target audience.

And we kind of tried

this on all the different games, and it really

works because all the Save the King, now almost all the match free games are using Save someone from drowning, from Snake, from Lava, from whatever danger.

So,

saving someone or near-the-experience, you should definitely try that out with your game, especially if you have a casual puzzle game.

And I mentioned also the creator-generated content.

I'm trying to stay away from user-generated content

because it's slightly different time now when, and I mentioned this, and I will mention this a million times.

It's not about, oh my god, this is the best game ever, you should download it.

You need to still catch the attention,

have a hook, and be different.

Try to really grab the attention and talk about the game and

be very straightforward.

Sometimes, even like this lady over here, even then, kind of showcase the gameplay.

But here, you still have the first, I don't know, like 10, 15 seconds where you see, oh, it's a, you know, it's a brand new game.

No ads, okay, interesting, whatever.

Oh, no Wi-Fi needed,

but she's doing it in a very interesting way, and I definitely grabbed the attention of everybody.

It was one of the best performing

videos,

and

who is this?

It's me, of course, because I'm always trying to experiment, obviously, with a lot of different things.

And I try to understand how

it works.

When I'm talking about the creator-generated content, then obviously I want to try.

And I'm also trying to do ugly ads, and I'm also trying to do different things.

Because when I understand how it works,

it always makes things better.

Also, here with recording

a video with my friend, dear friend John Wright, and it was performing decently well on all the social media channels and even in some ads that I ran as an experiment for the playable trends that I posted on my sub stack.

So,

when I talk about the creative-generated content, again,

yes,

it still kind of

has the same

requirements as the best kind of like the anatomy of the best performing ad.

Great hook, great kind of meat, which is the middle of the creative, and then the the call to action.

And then I was talking about ugly ads that are native and look

really not like an ad.

And this is something that shot at home.

You need to see, you need to hear all the music, and pew, pew, pew, playing with all the

toy soldiers, and then kind of seeing all the gameplay from this.

It was performing super well because people are not really

they didn't really understand

it's actually an ad.

And it was.

And this also was shot by

my dear friend Laurie from Longleaf Valley as a UA manager.

Firstly, it was run on TikTok as an kind of organic post, but it was getting so

good

feedback in terms of the comments, likes, and the views.

Then I kind of

thought, oh, okay, let's run it as a UA video.

Immediately, the best performing video on Facebook for months, because it doesn't really look like an ad, right?

And it's very native

and it's very believable as well.

So this is something that

you should start doing.

And e-commerce are already right there.

Some apps are doing it as well.

So good.

All the trends, obviously, there are so much more trends that are kind of changing on a monthly basis.

So, check out all the trends on our YouTube channel.

But then we come back to the actual creative testing and the benchmarks and

how different people test creatives and how much money they kind of allocate to creative testing.

It's interesting that people are doing like 10% of the UA budget on creative testing.

almost like all like 75%

which is something that I usually do

depends on the level of budget, but still 10% is a good amount.

What is shocking is that people are producing less than two videos a week.

That's not enough.

And only 18% produce five and more videos a week, which is a great start.

You should start there.

And also, as soon as you branch out of Facebook and Google, go for playables, obviously.

So

go for at least three to five videos a week, ideally five plus.

And you can use AI to produce more.

You know, just think about the iterations that I was talking about, but then you launch the creatives into the campaigns.

You immediately see if it's performing well, if it's not performing well.

I mean, you have

in the campaigns on a daily basis.

You see it.

You have the gut feeling.

You have the numbers.

And after $10

spent, you see if the creative is going to work or not.

You are in the campaigns, you know what I'm talking about.

So, how can you actually test different creatives?

Million different frameworks.

It's all like different things on Facebook, Google, and Appleby and other

kind of channels.

For me, right now, as I was talking about the volume, I don't really have a creative testing campaign if I don't really need to.

If I need to, it means I have 20, 40 creatives a week, so I need to kind of test them in a separate campaign.

But what I

do now is just launch or add new creatives into the business as usual campaign twice a week.

I don't really care.

It's not really restarting the or yeah, restarting the learning phase anymore.

So it's fine.

So I just launched new creatives into the campaign twice a week.

There you go.

If you don't want to do that, here's the option, which is still one Geos

150 per day and/500 or 1000 per day.

It depends obviously on the scale of your UA operations.

But just add new creatives there after four days,

post them, and the best performing move to business as usual.

And repeat every time or every week, every time you have new creatives.

We have some other kind of UI channels, obviously.

Unity has its own creative testing campaign.

Google doesn't really have this, but I just add new creatives as a new ad group in the business usual campaigns.

TikTok, oh my god, TikTok is just creative hungry.

10 videos or eight videos, like three times a week, as soon as I have so many of them

from the TikTok creative challenge.

And every time there is a new ad group and a new winner, just promote the winners into the main ad group, rinse and repeat.

The main interesting thing from the uplavin episode was the was the fact that uploading actually allocates like 5% of the campaign budget to the new creatives.

So you don't really need to have a new creative testing campaign you just launch again a new camp new creatives into the campaign and it will test and if it's if it gets

decent results it it's gonna get more and more traffic i've seen this recently we produced few creatives for for one game they tested it on facebook none of them got any type of any traffic.

Like, okay, interesting.

Then they launched it on Applavin and six out of eight immediately got traction, and two of them became a creative winners afterwards.

So

not all the creatives are suitable for all the channels.

And the testing, you know, creative winner on Facebook, it's gonna outperf almost, almost always outperform

everything that you can.

So

AI.

Let's hear this out.

AI is gonna take your job.

If not AI, I will.

Jokes Jokes aside, 68% of respondents in my survey use AI to assist in the creative framework.

Gemini, ChatGPT, Midjourney, CapCut, and Synthesia are common tools for producing creatives.

Cling AI, Runway Sora, and Google VO2 are also used.

Use cases for AI, UGC scripts, thumbnail optimization, first-frame hooks, iteration, audience research, and many more.

Yes, if you are not using AI, you are already 2.5 steps behind.

Pun intended.

Okay, let's get back to the real presentation.

This is so meta, because this is not me, but it's my AI avatar.

Exactly.

This is the future.

But jokes aside, it actually is an integral part of our creative process, even with my team.

As soon as I introduce the AI, MidJourney, Clink,

or even ChatGPT to the

image regeneration or anything related to the scripts or whatever,

text the whole production.

I mean, it's inevitable.

So, start using it yesterday.

It's insane.

And it also kind of goes into the UGC,

AI UGC, because it's not always only about the

talking head anymore.

It's the in-position kind of

UGC stuff

because

it also evolves, right?

We don't really

we can't even say who is real and who is not real.

I mean,

if it's good or bad, I don't really know, but still, it really speeds up the whole production.

And

this lady in the yoga position, one of the best performing videos from the past.

So, again, if you need a new UGC and the UGC already works in the games that you manage, help yourself out with the AI UGC.

There's so many different tools.

Every week there's a new one.

So just go for something which is not really the same as everybody else is using because nobody wants to see that.

And when it comes to AI in general, we have so many games using this, and we have, I mean, we might have

the AI corner that the Yakube is always pushing for with the AI because it's becoming more insane and more and more insane.

But these are just some examples from the Forex games that they are using this.

And I think the Arcor started using this in last year in september in november and since then never stopped and it's very very attention grabbing very novel people want to watch this and it works really well on facebook for some reason

especially on facebook i have more forex kind of examples these are some of our very favorite ones

because it's I mean, it's very unique, let's call it that way.

And I still want to watch it.

So as soon as you want to watch and you watch the video for almost a minute and 15 seconds, that's a very positive signal for Facebook.

And I would say also for other your channels.

But it's not only about

Forex games.

This is again an AI,

oops, AI UGC that kind of combine

AI that combines UGC,

link AI,

and some other kind of trends.

And again, like performing super well.

Because it's again, it's something that

you want to watch until the very end.

And nothing in this is actual real person.

So I think it's

based on AI, which drives me to

the whole kind of skill set in the UA field for 2025.

And it's not the same as it was before because I mean, even monkey can just push buttons, right?

But these days,

your manager really needs to

know how to tell a story.

It needs to kind of know how to write a good prompt for any type of AI.

And that's where everything is heading, right?

So you you need to go from just clicking the buttons to be very creative-oriented and very kind of engineering the prompts.

That's just to get the best outputs.

So, if you

if you want to ask your team and push them after this episode, ask them how many creatives are you producing or they are producing every week?

What's the CPI or the ROAS by

platform and the campaigns?

What's the spend goal?

If you hit the target, what's the payback?

And what kind of AI tools they use?

And if they are kind of

testing new channels frequently.

If they are, perfect.

If they are not, well, it's

something they should definitely do.

So don't use the good old tactics, but kind of use

the new stuff and new storytelling, testing new creatives producing new creatives testing new channels diversify diversifying from from the facebook and googles to all the ad networks and testing the reward at ui play

because that's where all of us are going at the moment so guys thank you very much for listening and to tuning into this episode this week

if you have any comments or any

anything you want to share with me on or the crew or Felix and Yaku, please share it under the video or join the Slack channel or see you somewhere later on.

Thank you very much.

Cheers.

Bye-bye.