Note: We’ve had two BIG launches last month!

1. The Mobile Growth Handbook 2022 is packed with incredible insights along with tried and tested strategies, this book is the perfect tool to learn from and hone your growth plans for this year. We feature hand-selected insights from the smartest mobile growth marketers.

We have over 100 pages of insights that can be accessed at all times (as soon as you sign up). Download the book now and get a head start on your growth strategies.

Download the book now and get a head start on your growth strategies. https://mobileuseracquisitionshow.com/mobile-growth-handbook-2022/

2. The Mobile Growth Slack: A community that was a part of our workshop series – The Mobile Growth Lab, is now open to the general public. Join over 150 mobile marketers to discuss challenges and share your expertise. More details are available here: https://mobileuseracquisitionshow.com/slack/

If you’re ready to join the growing community, fill this form: https://forms.gle/cRCYM4gT1tdXgg6u5

ABOUT MATEJ: LinkedIn | Lancaric Website | two & a half gamers podcast

ABOUT ROCKETSHIP HQ: Website | LinkedIn  | Twitter | YouTube

KEY HIGHLIGHTS

⛏ What happens if you’re running your UA on only one channel

❗ Why you can’t rely only on Facebook anymore for UA

💪 The scale at which you should start diversifying your UA portfolio

🆕 The basis on which you should start adding new channels

💸 How your monetization model defines your diversification strategy

✅ The right approach to diversifying your optimization strategy

🔗 Things to remember when choosing an ad network

🥇 Which diversification should come first – channel or optimization 

👀 Working with SKAN in your diversified UA mix

🔧 How to build a custom schema within SKAN

🧐 Comparing UA channels post IDFA

🔌 How incrementality helps when diversifying your UA portfolio

📝 What to keep in mind when running incrementality tests

KEY QUOTES

One UA channel is not the safest bet

There were times when Facebook was doing well, and then the quality decreased—everything was changing. So if you put all your eggs in one UA basket, you are playing a very dangerous game, I would say. If you’re running your UA on one channel and that’s the only channel, what if that stops working?

When to diversify your UA portfolio

If we were having this discussion three years ago, I would say Facebook only, up until you hit around $1 million a month in spend. But things have changed quite rapidly. I start thinking about diversifying the UA portfolio after hitting like $20,000 a month.

How to go about the diversification process

I always tried to maximize the spend in one channel. So, doing the diversification of the optimization strategy first and then moving slowly to other channels, step-by-step.

Trials and errors post IDFA

I tried basically everything when these IDFA shenanigans happened, from funnel, engagement schema to revenue schema. I’m now working mostly with custom schema, where I use four to five engagement events, depending again on the game. Then I add the revenue brackets on top of it.

How to set up revenue brackets

If there is a possibility for players to buy something within 1,000 euros or dollars on the first day, then it won’t make sense if I set up brackets by $1, because it’s a huge gap. That’s something that I keep in mind – based on the maximum amount of money that can be purchased or can be spent on the first day, I set up the revenue brackets.

Incrementality is nice to have

I opened up a Google campaign, and suddenly I was able to see a huge install and revenue increase in organics. And there was nothing else that changed.

Obviously, I could see this even without running an incrementality test. But the incrementality test just said, “Okay, so you’ve arrived, but the question is, do you want to run this campaign or not?” Of course you want to run it, because this is definitely having a positive effect on both installs and revenues!

FULL TRANSCRIPT BELOW

Shamanth

I’m very excited to welcome back Matej Lancaric to the Mobile User Acquisition Show. Matej, welcome back to the show.

Matej

Thanks for having me again. This is the third time, I think. I’m super excited.

Shamanth

The first time was very early on in the podcast’s life, and certainly you have been very popular even then, early on.

Matej

Nice!

Shamanth

I’ve followed and admired a lot of the work you’ve done. We’ve done a little bit of work together as well, and I definitely admire everything you’ve shown and taught to people.

For the episode today, I’m excited to have your inputs and insights about a lot of what’s changed in the post identifier world today. And we’re going to talk about diversifying the user acquisition mix in this post-ATT world. So let’s start with the basics. Why think of diversifying your UA channel mix at all?

Matej

These days, everything is changing so rapidly. Last year, I sometimes felt like I was riding a roller coaster!

There were times when Facebook was doing well, and then the quality decreased—everything was changing. So if you put all your eggs in one UA basket, you are playing a very dangerous game, I would say. If you’re running your UA on one channel and that’s the only channel, what if that stops working?

I started to see this kind of behavior when I was soft-launching games last year. I’ve always used Facebook as my go-to channel, usually very early on during the soft launching. And then I saw a very big decline in terms of the quality of players. So I started adding Google and Unity campaigns early on, and I was able to get more data points for benchmarking in a soft launch. It was definitely very beneficial for making decisions, because for example, we were on the verge of killing some games. And then, when we suddenly started running Unity campaigns, the numbers definitely looked better than what we saw on Facebook.

That got me thinking. I couldn’t rely only on Facebook anymore, and I definitely needed to start running more UAC. I’ll give you an example of a situation from last year. During one week, the Google ads campaigns, all the tROAS campaigns, were performing super well. Then in another week, Unity was scaling amazingly because we found a new creative winner.

Then Facebook AAA campaigns were down, but on the other hand, interest and lookalike suddenly worked! So I was left wondering what was happening! So we definitely need to be on top of our game, always.

Shamanth

So what I’m hearing you say—and I think we’re talking primarily iOS here—just because Facebook has been so up and down, it’s not been reliable enough. That’s certainly been our experience too. And like you said, it made the difference between killing a game and keeping it; so there are very high stakes there.

Matej

Exactly.

Shamanth

Let’s assume a game is out of soft launch. It is live. At what level of scale should marketers start thinking about diversifying?

Matej

If we were having this discussion three years ago, I would say Facebook only, up until you hit around $1 million a month in spend. But things have changed quite rapidly.

I start thinking about diversifying the UA portfolio after hitting like $20,000 a month.

Now don’t get me wrong. Facebook is still my go-to channel, even though the quality has decreased heavily over the last few months. But I just don’t want to rely on one channel.

If you’re running small campaigns here and there, then you can still stay with one channel up until $20,000. Because let’s face it – you need to spend your small budget wisely. And if you spread that $20,000 into, say, five channels, then you will get a very, very small amount of data that won’t really have any significance.

Shamanth

So at $20,000 you start looking at additional channels. What happens as you scale from $20,000 to $50,000 or $100,000 and so on? When do you introduce additional channels from there?

Matej

It will always vary from game to game. But I’d start adding new channels based on the CPI and LTV equation that I can see for that specific game.

So there would be a new channel at $20,000, then another at $50,000, and then after that, it would depend on what the campaign structure looks like, and on the kind of monetization model in the game. If we were running, say, 50-50 in-app purchases versus ads, that would open up some very nice new possibilities. So I wouldn’t need to diversify the UA channels, but I could diversify the event optimization or the optimization strategy.

If it’s an IAP-heavy game, then anywhere around $50-80,000 and then $100,000 and above, I’d start adding new channels right away.

Shamanth

For folks who may not be familiar, can you tell us a little bit more about what you mean by diversifying the optimization strategy?

Matej

For different types of monetization in the game, there are different events we can optimize for. So for example, if you have an ad-monetized game, you can run ad-ROAS campaigns, which is basically the same thing as value optimization on Facebook or tROAS optimization on Google, but for ads revenue.

Let’s take a simplified example – you can optimize for ad whales. And then you can also optimize for different ads watched events as well, when players see five rewarded videos, or 10 or 20 or 50 or however many you set up. But then you can optimize for different engagement events in the game, like level 20 achieved, tutorial completion or something else down the line.

So that’s the diversification of the optimization. Purchase optimized campaigns and value optimized campaigns also come to mind.

Shamanth

So you’re saying, look, optimize for different ad events, purchase events, and maybe there’s a non-purchase non-ad event that’s worth optimizing for as well.

Matej

Oh yeah, definitely. You can see different LTVs on different UA channels. The same thing goes with different optimizations and different retention profiles as well. 

Shamanth

You’re doing this within Facebook, within each ad network separately?

Matej

Yes, it’s definitely Facebook and Google. Then AppLovin can run the CPE campaign, which is the campaign for event optimization. You can use different engagement events — ad ROAS campaigns on AppLovin, purchase optimized campaigns on AppLovin and on Unity as well.

There are a lot of similarities between the ad networks. But you need to keep in mind that there are some differences. For example, with Unity, if you have an MMP like Singular, you can’t run ad ROAS campaigns because they are not connected. But you can do it if you have AppsFlyer, for instance.

Shamanth

So there are still a lot of options and possibilities for diversification within channels on iOS. And presumably that’s because on Facebook, you still have some modeled conversions for the optimization to work. Unity, ironSource, AppLovin have probabilistic for the optimization to work very well.

And when you’re at, say $20,000, are you doing different optimization types within Facebook? How do you approach different optimization types on such small scales?

Matej

I always tried to maximize the spend in one channel. So, doing the diversification of the optimization strategy first and then moving slowly to other channels, step-by-step.

Shamanth

So if you’re starting on Facebook, you’d start with purchase optimization, ROAS optimization, ad monetization optimization. Once you max that out completely, you move to channel number two?

Matej

Exactly. It also depends on whether you have loads of optimization opportunities. In that case, the $20,000 can change to $50,000 or $100,000.

Shamanth

Yes, and I imagine that depends on the app’s monetization models. And like you said, if the model is a hybrid of IAP plus ads, more optimizations are obviously possible, but if it’s something like, say, social casino and almost everything’s IAP, then, you can’t even optimize for installs.

Matej

Yes, in that case it’s very straightforward.

Shamanth

The other dimension of setting up channels post ATT is SKAN. Within SKAN, what sort of schemas do you find effective and what others do you find ineffective?

Matej

I don’t think it’s all about what sort of schemas are effective or not effective. You just need to find one that works best for you.

I tried basically everything when these IDFA shenanigans happened, from funnel, engagement schema to revenue schema. I’m now working mostly with custom schema, where I use four to five engagement events, depending again on the game. Then I add the revenue brackets on top of it.

So we measure and send as many signals as possible back to Facebook and Google or any other ad network. For eight games out of 10, the custom schema and revenue brackets worked. But yes, there’s the caveat that we were able to get a really large number of purchase events in the first 24 hours. So that’s why I was able to scale the iOS campaigns.

And this is the part where the UA loves the product and game teams. We need to be very close to each other and discuss these opportunities together and try to find out how to get as many events as possible into the first day, and send these events back to the paid UA channels and not destroy the economy of the game.

It’s all about these little tricks and hacks and then the discussions with the teams.

Shamanth

And again, for folks who may not be very familiar, can you define what you mean by custom schema?

Matej

Sure. The custom schema is a combination of multiple schemas actually. So you can use the engagement schema and the revenue schema in one schema together. So that’s why I use the funnel version with several events. And then on top of it, there is the revenue schema with the revenue brackets.

Shamanth

Which means, the first few bids are the early stages of the funnel – register, complete tutorial, level one, level five, and subsequently, after bid six or seven, you just have purchases?

Matej

Yes, exactly. And based on the level of depth of the purchasing power of day one, I choose the brackets. For instance,

if there is a possibility for players to buy something within 1,000 euros or dollars on the first day, then it won’t make sense if I set up brackets by $1, because it’s a huge gap. That’s something that I keep in mind – based on the maximum amount of money that can be purchased or can be spent on the first day, I set up the revenue brackets.

Shamanth

That makes sense. Obviously, it’s because the first 24 hours are very, very critical, as that’s just the way SKAN works. So presumably, the funnel events you set up before purchase all have to be within the first 24 hours.

Matej

Yes, ideally.

Shamanth

The other important thing around SKAN is the privacy thresholds. I’ve seen Facebook campaigns where cost per unique purchaser is like $1,000, $2,000, $3,000, which doesn’t make sense. On the other hand, since something like Unity or ironSource runs probabilistic, it’s just a completely different method of measurement.

So, when you are diversifying, how do you compare these channels?

Matej

Since IDFA, it’s been really hard to compare those channels one-to-one and that is just how it works.

I always try to check what the revenue baseline looks like and then start adding the channels. So start one channel, then check the revenue baseline after a week, or try to get some level of data and then start adding new channels one after another. Please bear in mind that this is definitely not 100% correct, but it gives me some kind of revenue trend.

So if I’m running some campaigns on Facebook and I see a big revenue spike which is higher than what I’m spending, the margin exists there. That’s always a good indicator that I’m on the right track.

To address your point, if you see $1,000 or $2,000 CPA for purchases on Facebook, that can happen when there is around 30–40% of the UA numbers or UA players go into organic. How was I able to do that math? I used a tool called Incrmntal, which was able to help me identify these kinds of spillover for me. So I was actually aware of that as it was happening. I thought that around 20% were going into organic, because that’s what the data told me, but I wasn’t sure completely. So, they ran an incrementality test for me and then identified that 30% or even 40% goes to organic. And if 40% of your UA is going to organic, then yes, you’re looking at $1,000 CPA. That kind of sucks, but as you can see, that’s not the true CPA.

Shamanth

Absolutely. With SKAN, obviously there’s the inaccuracy because of the privacy threshold. With probabilistic, are you seeing signal loss and if so, how much are you seeing? Or is it just a matter of ‘what you see is what you get’?

Matej

There are differences between the games, but what you see is what you get. I know, again, that this is not 100% correct, and I’m sure that this will never be 100% correct. It’s a big black box. Of course, everything was a big black box even before, but now it’s just another one!

And what can we do? It’s still very viable to do UA on iOS. The payers didn’t disappear – they’re still there. It’s just a little bit harder to acquire them, but you just need to be a bit smarter about it or think about it slightly differently than before.

Shamanth

You’ve briefly talked about incrementality analysis. How effective have you found it? I know you gave an example of how you found the organics that resulted from SKAN. But again, at what level of scale or what number of channels is incrementality analysis meaningful?

Matej

Incrementality is something that is nice to have. Obviously, there’s no doubt you can work without it. But it’s nice to have. All these analyses just give you more data points, so you can make better decisions.

You can use it even when you have one channel, and you’re trying to measure how your UA activities affect the organic installs and revenue. You can use it when you are running multiple sources, and you can identify cannibalization with incrementality.

For example, let’s say you’re running a Facebook worldwide campaign. And then you decide to run a Unity campaign in Slovakia. So the Unity campaign will affect your Facebook campaign since you started it, because obviously, if you run two channels in one country, you want to see more installs and higher revenues.

But the incrementality test discovered that the Unity campaign is basically destroying the whole Facebook activity you are doing. So it doesn’t make any sense to run it. So instead of seeing more installs, you actually see either the same amount of installs for two channels, or even lower because it can cannibalize. That’s one example.

Another example is definitely more positive, and it happened to me in South Korea.

I opened up a Google campaign, and suddenly I was able to see a huge install and revenue increase in organics. And there was nothing else that changed.

Obviously, I could see this even without running an incrementality test. But the incrementality test just said, “Okay, so you’ve arrived, but the question is, do you want to run this campaign or not?” Of course you want to run it, because this is definitely having a positive effect on both installs and revenues!

So it’s a pretty helpful tool. I know it can be misleading sometimes, which is why you still need to be mindful of your own thoughts and experiences and what you see in terms of the data. But it can be helpful, definitely.

Shamanth

Absolutely. And that can definitely be powerful insights, but, like you said, you need to be careful about how you’re interpreting the data.

Matej

Yes, for sure.

Shamanth

And with an incrementality there’s like different techniques. You can build basic spreadsheet-based models. You could have relatively complex data science-based models. Are there specific techniques which you find more effective than the others?

Matej

For me, the most effective was using the tool because sometimes, when you’re diversifying the UA portfolio, you definitely need to put more work into the UA campaigns. And if there is a tool that can help you build the incrementality tests, and you don’t need to use any spreadsheets and build it on your own? That’s super effective for me!

Shamanth

True, plug and play oftentimes can work. And like you also said, you just need to be careful about how you interpret the output and data.

Matej

Yes. Of course, there’s one thing I need to mention – not every company can buy the tool, and they need to do it on their own. I totally get that. But if you are in a position to actually leverage these types of tools or any other tools, that’s perfect. Saves a lot of time!

Shamanth

Certainly. Matej, I think this has been very instructive, as has been every time I’ve spoken to you. You certainly pointed out a very important facet of UA after ATT, which is that you can’t really rely on just the big channels anymore. You do have to diversify.

This is perhaps a good place for us to wrap. But before we do, could you tell folks how they can find out more about you and everything you do?

Matej

You can find me on LinkedIn or my website, which is lancaric.me. Or if you are a podcast fan, we have started a new podcast called ‘two and half gamers’. You can tune in any time.

Shamanth

You were also an instructor in our workshop series, Mobile Growth Lab, the last edition. We are still in the planning stages of the next one. I hope you’ll join us for that one.

Matej

I will be glad to do it!

Shamanth

Excellent. We will append a link to your website and everything you do in the show notes. Thank you so much for being on the show, Matej.

Matej

Thank you, Shamanth, for having me again. Thank you. Bye-bye.

A REQUEST BEFORE YOU GO

I have a very important favor to ask, which as those of you who know me know I don’t do often. If you get any pleasure or inspiration from this episode, could you PLEASE leave a review on your favorite podcasting platform – be it iTunes, Overcast, Spotify, Google Podcasts or wherever you get your podcast fix. This podcast is very much a labor of love – and each episode takes many many hours to put together. When you write a review, it will not only be a great deal of encouragement to us, but it will also support getting the word out about the Mobile User Acquisition Show.

Constructive criticism and suggestions for improvement are welcome, whether on podcasting platforms – or by email to shamanth at rocketshiphq.com. We read all reviews & I want to make this podcast better.

Thank you – and I look forward to seeing you with the next episode!