🤩How to prepare for a post-IDFA world as a marketer on iOS 14 – with Eric Seufert, Thomas Petit, and Nebojsa Radovic💥

In today’s episode, we have an all star panel featuring the smartest folks we know. The episode features Eric Seufert from Mobile Dev Memo, mobile growth consultant Thomas Petit, and Nebojsa Radovic from N3TWORK.

What are advertisers solving for? What kind of advertisers will be the most and least impacted? Of the solutions proposed, what’s real and what isn’t? Each of our superstar guests offers their tips, strategies, and opinions on what the mobile marketing world will look like post-IDFA – and more importantly how to prepare.

KEY HIGHLIGHTS

🤔What strategies are realistic and which strategies are not?

🔄Is Apple going to provide loopholes for advertisers and ad tech companies?

✨Deterministic measurement has been an illusion all along.

🙅🏻‍♀️Why investing in a short-term, potentially fleeting, strategy is not advisable

🕸The importance of embracing the complexity of attribution by nature.

🗓How figuring out workarounds might help in the short term, but there is a need for a more scalable solution in the long-term

📝The importance of a publisher that’s really good at technical UA and understands incrementality.

👀If you’re a small developer, what sort of publisher should you look for?

💰What will be the impact of this change on ad monetization dollars – and specifically hypercasual developers?

💡What are some of the reasons other than privacy behind Apple implementing this shift away from iOS 14?

🌀What the transition period right after the introduction of iOS 14 could look like.

⬇️Why a small drop in costs and efficiency might mean you aren’t able to spend anything.

🕶What will the opt-in rates and app revenues really look like?

👩‍👧‍👧How apps for kids are already dealing with some of these regulations and challenges.

🤑Will the spend shift to Android?

⚡️Why subscription based apps might have an advantage over others.

👊🏽Subscription apps might be much more aggressive on their first time experience.

😮Where Apple may iterate with SKAdNetwork.

✍🏽Why subscription apps might add signups to their app onboarding flow.

🔨One workaround might be to get the IDFAs of only high-value users.

🤖What sort of games might be most impacted post iOS 14.

🙌🏽How apps might lean into the SKAdNetwork approach.

KEY QUOTES

Wait for things to settle down

As an advertiser, if you’re building your infrastructure, you want to build it once, you don’t want to build it 6 times or 8 times to accommodate these short term solutions. I think, as an advertiser, you probably just think about, “Okay, what can we build that’s gonna withstand the test of time here?” I think, just from a pure efficiency of resources standpoint, you probably want to be focused on the longer term solution versus a bunch of short term solutions.

Attribution was never perfect

What we believe with the deterministic solution, we had a lot of people believe that this was attribution. It’s partly fallacy, attribution is very complex. That’s why when you look in App Store connect, and Apple is telling you that 70% of your downloads are organic and then you look in your MMP and it’s 30% organic — something is wrong here.

What takes the place of payer profiles

Publishers that have a lot of payer profiles, the impact of those payer profiles will diminish — it just won’t be as big as it was before. So it might make more sense to work with a publisher that’s really good with technical UA, understands incrementality, understands how to do a probabilistic distribution as Eric and Thomas talked about than the work with a publisher that has a lot IDFAs that they could retarget because that won’t be possible anymore.

iOS 14 has larger goals beyond privacy

I argue this is one of the big reasons Apple is making this move at the end of the day because they never liked the ad monetization system, but also because their goal is to grow the services category. That goes through selling more search ads, but that goes mainly through selling an IAP– search ads are 1 and IAPs are 100. So I think pushing people towards IAP and subscription is one of the goals behind this change.

There will be efficiency challenges

If the efficiency drops by 20%, you might not be able to spend anything because you’re no longer able to target those high monetizers, and you’re trying to average out their contribution via 5% of a broader pool of people that you’re onboarding versus now where Facebook is able to increase that representation. And it’s still really hard to scale.

This is not the end for advertizing to iOS users

I had this comment on LinkedIn recently, somebody saying, “No, everybody’s going to switch all their budget to Android.” I don’t think that’s gonna happen for the main reason that iOS users are just spending so much more.

Subscription models are more dependable

The fact is a majority of subscriptions happen on day one. That will make reporting a little bit easier because you can extrapolate, in a slightly more reliable manner, your day 1 events to LTV. Also, because subscription doesn’t have this whale structure of revenue, and pretty much everybody pays the same.

Work with what is available

My proposal right now for the team and for some other people is basically to build a solution that will just ask for the IDFA of payers where then you would be able to share the revenue and payer information with the platforms and potentially create a workaround that way. So it might not be 10% it might be 20%, which is enough to get just the IDFAs of high value users.

FULL TRANSCRIPT BELOW

Shamanth: I’m very excited to host this all star panel on the Mobile User Acquisition Show. I have with us, Eric Seufert from Mobile Dev Memo, mobile growth consultant Thomas Petit, and Nebojsa Radovic from N3TWORK. All three of you have been fairly vocal, fairly opinionated around everything that we are seeing and everything that will unfold in the next couple of months. I couldn’t think of smarter folks to have on the show to really distill what the short and medium term looks like for mobile marketing and mobile measurement, specifically how we all look at things.

Eric you certainly popularized calling this the apocalypse, but what does the apocalypse look like in the specifics? We’re very thrilled to dive into all of this today. To start off, there’s been a lot of people talking and writing about what’s happening out there — certainly ad tech companies and MMPs that have benefited from the old paradigm of how mobile marketing functioned.

They’ve put out opinion pieces about, they’ve spoken about how they’re going to address the post iOS 14 landscape. Now, of all of the things you’re seeing out there, what’s realistic? What’s realistically possible? What just isn’t going to work? And Eric, if you want to take that, because I know this is something you’ve had opinions on, and would love to kick things off there.

Eric: I think if I try to evaluate all the proposals I’ve seen just on the basis of feasibility and viability in the context of understanding how Apple plans to enforce this change, which I think is very aggressive.

I don’t get the sense from Apple that the change was announced, but they’re going to provide a bunch of loopholes and work around that, it was just sort of a symbolic gesture, but it doesn’t really have any teeth, and they’re not gonna enforce it full throatedly. Within that context, I think the products or approaches I’ve seen that I believe are the most viable are the ones that fully lean into an actual paradigm shift, meaning that it is going to require people to start doing measurement in a probabilistic way.

I’ve seen some kind of media mix models be proposed, and my sense is that that is the way forward. First of all, maybe you can come up with a viable workaround. You have to acknowledge there’s going to be some sort of cat and mouse dynamic here that these ad tech companies are gonna put forth the solutions that aren’t fully compliant with the spirit of these privacy changes, and then Apple’s gonna shut them down, and they’ll come up with a different workaround. There’s going to be maybe a year or 18 months of that kind of back and forth.

My sense is that it’s mostly having to do with the analysis and measurement. Deterministic has been kind of an illusion for a long time — you had LAT, you had SANs that basically front run that sort of conversion claims. If you acknowledged that that was an illusion all along. Just look at some of the bigger advertisers moving this direction anyway, even before this was announced, even before they could guess that this would come, even before they could anticipate this. A lot of people are moving that direction anyway. So it feels like given that there was already momentum in this direction, it’s probably the right direction.

Shamanth: So definitely this is going to be more and more of a shift towards probabilistic measurement. The short term fixes people claim, definitely ad tech companies, because that’s where their bread and butter lies, they claim these short term solutions, but you don’t want to be building on that. I’m curious, Nebo, Thomas, do you guys have any thoughts on this? What is realistically possible vs what isn’t realistically possible?

Thomas: I’d agree with Eric that the workarounds are not going to stick forever and investing in something that might get shut down anytime is a bit of a wrong bet. Sort of like you’re Criteo in 2017 — that’s not a good position to be in. To be honest. I think a lot of it is because we’re before the launch. Something I’m worried about is the timing because we have very little time to prepare like us as advertisers, but also vendors.

I think Apple could have done something a little bit better, like releasing the SKAdNetwork before iOS 14. No, What I’m thinking is a lot of ad tech is trying to see where they can move, but I think the sky is going to clear very fast. We’re in this bizarre situation now because it’s July, August. Two months from now, all of these loopholes are not going to stick.

Shamanth: We’ll know where Apple stands on the loopholes and how things go.

Thomas: I would also agree with Eric that

what we believe with the deterministic solution, we had a lot of people believe that this was attribution. It’s partly fallacy, attribution is very complex. That’s why when you look in App Store connect, and Apple is telling you that 70% of your downloads are organic and then you look in your MMP and it’s 30% organic — something is wrong here.

I don’t know who’s right, but I know something is wrong. I’m not saying Apple got it right either. But in a way, it might be good because maybe it will make us rethink in a very different way.

Attribution is a very complex topic, not only on the technical side, but on how you interpret it. I believe, personally, what the MMP should do is embrace the fact that some people recognize them as attribution specialists, not deterministic specialists. There will be complexity in mixing up what comes through SKAdNetwork, what comes through media mix models, and it’s not going to be easy at all. It might not be as profitable as it used to be, but I think they should be the ones who are going to educate people around the fact that it’s going to be more complex to approach attribution. The previous one was simple, and that’s why people loved it. The new one might not be as accurate, but it will force us into embracing the complexity that attribution has by nature. In the short term, it’s a problem. In the long term, it might be an opportunity.

Shamanth: We’re making sense of an imperfect world. Whoever does that best will really win in the medium to long term.

Nebo: I agree with everything Thomas and Eric said. Figuring out workarounds might help in the short term, but we need to come up with a more scalable solution that will work for months or years and not days. I think we’re all on the same page there.

Shamanth: Right. Nebo, I know your team works on publishing and scaling games, so you perhaps have a more nuanced understanding of how this can impact many app developers, especially the small and medium sized developers that typically benefit from publishing. I do know that a lot of mobile publishing has been very UA focused. So how do you think iOS 14 changes things for publishers, developers who rely on publishers, and specifically gaming. What’s your sense about how this will play out, especially for the smaller app developers?

Neob: I have a few theories. The first one is that

publishers that have a lot of payer profiles, the impact of those payer profiles will diminish — it just won’t be as big as it was before. So it might make more sense to work with a publisher that’s really good with technical UA, understands incrementality, understands how to do a probabilistic distribution as Eric and Thomas talked about than the work with a publisher that has a lot IDFAs that they could retarget because that won’t be possible anymore.

The good thing about payer profiles is that Facebook and Google are already moving away from them. Facebook is testing triple A ads that don’t really leverage look-alikes. That means that they won’t be as necessary.

I think the key point here is that if you’re a small developer, you should really look for a publisher that’s really good at technical UA, that understands how to find players without AO and VO, how to do the proper campaign setup in the post SKAdNetwork world, and understands incrementality lift studies, etc. Another interesting thing is that it’s questionable whether DSPs will exist. They will to a certain extent, but without the ability to target specific IDFAs or to be more precise to filter out specific IDFAs that are never going to install your app, it’s likely that the cost is going to go up. Publishers with a DSP, at least in the short term, might not really have that competitive advantage.

The last bit is when it comes to the type of games that small developers want to build, I don’t really understand yet what’s the real impact on ad monetization, and whether the fact that the big networks, especially Facebook and Google, who are kind of good at cherry picking impressions, are still interested in paying premium CPMs for those impressions. Now that they’re really not able to identify who the user behind the impression is. And what’s basically the effect of that on impression level bidding and impression level revenue reporting because that’s going to be crucial for hyper-casual developers in understanding how good the monetization is for the games they’re building, so that might cause a certain shift in the types of games that are being built. Developers might move away from and monetize games that monetize more with IAP, but they’re just the guests obviously.

Thomas: Which

One of them — maybe the only one — but so that makes sense.

Nebo: Yeah, Apple arcade as well is directly competing with hyper-casual developers, and they want to have a bigger piece of that pie. So that’s definitely one of the reasons for this shift.

Shamanth: So you’re saying, looking at the macro picture, Apple isn’t getting a 30% cut of ad monetization revenues, and that’s really one key motivation. Do you guys anticipate come September-October, ad CPMs are going to crater? Do you think that’s going to somehow make up for that reduced targeting and reduced ROAS numbers that we are going to see?

Thomas: I believe we will see at least a temporary drop because some people will freak out and at least assess their position. I think that there are really two periods here like the transition period in the first 1-3 months, and then people will readjust. I’m not sure about next year. For this year, my question is not so much about how much they will drop, but is it gonna be enough to compensate for the lower side of optimization?

I think this is one of the big unknowns of this problem. We know the machine learning algorithms are going to operate in a weakened manner, but I’m not sure CPM’s are gonna overcompensate it.

My experience in UA is when there was a smaller shift than this one, things adjust. At the end of the day, the big advertisers are and then make the bids based on what they returned. I guess very soon we will find a new balance, but there’s a big unknown on the transition period for sure.

Shamanth: Sure, and Eric, you wrote this week about how at least 50% of Facebook’s efficiency is attributable to impression level monetization. You quoted a paper from Facebook. Curious, how do you see this play out?

Eric: Well, they said that — that was not me. On FAN, the personalization of the ad placement made up for 50% of the CPM price. I think the problem here is that targeting is the product, and so that’s what validates the price. If you get rid of the targeting, I don’t know that you could say that now we’re going to a really generic more broad level of aggregating revenue around these different placements and then seeing which of those were worth it. I think you can’t target these users individually.

We know that the distribution of spend is a super long tail, especially in gaming, people that monetize to a very high degree. We know that there’s a big fat portion of the curve at zero, and so then when you say we can no longer target that super long tail, it’s not like efficiency drops by 20%. Maybe say, “Hey, our costs drop by 20%.”

It’s still really hard for gaming companies to scale even though they’re using the most sophisticated sort of machinery in the world to hyper target the people that are most likely to spend in your apps — it’s still hard to scale. So without over indexing on those people that do spend money, I don’t think you can say, “Well, we know that we’re bringing in 100 users, and that for every hundred, one will spend.” I just don’t think user economics work that way.

Now, I think the big question for me, and a lot of the pushback I’ve gotten in that article, is a couple things. One is people think I’m being overly pessimistic around the opt-in rates, and they think the opt-in rates will be 70%. To me, that seems delusional, but I would love to get everyone’s opinion on that. When I’ve used an estimate for opt-in rates, they’ve been indexed at the poll that I did on the Mobile Dev Memo Slack where most people said it was gonna be between 10 and 20%. The vast majority of respondents said between 10 and 20% — that’s where I’m going.

Also, I’ve taken some guidance from the actual GDPR opt-in rates, and when the opt-in mechanic is not gamed, in a lot of cases there’s dark patterns that get you to click because you want to get it out of your face. When the GDPR opt-ins have conformed to the spec that they’re supposed to use when they build these tools, opt-in rates are about 10%. That’s not a guess on my part, that was a paper that was floating around like two weeks ago. Apple is not allowing anybody to make games that pop up because they control the design of it and the headline text. So people said, “No, you’re overly pessimistic. We think it is going to be 70%. It’ll have an impact, but it won’t be so dramatic.” Then, some of the people said, “You’re way overestimating the share of app revenue for Facebook.” For Facebook, maybe 10% or 20% of the revenue comes from app installs. It’s not a majority, so it’s not going to hurt Facebook that much. It’ll hurt app advertisers, but Facebook as a company won’t be that bad.

I would love to get people’s feedback on those two points. On the Facebook point, especially, it just seems not credible to me that their app install revenue or their app advertising revenue is much less than half because they’ve said publicly that 94% of revenue comes from mobile. So where is it? How is it all just web based DTC? How is the majority of the revenue not app if the majority of all of it is mobile?

Shamanth: Exactly. With games, we know how much mobile ad spend is — it’s expected to hit $240 billion this year. It’s reasonable and safe to assume that a lot of that has gone to Facebook. I’m sure that there are estimates that say what percentage of that goes go to Facebook. To switch gears a bit, Thomas, you’ve worked on many different kinds of apps, many subscription based apps, certainly a couple of apps for kids, and on Apple search. What are some of the pieces of advice that you’re giving to app developers that you’re working with, considering they are so heterogenous?

Thomas: The main advice I tell them is be ready. We can’t do anything, but we need to act super quickly as soon as we are. Everybody is trying to figure out how we’re going to do it, but it’s not like we can already deploy infrastructure that we will need by September. So what I’m telling them is be ready because in two months, we’re going to need to act right on. This thing is going to happen immediately. We need to plan it in Q3 and Q4 and the roadmap that we will need to shift our whole infrastructure there, but we can’t even really get started. You can poke around the documentation but it isn’t the same. The first thing I tell them is to put this thing on the road map because it’s huge. It’s changing a lot of things. That’s the first one.

There is a particular case for the kids’ apps that Apple, after doing some bizarre change of dates around, started enforcing that the IDFA is dead for kids by the 30th of June. It’s been about three weeks, and I can tell you it’s not fun. There’s one particular situation where it’s a war room on fire every day. I believe that they will end this with an advantage in the sense that they had to fast forward their thinking and shifting strategy. How do we prepare earlier than the other people? When September and October come, there’s a lot of research, and who’s in charge of that, and the endless discussions and reading that will have been done. It’s a bit of a tough situation now because they can’t use the IDFA and they can’t use the SKAdNetwork either, which brings me to an interesting topic.

If the pressure on the Android inventory is raised by that much, it’s not going to be profitable for people, but this is the kind of thing that we have to think about for the very short term because for the kids apps that are in this intermediary situation where we don’t have the new solution, but we don’t have the old one either. It’s a bizarre thing. It’s not a great position to be in. I chatted with kids’ apps that are fingerprinting, some others who stopped UA, some others who ran blind. It’s a fun thing, but I think they will be more prepared by then, but it’s tough.

You mentioned subscription. I think here subscription services have a slight advantage to others, typically gaming with very long term retention, where you’ve got cohorts of people who keep retaining and paying for months, sometimes years — the very big hits of the App Store. They’re all four or five years old, and they have not all the users are new, those who are making money. Some of them are fanatics who have been playing for a very long time. This is going to be very hard due to the mechanics of SKAdNetwork. The thing that benefits subscription slightly, but I think is bad for Apple, is a lot of subscription happens on the very first day and a lot of them happen on your onboarding.

It depends if the developers have a bit more what I call aggressive with his onboarding paywall or not.

Still the fact that for most subscriptions, a majority happened on day one. That will make reporting a little bit easier because you can extrapolate, in a slightly more reliable manner, your day 1 events to LTV. Also, because subscription doesn’t have this whale structure of revenue, and pretty much everybody pays the same.

It’s not entirely true, so that makes it a little bit easier. The thing that I see as a negative there, and I don’t know how Apple is going to react in this, is that it pushes everybody to be more aggressive to make the most revenue and extract as much revenue and data as you can in the first time experience.

I’m a little bit afraid of how far these strategies can go because I know I’ve crossed the line a couple of times myself with some products where we had a really nice uplift, but the expense was not great. If everybody moves to a more aggressive position because we have this timing issue on the event, it might be a bit of a problem in the medium.

That’s typically where I think Apple will iterate on SKAdNetwork because they’re not going to go back on the fact — this is staying like for sure. I do believe they will iterate a bit. The first thing I would recommend them to iterate, if they want to listen, but where I think they will end up iterating, because of what I just said before, is the ability to receive different events for different periods of time than just having this first day and then this bizarre random timer that they put that basically makes the whole thing impractical. So I think that would change, but I don’t know. I’m just guessing here.

Shamanth: Yeah, and I think definitely subscription apps can be at an advantage because a trial oftentimes after the first couple of hours typically.

Thomas: Yeah, you have to monitor how you convert your trials after. I’ve said before with all these changes that a lot of subscription apps have moved to app event optimization based on free trial, but if you go too far with that sometimes it’s actually a counterproductive in the sense that you attract more and more cohorts that don’t convert to trial because some, especially younger audiences know exactly how to control trials. Here it will make things a little bit more complex.

Most subscriptions have signups, and they will do it internally from their own, so we won’t be able to attribute to that ad and that ad group. We can study how the cohort evolves and correct it from a product point of view, which by the way makes me think that a lot of people are gonna add signups to the app. A lot of gaming apps don’t use them because they want to ease the process of just coming in. The subscription is pretty much the opposite — almost everybody is forced to sign up before being able to even use the app. I think sign up is going to be more common in the future, which from a user perspective is a bit bad because it’s friction at the beginning, but I guess it would give options for more personalization and so on.

Shamanth: Definitely. I know we have almost come up on time. Perhaps as a good place to wrap, Eric and Nebo would love to hear from you guys about your advice. What are you guys advising people and apps you guys work with to close this out?

Nebo: My main advice is just to wait and understand the changes coming and learn about the issue as much as possible, but I think it’s really hard to know how things will look in two months from now when things go live. I am slightly less bearish on the IDFA opt-in rates. Now, we know that basically we can ask users for that permission later.

The other thing is, I’d be cautious when planning new products and make sure that gaming companies are having a well diversified portfolio because we don’t still know who’s going to be the most impacted. My thoughts are maybe social casino games will be most impacted because of the focus on high value players, but we don’t know that yet. So just learn about what’s going on, read all the newsletters, and what the proposals and MPs are providing. Then we’ll know in two months from now what the best solution is. But we’ll probably spend a few more months figuring this out.

Eric: I think if you want to lean in heavily to the SKAdNetwork approach, then you want to start thinking about how your monetization ties into this conversion event timer, and how you use that to the most possible advantage. If the ad network is still optimizing at the campaign level around the events you are getting back, you can craft that user experience in such a way that you get to the highest value conversion event within some reasonable amount of time after the install, and then you can still sort of get that data back.

That’s the problem with it — trying to delay the timer on the IDFA pop up is that then you don’t get access to the IDFA before that. You don’t get access to the IDFA until they say yes, and they don’t ever say yes, and they don’t ever get it. So the idea there is to convince the user that “Hey, there’s value here, and I should let them track me.” Maybe even incentivize it. Or you’re delaying that point at which you get the IDFA, and they can use that for ad optimization, so there’s a trade off.

But my sense is this sort of accelerates the trend of trying to maximize the opportunity that you get in the overlap between the period of onboarding when you still can send a lot of meaningful data back to the networks and that of monetization design. I feel like the kind of economy design is now going to be even more tightly integrated with the UA strategy, which was the direction things were heading. The SKAdNetwork sort of approach now magnifies the importance of that.

Shamanth: Definitely. And, Nebo, you definitely spoke about that in the episode on this show earlier. I think that’s definitely going to get magnified. I think this has a far reaching impact — not just for UA and marketing — but for products and how we build them. We will definitely keep a lookout for how things evolve and shake up. Certainly I’ll be looking at your input off channel. Everywhere you guys write and speak, I’ll be looking. This is perhaps a good place for us to wrap you guys. Thank you so much, again for being on the Mobile User Acquisition Show. We really appreciate your time today.

All: Cheers, guys. 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 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!