How does SKAdNetwork 4.0 work? Q&A with Tenjin’s product manager

SKAdNetwork 4.0 is here, and it stands out from its predecessors. For the first time, Apple has released a new version of the SKAdNetwork API with not one, not two, but several new features. And they did so without releasing a beta version for testing first.

We go into more detail on the subject in the Q&A with Tenjin’s product manager, Jaspreet Bassan, below.

Here are some of the crucial SKAdNetwork 4.0-related questions Jaspreet answers:

1. What surprised you the most about this SKAdNetwork update?

1. If I am an advertiser, what should I do now? What should I be paying attention to?

1. How do you think the industry is reacting to SKAdNetwork 4.0?

…and more.

Question: What surprised you the most about this SKAdNetwork update?

Answer: For me the most surprising thing was that a big update like SKAdNetwork 4.0 was released on time, and without a beta for testing. Typically, Apple has always added one valuable new feature in each update. In SKAdNetwork 2.2, this was fidelity type, in SKAdNetwork 3.0 it was ad networks getting a copy of non-winning postbacks, which was a did win parameter, and finally in iOS 15 developers were able to get a copy of the winning postbacks.

But none of these previous updates came with a variety of new features. So, it was surprising to learn that a) SKAN 4.0 has a lot of new features and b) that Apple released SKAN 4.0 with iOS 16.1 on time, and without a beta for testing.

Question: If I am an advertiser, what should I do now? What should I be paying attention to?

Answer: It might be a little too early for advertisers to start paying attention to SKAdNetwork 4.0. At the moment, I believe advertisers can continue focusing on SKAdNetwork 3.0 while MMP’s and ad networks build SKAdNetwork 4.0 out.

But when the industry starts adopting SKAdNetwork 4.0, I think as an advertiser I would be excited about getting additional information on my installs, but apprehensive at the same time. I would be apprehensive because the industry was just catching up to SKAdNetwork 3.0 and how to use it to optimize campaigns. But now, we have to learn new things to be able to optimize campaigns with SKAdNetwork 4.0 data.

Question: Can you give an overview of what has changed?Answer: There are a lot of new features in SKAdNetwork 4.0, and the main goal of this update is to send advertisers and ad networks more information about their installs while preserving their privacy. Let’s look at the following image that provides an overview of the crowd anonymity tiers to answer this question. It is important to understand the concept of crowd anonymity before diving into the new SKAdNetwork 4.0 features.

So, Apple has defined 4 tiers under crowd anonymity, tier 0, tier 1, tier 2, and tier 3. The concept is that an app is considered to be in tier 0 if it has a few (or the lowest) number of installs. Alternatively, an app is considered to be in tier 3 if it has a lot of installs. And based on what tier an app falls under, Apple will send different kinds of data.

Question: Why don’t we start with the conversion values that we get back?

Answer: We have two conversion values now; fine-grained and coarse-grained conversion values. Fine-grained conversion values are the same as before. They are the 64 bit conversion values, and they remain unchanged when it comes to the crowd anonymity aspect and the random delay (which is 24 to 48 hours) that is used to send the value is also the same. However, we can now decrease the conversion value, whereas before we were only allowed to increase it. And the 24 hour timer, where advertisers used to get 24 hours to update the conversion value, is now gone.

Now let’s talk about coarse-grained conversion values. Advertisers can get these conversion values with the possibility of low, medium and high. Let’s look at the SKAdNetwork 4.0 image below to understand the coarse-grained and the fine-grained conversion values better, including when, and at what tiers we receive them.

From the image above, we can see that we receive a coarse-grained conversion value when we are in tier 1, and we receive a fine-grained conversion value when the app is in tiers 2 or 3. Before this, we did not receive any conversion value in tier 1. This is a new addition by Apple. Now we can get a coarse-grained conversion value in tier 1 when we have a low number of installs while preserving user privacy.

Question: How are coarse-grained conversion values even useful?

Answer: Before SKAdNetwork 4.0, apps had to wait to get into tier 2 to get any kind of information about their installs. But now with coarse-grained conversion values, apps can get some information about their installs much faster. They can get this information in tier 1 much faster as they don’t have to wait for their app to get into tier 2.

Question: How about source IDs, how are source IDs different from campaign IDs?

Answer: Source IDs are also hierarchical in nature. Let’s look at the following image from WWDC to understand this:

The hierarchy of source IDs is represented as the last 2 digits, the last 3 digits, and the last 4 digits. The image above provides an example of how advertisers and ad networks can encode these hierarchies however they want. The last 2 digits might be the same as the campaign ID in the previous versions, the last 3 digits can also represent location and the last 4 digits (or the first digit) represents placement. It is always up to the ad network or the advertiser to encode this however they want. And we believe that mostly it will be the ad networks that will encode this value.

Let’s go back to the SKAdNetwork 4.0 image below to see when we receive source IDs:

So when the app is in tier 0 and tier 1, we receive only the last two digits of the source ID. But as the app goes into tiers 2 and 3, that is when the number of installs increases, and the app can now receive the last two digits or all of the digits of the source ID – depending on how it is defined by the ad network.

Question: So you have gone over the different kinds of conversion values, and given us the distinguishing factors between source IDs and campaign IDs. Can you also cover the different postbacks that are shown in the table? What do these 3 postbacks really mean?

Answer: This is another big addition to SKAdNetwork 4.0. We can now receive 3 postbacks instead of 1. And these 3 postbacks are meant to measure different periods of time post install.

Let’s look at the image again below

We can see that postback 2 ends at day 7 after-install and postback 3 ends at day 35 after-install. We can lock these two postbacks, and as soon as they’re locked and that time window ends, Apple will send this postback with a random delay of 1 to 6 days. Additionally, Postback 1 ends at day 2 after install and we receive it with a random delay of 1 to 2 days.

Question: So, what are the advantages of postbacks 2 and 3?

Answer: Apps should be able to measure re-engagement now.

Question: And who is going to receive postbacks 2 and 3?

Answer: Only the developer and the winning ad network.

Finally, Apple has also added web to app attribution to SKAdNetwork 4.0. If we look at the image again, we see that when the app is at tier 3, Apple will send either the source app ID or the source domain if we’re using web to app attribution.

Question: How does Tenjin feel about the SKAdNetwork 4.0 changes?

Answer: I think we have mixed feelings about it. While SKAdNetwork 4.0 looks way more promising than SKAdNetwork 3.0 for advertisers, it is hard to tell if the industry can build a complete solution that accurately estimates ROI or ROAS – which would increase SKAdNetwork’s adoption. But looking at how SKAN has evolved over time, I think we are going towards a complete solution and since preserving privacy is very important to us, I feel optimistic about it. Given the complexity of SKAdNetwork 4.0 features, I think we will play a pivotal role in conveying these features in the most appropriate way that helps our customers make the right decisions. Not just us, but all MMPs. So the role of MMPs is only getting more crucial.

Question: What should Tenjin customers expect in terms of SKAdNetwork 4.0 related features?

Answer: We are working towards bringing SKAdNetwork 4.0 features in our reporting suite. We have been communicating with our partners to synchronize implementation and launch. We will continue to support older versions, so backwards compatibility is going to be important when we launch SKAdNetwork 4.0. Web to app attribution is out of scope right now but this is something that we’ve been brainstorming for some time now, so more details on it are to be discussed. I think our customers can expect something in the first quarter of 2023. But it’s a loose estimate so no promises there.

Question: Could you dive a little deeper into your SKAdNetwork solution and maybe talk a bit more about IDFA and SKAdNetwork?

Answer: SKAdNetwork does not differentiate between opt-in and opt-out users. We have always provided separate reporting for SKAdNetwork while keeping transparency in mind.

Question: What do you mean by transparency?

Answer: By transparency I mean we provide very detailed reporting using very different metrics. We’ve always prioritized providing transparency to our customers by providing them with as much information as possible for them to be able to make the most optimal decisions. We have different install metrics such as reported installs, tracked installs, SKAdNetwork installs, and installs from ad networks.

But sometimes this transparency can look like noise to our not-so-mature customers, who are just looking for a more guided setup.

While we do offer a SKAdNetwork dashboard today, we do see a need to offer a complete solution for iOS reporting in the near future that makes SKAdNetwork more readable, in a sense, and something that helps with correctly estimating ROI by keeping transparency in mind but at the same time presenting transparency as actionable information rather than confusion or noise.

We also want to make the conversion value setup process more intuitive for our customers. We have some customers who still don’t understand how to set conversion values, and we are looking to make this process more seamless between them and the ad networks.

Question: Could you talk a bit more about how your partners are planning to adopt SKAdNetwork 4.0?

Answer: We’ve been communicating with some of the major partners regarding SKAdNetwork 4.0 and they are still in the planning phase trying to get their ducks in a row, I think all of us are. But given our past experiences we are expecting our partners to react differently to SKAdNetwork 4.0, and we’d have to build something different to accommodate all the nuances, but we will do our best to build something that is standardized for our customers.

Question: When should we see the changes take full effect?

Answer: Good question. Let’s try to answer this while examining the conditions that must be met in order to get a SKAdNetwork 4.0 postback:

Condition #1: The ad network needs to generate an ad signature for version 4.0

Condition #2: For ads that appear in an app, the app needs to be built with the iOS 16.1 SDK or later. For web ads, the ad appears in Safari 16.1 or later.

Condition #3: The advertised app is App Store-signed and should be running on a device with iOS 16.1 or later.

I think conditions 2 and 3 will most likely get adopted first and condition 1 will take the longest time. But “full effect” is a bit hard to define. Does it also mean that ad networks learn to use SKAdNetwork 4.0 data to optimize campaigns? If yes, then that could take up to a year or more given our past experience. But this is a loose estimate. Given that we get more signal now, which also means more signal to optimize campaigns, I think ad networks have an incentive to roll this out sooner rather than later.

Question: Do you have any thoughts on SKAdNetwork 4.0? Or how you think the industry is going to react to it (or is reacting to it right now)?

Answer: Personally, I am passionate about the privacy-centric way of advertising. I think SKAdNetwork could be the standardized way of doing ad attribution and reporting. But at the same time, we can still see that the market has not fully adopted SKAN yet. So we still have a long way to go. But I am excited that the industry is taking privacy more seriously and building solutions for a more privacy safe way of advertising. Doing something that not only helps our customers, but people in general is always great and excites a lot of us.