How conversion values work in SKAN 4.0

UPDATE - OCTOBER 2022

SKAdNetwork 4.0 was released on Oct 24, 2022 together with iOS 16.1. This article was updated according to Apple’s SKAN 4.0 release notes.

What’s up with SKAdNetwork 4.0?

At the beginning of June, Apple introduced many of the upcoming changes that would be available in SKAdNetwork 4.0—including new types of conversion values 😱. 

In this post, we answer the top questions about how conversion values work in SKAN 4.0, including:

1. How conversion values changed in SKAN 4.0?
2. What are coarse-grained conversion values?
3. What are fine-grained conversion values?
4. Will you get both coarse-grained and fine-grained conversion values in each postback?
5. Will you always see a conversion value in the postback?
6. What happened to SKAN’s infamous random timer?
7. Are these changes already available?
8. What are the next steps for app marketers looking to run SKAN 4.0 campaigns?
9. Are there other changes in SKAN 4.0 my team should be aware of?

 Feel free to skip ahead to any question you find relevant! 😉

‍

1. How are conversion values different in SKAN 4.0?

In the new SKAdNetwork version, conversion values are classified into two different types, coarse-grained conversion values and fine-grained conversion values, based on the number of downloads and the level of user privacy met by the campaign.

‍

2. What are fine-grained conversion values?

Fine-grained values are conversion values as we know them today, a 6-bit value allowing for up to 64 combinations (from 0 to 63) that can be set up to collect user behavior signals to optimize campaign performance.

3. What are coarse-grained conversion values?

Coarse-grained values are a new type of conversion values that can result in one of three different values, “low”, “medium” or “high”, according to the type of event completed by the user and the conversion value mapping of each advertiser. 

Note that the  method has been updated to reflect the introduction of coarse-grained conversion values.

4. Will I get both coarse-grained and fine-grained conversion values in each postback?

Not at all 👎. Even if both values are captured during the measurement window, advertisers get only one of these two values, either the coarse-grained value or the fine-grained value (and only when Apple's privacy threshold is met).

Note that in SKAN 4.0 multiple postbacks are available across different time windows, but marketers only get a fine-grained value in the first postback, and therefore, for the first attribution window (0-2 days).

‍

5. Will I always see a conversion value in the postback?

To decide whether you’ll receive a conversion value, Apple leverages crowd anonymity, defined as a system of four tiers that relies on counts of installs to “determine the level of privacy assured to the person using your app”. In their WWCD session, Apple states: “when the install count is low, we take extra steps to protect privacy by limiting the trackable information sent back. As the count scales up and the user’s uniqueness starts to blend into the crowd, we send more data back. Finally, as the count reaches the highest tier, we are able to send the most data back while still preserving privacy.”

In the image above, the stars symbolize users' uniqueness, showing how in Tier 0 it’s harder to preserve users’ privacy, while at the highest end, user uniqueness blends in with the crowd.  

Based on the four crowd anonymity tiers, there are three different scenarios you should expect when receiving your SKAN 4.0 postback:

1️⃣ You don’t receive a conversion value when the install count is low and therefore doesn’t meet the privacy threshold (Tier 0).

2️⃣ You receive a coarse-grained conversion value when the count of installs generated is high enough to meet the threshold but not enough to be on the higher end of the crowd anonymity spectrum. Data tier 1 might get this value in any of the three postbacks; data tiers 1 and 2 might get this value in the second or third postbacks. 

3️⃣ You receive a fine-grained conversion value if the count of installs generated by the campaign is high enough to access detailed attribution data without compromising crowd anonymity. You can only receive fine-grained conversion values in the first postback, and it’s only available for data tiers 2 and 3.

‍

6. What happened to SKAN’s infamous randomized timer?

Postback delivery times changed in SKAdNetwork 4.0: randomized timers are now shorter for the first postback and longer for postbacks number 1 and 2. Postback 1 corresponds to the first conversion window (0-2 days). After this window ends, the postback will be sent within the following 24-48 hours. For the second and third postbacks, the random delay is longer, taking between 24-144 hours.

The latest Apple documentation introduced a new change that allows advertisers to “lock” and send a conversion value before the conversion window ends. This method can be used on any or all conversion windows, and it can help advertisers get conversion values earlier. 

However, this doesn’t mean that when using Apple’s lockWindow parameter you will get the conversion value right away: the randomized timer will start after a conversion value is locked. This means that for postback 1, you will receive the postback 24-48 hours after the conversion value is locked, and for postbacks 2 and 3, you will get it after 24-144 hours.

Note that once a conversion is locked, the system will ignore any further user activity that happens in the remaining time of that same conversion window.

‍

7. Are these changes already available?

Yes! SKAdNetwork was released on Oct 24, 2022, so advertisers can already leverage these new changes.

‍

8. What are the next steps for app marketers looking to get ready for SKAN 4.0?

Work together with your DSP and MMP partners to rethink your overall conversion value strategy. Take into account that: 

• Your SKAdNetwork 4.0 postbacks have new fields for parsing coarse-conversion values, which didn’t exist in previous SKAN versions. When mapped correctly, these new values will help you get at least some indication of campaign quality even if the campaign didn't generate enough conversions to be at the higher end of crowd anonymity. 
• Conversion values can increase or decrease, whereas in previous SKAdNetwork versions they could only increase. 
• updatePostbackConversionValue has been deprecated, from SKAN 4.0 onwards you will need to use updatePostbackConversionValue(_:coarseValue:lockWindow:completionHandler:) instead.
• While in previous versions you received only one postback, in SKAN 4.0 you can get up to three postbacks across different conversion windows. This will require advertisers to rethink what will be the right conversion values to send and at what specific time. 
• The new Lock Window parameter can help you get postbacks before the end of the conversion window, leverage this new feature to receive campaign data earlier.

‍

9. Are there other changes in SKAN 4.0 my team should be aware of?

SKAdNetwork 4.0 brings many new updates, hierarchical source identifiers, including multiple postbacks, SKAN attributions for web, and more! We dive into how each of these changes will impact iOS mobile marketing campaigns in this article.

Ready to test SKAdNetwork 4.0?

Want to start testing SKAdNetwork? Are you already running SKAN campaigns but want to take your performance to the next level? Contact us to discuss how Jampp can help you drive incremental revenue in the post-IDFA era.