A brief history of SKAN: before and after

The early days of the SKAd Network

Before the implementation of the SKAN network, advertiser IDs were shared across all networks to see if a potential user clicked on a specific ad. This gave app marketers plenty of information that enabled them to measure and optimise ads based on data that reflected user activity. However, growing concerns over user privacy and data collection gave rise to the need for an alternative to traditional mobile app install attribution methods, such as cookies or device IDs - one with more privacy than before.

Initially introduced by Apple in 2018, the SKAd Network brought on a more privacy-friendly framework for mobile app install attribution. Then, in 2020, Apple announced its SKAdNetwork 2.0 and ATT (app tracking transparency) at WWDC, greatly affecting mobile app marketers and advertisers. While people were automatically tracked before, the ATT prompt now meant that users had to opt-in in order to be tracked both on the platform where the app was being advertised, and inside the app marketers were advertising for.

While the introduction of the SKAdNetwork certainly increased user privacy, it did make things harder for app marketers:

• While we still had the ability to track installs and some post-install activity, the new reports only included events that happened in the first few hours of app usage. 
• It took at least 48 hours to receive data. 

The limitation in only having information pertaining to the first day of usage in combination with the delay in receiving the data made optimising campaigns very difficult. 

SKAN 4.0: advantages and disadvantages

Back in October 2022, Apple finally released the next version of SKAN (SKAdNetwork 4.0), which gives more tracking possibilities at the expense of adding complexity to data interpretation. The most significant changes introduced help create a balance between ad networks being able to measure more, and securing user privacy, including:

Change 1: multiple postbacks with different conversion windows instead of one

Advertisers can now receive a second and third postbacks as opposed to one with SKAN 3.0. Each of the postbacks is focused on a particular activity window to enable advertisers to understand how users engage with their apps over time. The postback windows are:

Change 2: LockWindow

This allows app developers and advertisers to finalise their conversion values, and lock in the measurement window to receive their postbacks sooner.

Change 3: Crowd anonymity

This is Apple’s way of describing how SKAN delivers attribution data while preserving user privacy. The more installs you get, the more data you’ll actually receive.

Change 4: Coarse conversions values

In addition to the fine conversion value, Apple added a second, less precise conversion value, with only three potential values: 0, 1 and 2.

Change 5: Source identifier of 4 digits

In SKAN 4.0, this identifier consists of four digits instead of 2. It will help to track performance at a more granular level with 3 dimensions of the ad network. What’s more, it could be used to track installs at the campaign, adset, and creative levels.

Change 6: Web-to-app attribution

SKAN 4.0 enables web to app tracking via their Web ads API. It’s only available for Safari.

It can be a lot to take in at first, however, your measurement partner will be a good help in navigating these changes and implementing the new SDK properly.

3 lessons on SKAN 4.0 from Kovalee’s Growth team

Lesson 1: Passing the minimum privacy threshold

To ensure user privacy, Apple requires you to pass the privacy threshold by generating a minimum number of installs per day and per campaign. As a result, app advertisers need to structure app marketing campaigns around receiving larger volumes of data. 

Privacy thresholds vary depending on the ad network used and on how the SKAN campaign IDs are managed. For instance, the tests we conducted on Meta with SKAN 3 showed that you need at least 30 installs per day per ad group to start passing the privacy threshold. With crowd anonymity as a new feature, the privacy threshold may be even harder to fully pass.

What steps can you take to pass the privacy threshold? 

1. Prioritise passing the minimum threshold and receiving the data you need above hyper-focusing on very specific targeting.
2. Cluster markets with the same values and targets for specific audiences that are large enough. For example, exclude young audiences if they're not your core demographic. 
3. Once you’ve done that, then you can look into scaling your campaign. 

Lesson 2: Focus on events that really matter

Before SKAN, we tended to track all events conducted by users. Now, every additional event tracked adds a layer of complexity. More importantly, it could mislead ad networks on how to optimise the diffusion of your campaign if not well transcribed. 

Here are three simple steps you can take to avoid this:

1. Identify your most important events or metrics (ex. a purchase, a subscription, a registration, etc.). Remember that this event needs to be done in the first 24 hours of the user journey and avoid events that often occur 3 days after the download as you will not be able to use it for SKAN first postback.
   
   For example, you’re running a UA campaign for a subscription based app and you offer a three- day free trial period. In this situation, “trial started” as your main event and not the subscription event for the SKAN campaign.
   
   
2. Single out secondary events or usage that show a significant uplift in user value (ex. time spent in the first session, the number of actions taken, whether multiple events were completed in the same session etc.). This will help you to set a clear and powerful conversion value scheme.
   
   For example, if you see a high correlation between people that use the app for more than 5 minutes the first time and late purchases (purchases that don’t happen in the first 24h), then make sure to use a conversion value slot for this.
   
   

3. Ensure you’re able to interpret the data received and give clear guidance to the ad networks on which conversion values they should optimise the campaign for.

We opted for this approach to improve our campaign performance as it enables us to correctly interpret the data without getting lost in the sea of conversion values and information that wouldn’t be useful in optimising our campaigns. This approach also gives the network the opportunity to understand which users are the most valuable for us and the app we’re marketing. The algorithm identifies similar patterns in what we’re tracking, and can then target those user patterns for us.

Lesson 3: Optimize towards your main goal, not just installs

It’s important not to get caught in the potential pitfalls of taking a short-term approach, 

You should always align your ad campaign strategy with your overall business goals, by:

• Running campaign optimisations toward your main goal from the start
• Optimising and fine-tune your campaign to make it work by testing large amounts of creatives, and adjusting your target audience
• Only scaling once you’ve achieved success with your primary goal

It’s very easy to get caught in the pitfalls of taking a short-term approach, such as:

• Attracting low-value audiences: optimising for installs alone can attract users who may not be valuable to your business. Having a high number of installs doesn't necessarily translate into higher revenue or user engagement.

• The need for a second campaign: you may find yourself creating a new campaign from scratch when trying to optimise for a specific event. This is time-consuming and you will not fully leverage the knowledge you gathered from your first campaign.
• Misleading insights: what works for gaining a high volume of users may not necessarily work for attracting high-value users, especially on the creative side of your campaigns.

Can Kovalee help me navigate my apps’ growth?

At Kovalee, we also rely on our in-house tools that use data science to better identify the events that really matter and generate user forecasted lifetime value. Connected to Krome, our automated marketing campaign management tool, we optimize bids and budget based on forecasted ROAS. 

If you’ve got an app within the iOS ecosystem, we can and will help you. We’re a team of app experts with over 10 years of experience working with mobile app developers on Apple apps, and we’ve helped several apps crack the global top 10 in their respective categories. Our partnership model is simple - we free up app creators to focus on what they do best, content & code, while we take care of everything else. From app store optimization (ASO), monetization, and UX/UI creative design. We invest in the apps we believe in, and forge a strong and lasting partnership with them.

Learn more about app publishing here, or reach out by filling in this form below and we’ll get in touch with you as soon as we can! Or, if you'd like to become a Kovalee Kreator and build some awesome UGC for our partner apps, reach out to us here!