🔍 How to win with Apple Search Ads in a post-ATT world: with Thomas Petit, Growth Consultant 🎭

Our guest today is Thomas Petit. Thomas is an independent mobile growth consultant working primarily with non-gaming B2C apps. He is an external consultant for large apps (including 2 unicorns), a collaborator assisting several app agencies and an advisor for very early-stage startups. Thomas has run campaigns since the first day of Apple Search, spending 7 figures on the platform directly. He has also run audits on many other accounts. He is certified by Apple, SearchAdsHQ & ASOdesk – and is a regular public speaker on the topic of mobile growth.

KEY HIGHLIGHTS

🤺 How you needed to adapt to LAT remained unchanged since 2016 until this year.

🦙 Apple’s Ad Services framework – and how it bypasses the IDFAs and users’ ATT status.

🖼️ How Apple defines tracking – and why the way attribution works for Apple Search Ads isnt considered tracking by Apple.

🧪 The mechanics of user level data measurement and tracking without the IDFAs.

🔍 The Ad Services framework doesn’t require an MMP.

🐶 The fact that LAT users’ performance wasn’t measurable in the past was a challenge – and that goes away now.

🌳 Why it makes sense to keep the older campaign structure even after the recent changes to ASA.

🤽 Search tab campaigns can actually work for returning users.

🛍️ The somewhat ambiguous Ads Personalization setting.

🥝 How the new-returning-open structure that helps maximize scale on ASA.

🧱 What’s coming with iOS 15.

⚙️ Are budgets shifting to ASA after ATT enforcement?

KEY QUOTES

How Apple defines tracking

In Apple’s vocabulary, we’re not tracking these users, we’re measuring the performance of a campaign, because what Apple defines as tracking is mixing information from your company with another company. So, if a campaign comes from the Facebook app, or the YouTube app, or another gaming app, or whatever app that has some ads in it, you’re tracking when you’re measuring between apps. And here, what’s very peculiar with Apple Search Ads is you’re measuring with an OS app, with this operating system app. This is not considered by Apple as tracking, because you’re not crossing data with any other app.

The mechanics of attribution under the Ad Services framework

When a new user goes to open the app for the first time, the measurement framework basically pings an Apple server and says, “oh, I’ve got this new user, does he come from one of my ASA campaigns or not?” So, sometimes the payload will come back empty, like, “nope, didn’t come from any of your campaigns”. Sometimes the payload will come with a lot of details about oh yeah, it was that campaign and blah, blah, and blah, blah. And this payload is actually pretty granular in the sense that you can get keyword level data, plus all the creative performance. From that moment you can assign, like, “that user who just opened the app comes from that campaign”

The ideal ad group structure within campaigns

So once you decide, maybe you work on ad groups, so you’ve got high LTV keyword, a medium LTV keyword and a low LTV keyword, maybe, and these are your three ad groups. What I would have recommended is to have each of these three ad groups move into three, I call them concepts. Like, each keyword block would be moved into three ad groups that have the same keyword within, but different targeting based on is this user: a new user LAT-off, returning user LAT-off or anyone.

Where the search tab placement could actually work

I’ve seen a few advertisers that told me like, “Oh, yeah, it looks dreadful. But actually, on some segment of returning users, it’s either very important for us strategically, or it actually brings decent metrics.” And that shows to what extent returning users can really vary from new users. And if you would never have split, between new and returning in this way, you would never have seen that maybe in the search tab that works.

The cost of not targeting the ‘open’ group

If you only target new and returning, you need 18 Plus, you’re going to leave on the table all these users: ads Personalization, underage, and the bizarre other edge cases like education device.

Why ASA is hard to scale

A lot of people were already maxing out what they can get on Search Ads, and it’s a search inventory. It’s not like a feed where you can put the ads to anyone, if somebody is not looking for something relevant for you. Even if you managed to show your ads to them, it’s not going to produce much of an impact. And so, I never thought it was really possible to shift your Google and Facebook budget onto Apple. And, what I’ve seen over the last two months or three months since SKAd Network is ramping up is that actually the inventory prices haven’t changed much. And I think this is the signal that this budget shift didn’t really happen.

The difference between Google search and Apple search

When you look at a search on google.com or google dot, whatever, the searches are a lot more, longtail. And there’s a lot more of it. So actually, I would even say, it’s easier to scale up your spend on Google Search alone. And then you’ve got display and YouTube and whatnot. Well, a number of search ads, you don’t have this extra inventory. But even on the search itself, you’re already more limited by the nature of the inventory. And the fact that it’s just this pool of people looking. The search behavior is really different in an app store than it is on a browser. And that limits a lot.

FULL TRANSCRIPT BELOW

Shamanth : I’m very excited to welcome and welcome back Thomas Petit. Thomas, welcome back to the show.

Thomas: I’m very glad to be here. I’m a very regular listener of the show. And I have to say, you push the bar very high. I hope I can keep that up.

Shamanth: Your past episode was one of the most popular ones we’ve had. And as a mutual friend of ours said, the thermometer exploded when you showed up. No pressure this time!

You know, we’re going to talk about what could be a logical extension of the last thing we talked about. The last time we spoke, you talked about LAT, and how that was changing things. In another episode, you spoke about what’s changing with iOS 14.5. We’ve had you twice, because this feels like a very logical extension of those episodes.

Let’s start with the high level. What’s changing in terms of attribution in Apple search?

Thomas: It’s an interesting question. The LAT cohort was always very peculiar with Apple search ads, just because they deal with it completely differently to the other networks. I was happy to discuss some of the most detailed intricacies on the show. But, I actually had a few other contents and explanations around what it is, how to deal with it, measurement, which, to my surprise, remained fairly valid for like, three years or four years or so. From Apple Search Ads opening in 2016, until this year. You can listen to the 2017 content around that, and it would be valid up to three months ago, more or less. Which was surprising.

If you do listen to the content I published back then on Apple such as on LAT, actually, it is different now. And that’s why we’re here today. On a high level, what happened before is, Apple decided that the ads were going to be shown to people who had the IDFA as zero, LAT people, limited ad tracking people. But those would not be measured by the attribution framework that Apple prepared in the past. So we had part of our traffic and the impact of the money we spent going into organics. And that was a big topic for a long time.

Now that Apple changed the rules of the game for everybody regarding IDFA, and that we need consent to get the IDFA, the way this framework was designed was going to break. And so Apple designed a new measurement framework that is called Ad Services that was announced in January this year, and eventually started to be rolled out somewhere between February and April, quite silently actually. All MMPs adjusted. And for a lot of people, it was almost silent, just a little tweak.

The big change here is that attribution of Apple search ads doesn’t rely on the IDFA anymore, at all. Which is a big change compared to what we had before. But actually in a good way, which is, for the first time we can measure the impact of people who were LAT before and who are now IDFA zero no matter what. So, that’s the first positive.

The second positive being Apple attribution doesn’t depend on the IDFA. But it doesn’t depend on your ATT status. This is a change from before but it’s also a change compared to other channels. That means Apple Search Ads, ASA, needs to be handled slightly differently.

There were other changes on the platform, there was a new format introduced with search tabs, China is open for some advertisers, changing how people look at the channel, but those were all pretty marginal. And the big change is the fact that the way we measure Apple Search Ads has changed in the last three months. Here, for the better. For most channels, the changes of ATT are kind of a nightmare. But here it’s actually for the better. So, good for those who work with that channel.

Shamanth: In the past LAT users were not trackable just because IDFA was the main mechanism. And now, from what you’re saying, everybody can be tracked. Deterministically as you will.

Thomas: Yeah, when you say everybody can be tracked deterministically, I would say this is correct, but it needs a bit of wording specification. When we say deterministically, we say user level. So we can know each user and not in aggregate. And we can follow retention metrics, LTV, we can know if that person has made an IAP purchase after three months, or if they renew a subscription, or whatnot. Because we’ve got user level information, and that’s what deterministic means here. But I think here the confusion for some people might come from the word tracked. From a marketer’s perspective, yeah, that’s what we do. We track users. That means we measure that these people come from this campaign.

You’re just crossing data with the system. So yes, for us, marketers, we’re tracking this deterministically. For Apple, we’re only measuring first party data. It can lead to some confusion, because what Apple calls tracking is not exactly what most marketers would call tracking. But the answer is yes. We can measure the impact of campaigns on user level data without ATT consent, which is a big difference from all other networks out there.

Shamanth: The way Apple defines privacy related terms, is very, very strategic. Right? They say, Oh, this is basically first party marketing. We are showing an ad on our device. For installing an ad on our device, it’s not tracking.

Thomas: I was not giving too much opinion on how legit this definition is, or how fair it is, and what I would define as tracking. It’s clear this is how Apple defined the rules of the game. As marketers we’re playing within those rules. Whether we agree or not, we have to know them, and we have to know how to deal with them.

Shamanth: Certainly.

Can you talk about the mechanics of how user level data measurement or tracking happens without the IDFA? And the mechanics of exactly how ATT opt-in and opt-out users are treated by ASA just now.

Thomas: I’m going to answer you in a trivial way, which is not either technical or exactly how it happened. But it makes it easy to understand. Some users are going to see some ad impressions from search ads, when they search in the store. They search for fitness, they search for Tinder, they search for Tik tok, whatever, they’re going to see an ad. And maybe they’re going to install the app or not. On your side, regardless if you paid for those ad impressions — and regardless if this ad impression had some consequences —

when a new user goes to open the app for the first time, the measurement framework basically pings an Apple server and says, “oh, I’ve got this new user, does he come from one of my ASA campaigns or not?” So, sometimes the payload will come back empty, like, “nope, didn’t come from any of your campaigns”. Sometimes the payload will come with a lot of details about oh yeah, it was that campaign and blah, blah, and blah, blah. And this payload is actually pretty granular in the sense that you can get keyword level data, plus all the creative performance. From that moment you can assign, like, “that user who just opened the app comes from that campaign”.

And then you can technically, and you can legally, follow whatever he does after, because this is first party. That’s his retention metrics. Do they buy stuff? You don’t have a limitation of 24 hours or anything like that.

So basically it’s a ping at first open and you get the payload: “is this from Search Ads, which campaign, which ad group, which keyword” blah, blah, blah.

Technically, there is a small difference if a user has opted in to ATT. And you have this option early enough, which is another tricky bit, but Apple gives you the exact time stamp for the click, only for the opted-in cohort. Here, I believe for 99% of people, this has zero implication whatsoever. Like, okay, this guy comes from that campaign, maybe the click was five minutes ago, maybe the click was yesterday, it’s actually very, very minor. So there is what they call the advanced payload, which is when there is opt-in, and the basic payload, which is when there is no opt-in. But the information that comes from the basic payload, and what you can do with it in terms of LTV and cohort analysis, is actually more than enough, more than you need for most of the applications. Maybe some people will use the timestamp information for some major mix models and things like that. I personally think if you are at that level, maybe you should be a guest of the Mobile User Acquisition Show.

Shamanth: I think the mechanics make sense. So, if the Apple server pings you and says, “Yes, this is your user”, you tie it to your internal ID rather than the IDFA.

Thomas: Exactly. At the moment you receive this information, like, “Okay, oh, that user will just come from campaign x and keyword y and blah, blah”, you have internal tools that do that automatically without you doing it. But maybe you have your own system and database and analytics, you basically assign an internal user ID something that only you can use. If you share this with another app, or with any vendor, it’s pretty much useless, because it’s just yours. The same way analytics tools work. Of course, you don’t mix sources of new users, they assign you an internal user ID that is only for you. And this is exactly how it works here for Apple Search Ads., It’s like, oh, I recognize this user and I can follow whatever they do after. And that’s why we say it’s a user level and not aggregated or promoted. You know exactly who this user is so you can follow whatever they do inside your app. It’s not like you can track them on the internet and know that they buy some shoes on that ecommerce. But at least for you, you know, if they reopen the app, if they opt-in the push notification, if they purchase, if they retain, if they invite friends, blah, blah, blah.

Shamanth: And even for an MMP, they would get some sort of user ID. Which we don’t always see if we’re using a MMP interface. You don’t always see that in this one. But if you have a BI system, you will probably have an ID for that.

Thomas: Yeah, exactly. The MMP would do the exact same as analytics tools. They assign some kind of user ID internally to that user, and then you can use that. Which, actually, brings an interesting detail about this new framework, Ad Services, which is, in the past, you could implement the old framework by yourself or through an MMP. A lot of apps have an MMP for other reasons so it was just simpler to go through the MMP. In this new era, with Ad Services, the same thing remains true. You can rely on your MMP to do this work. And, if you have an MMP, I mean, why not? You know, it’s there and just use it. Let them do the job, the job you’re paying them for. And that works just great. I didn’t have any problem with any of the major MMPs regarding Ad Services work. Yeah, I can’t remember any huge issues with that.

You can also use this on your own. I would say this is mostly either for people who are too small or early to have an MMP, or for another reason. But, maybe they’re starting with Apple Search Ads. You can implement this framework on your own without an MMP. Some people might want to implement it by themselves, because they love to use internal tools, and maybe for the long term, you don’t rely on anyone if you want to switch MMP or whatever. But there’s no big benefit or big drawbacks of doing it yourself. And from what I’ve seen, most people just do the plug and play integration through the MMP. It just does make a lot of sense. And there’s nothing better in doing yourself in that particular case.

Shamanth: And in the past, we factored in LAT and we had an entire episode where you spoke about that: exactly how you factor in LAT in your campaign strategy. Now post ATT, what changes?

Thomas: Before, these users who couldn’t be measured, obviously, a lot of people assign a different value to it. Either because they were extrapolating and the uncertainty is higher, but also because a lot of advertisers thought, “Oh, if this is not going to be measured, I’m not going to target them. And I’m just going to go for LAT off only.” Which has pushed the price of the LAT-off inventory up. And as a reaction, many advertisers had their campaigns structured on, “Oh, if those are LAT off users, I’m probably going to bid slightly differently, because I get different information. But also, because the inventory price is different. So I need to adjust to that.” Now, this factor seems to disappear, because we can measure users whether they’re LAT or not. Actually, there’s no more LAT. So that it’s become a little bit irrelevant. But at the same time, a lot of people haven’t changed the way their campaigns are run, because this is fairly new. And because you still have tricky bits of targeting between knowing if a user is a new user or a returning user. Most people will have kept their old campaign structure. And this is actually what I do. And what I recommend for many cases, maybe not all cases, but in most cases. I don’t think it had a huge implication. It’s kind of curious that something that has established itself a little bit as standard, because there was this LAT thingie, actually remains fairly valid in a post-LAT era. Inventory price difference is still a little bit there. And I mean, if your company already structured things that way, you can restructure, but it’s not going to be a game changer for sure. So a lot of people just kept the way they were doing before. This is what I’m doing personally.

Shamanth: Can you refresh people’s memories? Or if people are not familiar with it, can you tell folks about what the campaign structure that you had recommended in the past was? And why was it structured in that way? Why, and what changes? If at all?

Thomas: So, there are three levels to the structure. The first level is at campaign level, due to how you can set up and put the settings in your campaign. And this is pretty much the same before and after. What I’ve recommended, what Apple has recommended, and what most agencies out there, or even internal teams have recommended, is actually pretty much the same. Usually, split campaigns by, Is it brand terms or not? Is it competitors’? Is it generic? Or is it discovery? So, you more or less have four campaign types. I know people who merge discovery with generic, it’s kind of not critical. But putting brand aside is very critical. But that’s another one. So basically, you have four types of campaign because the performance is slightly different, strategically, and so on.

And then some people decide to bundle some countries to have less of a burden. Some people decide to have one of each of these four for each country, which is around 60 countries. So that would be a maximum of 240 campaigns. And I guess that would be the most developed structure. And this is pretty much the same as it was before. And it’s pretty standard, more or less everybody does that. The only choice is are you going to build on more countries or not, depending on how you organize internally.

When you move down to the ad group level. Here, I think there’s one thing that can be done in very different ways. Some people like to group keywords by their meaning. Some people like to group keywords by their LTV, some people like to group keywords by “Am I actually spending or not?” So maybe a high spending keyword would have its own ad group on its own. And then you’ve got an ad group with a long tail of keywords, because you’ve bundled them all. I’ve talked to a bunch of people working with one and the other method. And I’ve seen various things again on that. I don’t believe there’s one that’s better than the other for performance. So it’s mostly how you like to get organized yourself, and what the person managing the channel is comfortable with for measurement. But there are various ways to do that. And also on that level, I think nothing’s changed just the same as before.

But then, the last part is something that I believe is changing a little bit for some. Which I’ve often recommended.

You don’t really get a clean split, because as soon as you don’t put any kind of age or gender criteria. Usually we use 18-plus to split LAT off, and not because Apple doesn’t want to share the age of users who are, were, LAT on. So basically, we end up with three groups.

The open group, the one with zero targeting, does overlap with the other two, which sometimes could get messy. And in a way, I would understand if some people say, “Oh, I don’t need to go through all of that anymore, I can just bundle them all because there’s no measuring, like, measuring issue anymore with this LAT”. And that would make sense. And I would simplify the structure a little bit.

Personally, I like to keep this separation, the split between these three groups. Because I’ve seen the conversion or retention performance of new users and returning users being entirely different. Also, a lot of companies would have a budget that is exclusively for acquisition and some for returning users not exactly remarketing. But let’s say they often tag it this way, because it’s simpler in concept. And me, I like to separate it because I often see different performance. So I don’t like to bundle it too much. But we have to admit that this separation is not entirely clean, because Apple made the distinction.

You can’t have a clean ad group, like the open ad group would overlap with the other two all the time. If you’re a small advertiser, if you don’t want to go into this kind of detail, it’s totally fine to have them all together. Personally, I don’t know, I like to have my things split. And also it helps when you’ve got budget changes, like, “I’m just gonna keep the returning user or just the new users”.

And, actually, I know I make long answers. But that’s something that we saw in the new format, like the search tab format, that Apple released about two, three months ago. For almost, for pretty much everybody, I was told the performance of these new campaigns were really dreadful. And I’ve seen a lot of people pull back and just stop them entirely, because they didn’t make sense.

But I’ve seen a few advertisers that told me like, “Oh, yeah, it looks dreadful. But actually, on some segment of returning users, it’s either very important for us strategically, or it actually brings decent metrics.” And that shows to what extent returning users can really vary from new users. And if you would never have split, between new and returning in this way, you would never have seen that maybe in the search tab that works.

And I think this kind of logic can apply outside of search time for other campaigns. And for me, that’s the benefit of splitting them out. Having a little bit better of a view on each, even though you know, it’s not 100%, because the open group is always going to mess a little bit with your data.

Shamanth: Sure. And considering how different the new and returning users will be, it certainly makes sense to separate them out.

Can you help us understand why we have the open group at all, because you have new users, you have returning users? Some would say that covers your entire universe of users.

Thomas: Yes, on paper. Before there was LAT. So if you only get the new and the returning one, basically you were leaving off maybe around 40% of the inventory. It depends on the country, and the vertical, sometimes 20, sometimes 50. But basically, a third of the inventory, you wouldn’t capture if you only add new and returning. So that’s why before we used to have the open group. There’s no more LAT, but Apple still labels some users differently. That’s the case for people who are under 18. Apple said we don’t want to share if it’s the new or returning user. Actually, we don’t want to even know ourselves, which I just covered on the tricky bit. That is actually not true. Apple needs to know if they’re new or returning because they’re reported elsewhere. So it’s kind of weird, but anyway. That’s the kind for underage users. There are some kinds of blocked users because of things like education device or some bizarre criteria that they add. But there is also this new option, which Apple called Apple Ads Personalization, that a small portion of users would go and deactivate, which only applies to Apple ads. But in this case, Apple doesn’t want to share at all if those users are new or returning.

And so if you only target new and returning, you need 18 Plus, you’re going to leave on the table all these users: ads Personalization, underage, and the bizarre other edge cases like education device.

Shamanth: Yeah, yeah. Yeah. So in other words, Ads Personalization is the new LAT.

Thomas: In a way, yes. Technically, they are defined differently, but for the manner of targeting on search ads. Yes, this is pretty much the equivalent. Yeah.

Shamanth: For the purpose of campaign construction and strategy, it sounds like that is the case.

Thomas: Absolutely right.

Shamanth: And it also sounds like the ads personalization numbers aren’t a lot but it is still nonzero. Or is that right?

Thomas: Here, that’s a very tricky question. Because the assumption is that only a tiny minority of users are actually going to go in this setting, that is quite buried down in Settings on iPhones. And I would believe it if it was clean. And only people who decide to opt out, it would be a tiny, tiny percentage. But the tricky bit is one, nobody knows about this percentage, like there was never any communication about it. And you can’t actually follow it precisely. Because the same way before we had the information LAT, not LAT, in the Apple interface, you don’t get this information about Ads Personalization. So we don’t really know the percentage. The other one is, on paper, it said that users that were on LAT before got moved automatically when they upgraded to iOS 14.5 to Ads Personalization off. But I have no way to prove that this was true in 100% of the cases or not. We know by edge cases that there was bizarre stuff happening there, like people who were LAT off and suddenly had this option on by default, and some people had the option grayed out. That it’s not exactly the same, but from far away, we might be able to assume that this percentage could be close to the old LAT one. Basically, because Apple ported the settings out from the older one. So, we’re only three months after ATT. There’s not so many new devices. I guess over time, when Apple releases new iPhones, this percentage might go down a little bit because on a new device, you’re not going to be asked for LAT or not. So it’s possible, the LAT percentage was always going up. It’s a bit of a guess here, really. But I have the opinion that the Ads Personalization percentage might actually go down a little bit as people get new devices. But let’s see if it’s gonna happen. And we can’t follow it very closely. So I couldn’t even tell you if it’s 2%, 40%. I have a rough idea that can fall away. Let’s say it’s close to the LAT percentage we had before.

Shamanth: And from what you’re saying, having new, returning, open lets you cover all your bases. So that’s the safest way, too.

Thomas: Well, safest, I don’t know, it’s more like if you’re trying to max out your spend on search ads, because you want to scale it. Or if, for some strategic reason, you want to have a very high share of voice. You want to defend your brand, for example. Limiting to new and returning could actually lower your potential reach. But for some people who are extremely focused on performance only, don’t have this perspective. Or maybe, if the performance is actually really bad, you might want to only target the percentage of people just because this is the only way you get performance. That depends on the case, actually. I’ve seen both cases of people who really want to go for everyone. And people were like, “Oh, no, the performance is bad there. Let’s go only for new users.” In some other cases only for returning users I can’t tell you straight up which is going to be the case, because I’ve seen that different apps have different behavior on that.

Shamanth: Just to switch gears a bit. I’d love to hear your perspective on what’s happening going forward, especially since iOS 15 was announced. In terms of being able to do testing with ASA in the future: there are a lot of changes coming. Can you speak to what’s changing?

Thomas: Yeah, that’s actually very exciting. And a bit of a hot topic this summer, when you don’t talk about IDFA and SKAd network, a lot of people, especially in the ASO world are talking about these new tools that are coming. I have been working with the AppStore maybe for eight years or something. And we’ve been waiting for an A/B test tool forever and it’s coming soon. So everybody’s super excited. Here, I guess a lot of people that were sometimes using creative sets for ASO or AB testing purposes are going to switch to this new tool. Creative sets were never great for this. There was a lot of bias introduced, and a lot of factors you didn’t control, and limitations. So actually, this is great news.

It doesn’t impact ASA directly, but creative sets will be less used for this because there will be a proper tool. But there is another feature that Apple has promised us that is super exciting. Which are custom product pages, which basically you can show different ASO listings, a different appearance, campaign by campaign. And here, there is really a potential of improving ASA performance by showing something more specific.

If it’s a new or returning user, you might show them different stuff like more advanced features to a returning user, for example. Or, a lot of apps have USPs that capture different, let’s say, intent. I work in the mental health space with apps that target meditation, and yoga, and stress and anxiety, and self improvement. And they will be able, by which keyword people are typing, to actually provide an experience that is custom to that.

So that’s actually super exciting. It’s not live yet. First, because Apple will only give us this with iOS 15, in full this year. But also because we’re waiting at this point. So this feature works on a custom URL, you need to say, “oh, if the user types this URL, then show that version.” And at the moment, we can’t really input any URL in Apple Search Ads. So we’re waiting not only for iOS 15, but also for an update from Apple in their own ads interface to open up this possibility. What I’m hearing, which is just a rumor, is that this is not going to be ready for iOS 15 launch. So, I obviously don’t have dates, and Apple typically doesn’t communicate on future releases. But maybe we’ll have to wait for three, six months for this. But it’s coming. The same thing applies to Google and Facebook. And I think this is going to be a very powerful feature. We’ve been waiting for it for years. And it really benefits both the user and the developer. The user is going to see something more relevant. And the developer will be able to capture his marketing message and his USP to what’s most relevant to the user. So I really think this feature is a win-win situation here.

Shamanth: Yeah, definitely. Certainly something we’re excited about. But also curious about, because it’s not very clear as to the mechanics of how this will work in the interface.

Thomas: But we already know there will be limitations here and there. Yeah. And there’s a number of it, and we don’t know when and yeah, but it’s still exciting.

Shamanth: It’s gonna be the whole privacy threshold that they’re going to throw at us.

Thomas: Yeah, I then thought about that, but I’m not sure it’s gonna fly Apple Search Ads, though. But let’s see.

Shamanth: Well, if you have a custom product page, it’s my understanding that you can take any URL and send traffic to it.

Thomas: Yes.

Shamanth: And if they let you take any URL, presumably you can use it for Google and Facebook. And if you do that, they will throw the privacy threshold at us. That’s one of the things I worry about.

Thomas: Yeah, but I guess this information you’re going to see on this custom page, you’re only going to have it in App Store analytics. It doesn’t really enable you to follow the cohort and so on. So that’s kind of a brand new thing. But yeah, like you said, there’s still a bit of uncertainty because nobody has seen it. We’ve seen screenshots and description is not the same as having all the options. It’s just, I think it’s a big leap forward. So even though it has limitations in privacy in privacy threshold, and other limitations, I heard it’s like 35 pages maximum or something like this. So there is a maximum, but it’s very exciting, I’m really looking forward to this one. Something new that is very relevant. It’s not like a site feature on the site, like I don’t know, recommendations or stuff like that, it’s a real new feature.

Shamanth: Exactly. Exactly. Exactly. Yep.

Just to switch gears to another topic that is very top of mind for a lot of marketers. Since SKAdnetwork has broken a lot of tracking in a lot of conventional channels, a lot of folks have been like, “oh, budgets are going to shift to ASA.” A lot of folks have speculated that Apple did this so they could capture budgets, so they could grow their services category. I think you’ve probably mentioned that too. What have you seen happen? Have you seen that budget shift to ASA? Do you think that’s likely? Has it happened already?

Thomas: That’s an interesting question. I’ve read a few, usually mainstream media, comments on this, before SKAdNetwork released. Like, “Oh, yeah, people are going to switch more of their budget to Apple ads, and they’re going to eat the Google and Facebook ad budget and they’re going to take it to Apple and Apple is doing this to gain market share.” I guess the logic behind it is valid. But if you look at the nature of Apple’s ads as a channel, this was never going to be able to happen. And it’s true that on one side, a lot of people get relatively good performance from search ads. And now after high uncertainty, or in most cases, horrible results from SKAd Network campaigns, it would make sense to think like, “Oh, yeah, let’s shift budget there”. The thing is that in practice, this is very rarely possible.

There was clearly a shift of budget towards Android, and more and more people are this summer, at least. Let’s see in the long term, and I’m not sure in the long term, this is gonna hold. At least for this quarter, the inventory pressure on Android is really high. I’ve read a couple of people saying Android is about 30% more expensive, the ad inventory since SKAdNetwork released. I don’t think this is true at all, for Apple ads. I would say actually, things have been pretty flat. And I put that on the fact that with a search channel, you can’t you can’t just push a button and get more of it. So unless your account was not really optimized for a big number of keywords, or all the countries, or targeting limitations like this, only new and returning and you can expand a little bit. But it’s not like from one day to another you can move from 10k to 100k of spend. That’s just never gonna happen on a search channel. It’s just the nature of it.

Shamanth: Right. And, even if you compare it to Google search on the web, it’s very different. People who want to scale Google search, they go to display and YouTube. Right? And it’s also not as good.

Thomas: As long as Apple doesn’t add inventory that is not search, of course, there will be this limitation. But I would say even when you look at Google search on the web, there’s a lot more long tail keywords. And we’re not talking about the same kind of volume either. There’s like 500 million visits in the App Store every week or something like this. But not everybody goes to the search. And when people go to the search spots, it’s very short tail actually, in the App Store.

Often you can scale a Google search budget a bit if you work deep enough on long tail and all kinds of extensions and stuff. You can always scale it a little bit. On [Apple] Search Ads it’s like, no. Most people who I consult on search are not telling me I’ve got terrible performance, they’re telling me how can I get more of it? And in some cases, just like you’re at 90%, 95% of what you can reach. We can optimize marginally. But there’s no way we’re gonna tweak the channel.

Shamanth: It’s definitely not the easiest channel to scale because there is a ceiling to it. That’s not going to change unless Apple introduces a non search placement. And like you said, the search tab was not very successful. We’ll see how that shakes out. Thomas, this has been incredibly comprehensive, as always, no surprises there. But this is perhaps a good place for us to start to wrap up. Thank you, again for being a guest on the Mobile User Acquisition Show. But before we let you carry on, could you tell folks how they can find out more about you and everything you do?

Thomas: I’m pretty visible publicly. It’s easy to find me on Twitter or LinkedIn. And I’m always happy to exchange on a few slack groups, open slacks where I’m participating, some of them being the ASO slack, or Mobile Dev Memo, or Mobile App Privacy. There’s a few of them on there. If you want to directly contact me, Twitter is fine, LinkedIn is fine. You can find my email online as well. And this would be your best place to find me.

Shamanth: Wonderful, we will link to all of that. And should we mention your email newsletter as well?

Thomas: I do have a newsletter. I didn’t mention it because I haven’t sent any episodes this summer. Because I was taking a bit of time off. And I was also looking at how the dust was gonna settle a bit on the SKAdNetwork. But it’s coming back this fall!

Shamanth: Wonderful. So we will send people there, too. madv.io. We will link to that. And everywhere else you are.

Thomas: We will be back, it will be back. I was waiting for sponsors, but they’re not there. they’re not kidding, actually. All right.

Shamanth: Okay. Cool. All right. Thomas, we will let you carry on. Thank you so much for being on the Mobile User Acquisition Show.

Thomas: Thanks for inviting me! And looking forward to the next show as well. So yeah, congrats for keeping up for keeping the quality up.

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!