Optimizing for the first 24 hours: Conversion values and iOS 14.5+

The GamesIndustry.biz Academy explored the impact of the release of iOS 14.5 for mobile game developers through a series of guest articles from experts, which you can find gathered on this page.

Attribution has undergone immense change with Apple's iOS 14.5+ privacy updates, and the shakeup of the mobile gaming ecosystem continues, as marketers and developers adapt their strategies and operations.

Speaking with advertisers and ad networks, it's clear that advertising is as fundamental to the industry now more than ever. The questions now are: how do we move forward with performance marketing? How do we work out the most efficient ways to drive the best users to apps and continue to offer them optimized and personalized experiences?

With less deterministic data to rely on, a certain level of guesswork that had been ironed out in recent years has now resurfaced

With less deterministic data to rely on, a certain level of guesswork that had been ironed out in recent years has now resurfaced. For mobile marketers, this means shifting to models and context to understand which channels to focus on to get new, high LTV (lifetime value) users. And when it comes to SKaDNetwork, those first, all-important 24 hours is where you can find an advantage.

Many apps, however, see key in-app actions occurring all throughout the first week -- not exclusively within the 24-hour window -- highlighting further how important it is to shift focus and identify behaviors and events you might not have been focusing on before.

Behavioral engagement and the first 24-hours

When working with Apple's SKAdNetwork, it's vital to leverage all behavioral data available in the first 24-hour window to its full potential.

Once optimized with a robust conversion value strategy, the number of installs being fed to you should be similar to what you were receiving pre iOS 14. Once you're happy with your install volume, it's time to focus on quality -- and on understanding which first-24-hour in-app behaviors lead to the highest-quality users.

The way SKAdNetwork allows marketers to collect information about user behavior within their app is with conversion values -- providing 6-bits of post-install numbers (a conversion value between 0 and 63).

Once the initial, post-install 24-hour window from a user expires without an update to the conversion value, a second 24-hour timer for attribution starts counting down. Within this second 24-hour window, the SKAdNetwork will randomly return the attribution data, obfuscating the time of install so that in-app event triggers such as log-ins and purchases can't be linked to individual users.

The way you set up your conversion value strategy is the key to success on SKAdNetwork.

While it's no longer possible to offer users the same level of targeted experiences based on their data, you can still drive the same number of installs. From this point onward, you've got everything you need to optimize your ads and ensure you move forward with acquiring users that offer high LTV. The journey may look different, but the results can be just as promising.

Getting the conversion value strategy right

Conversion value systems will look different depending on the needs of different verticals and different sized apps. A simpler conversion value system allows you to map six events to conversion values, where each bit of the conversion value maps to an event. This approach is suitable for many use-cases but doesn't meet the needs of all apps.

For those wanting to track as much behavioral engagement/as many events as possible within the first 24 hours -- focusing on count and value ranges allows advertisers to examine user behavior on a more granular level. For example, they may want to focus on how many sessions, events, in-app purchase revenue or ad revenue (or any combination of them) and then link these buckets to a conversion value.

To create the best possible conversion value strategy we recommend taking the following steps:

• Set your focus to either engagement or revenue.
• When looking at revenue, identify the clusters of revenue ranges that exist for current users.
• When looking at engagement, explore the engagement behaviors of users in the first 24 hours (independent from the source they were acquired). If you can easily identify differences, consider adding engagement-specific tracking into the conversion value coding. If not, determine what's missing and how you can gather the data points needed.
• Launch and observe the distribution of triggered conversion values. In the best-case scenario, you'll see clear peaks and troughs on the performance as opposed to a homogenous picture.

It's becoming increasingly difficult to identify whales and link them back to specific sources, which is why building a strategy focused on the first 24 hours is integral to identifying user groups and segments that will perform best. This is the way to move forward and is what you should be optimizing toward.

Don't get tripped up

Advertisers want to nail the conversion value, so they're (rightly so) focusing on designing unique and specific algorithms that pack in as much information as possible. Generally speaking, this is the most effective way to gather data when working with SKAdNetwork -- but there are a few pitfalls to avoid.

It's essential to base your strategy on behavioral engagement that can be effectively and reliably measured in those critical first 24 hours

Robust models, where certain key events or even key event ranges based on events that do not occur on a rolling 24-hour basis (for example, where the critical event being tracked is triggered more than 24 hours after install), mean that the conversion value will no longer be updated.

Regardless of how clever the algorithm is in this case, you won't get fed the information you've built it to deliver you. So, it's essential to base your strategy on behavioral engagement that can be effectively and reliably measured in those critical first 24 hours.

Push for engagement

Data that comes from SKAdNetwork is based exclusively on in-app actions completed in the first 24 hours. This means, simply, you need to extract as much information as possible in this time window. A few options to try include:

• Using a high-impact push notification strategy for reactivation to keep users coming back to your app to complete more actions.
• Crafting a more 'aggressive' push-to sign-up UX to encourage users.
• Focusing on catching different data points that help to separate your user segments. These might include gamified questionnaires, surveys, or free trials.

Whichever strategy you implement needs to be tested and validated to ensure it's not having a negative impact on overall in-app conversion rates -- and the way you tailor your approach varies from vertical to vertical. The principles, however, are the same: every data point that can be gathered will help you to identify engagement clusters that indicate higher LTV based on behaviors that take place in the first 24 hours.

You can then use these to build an optimized conversion value configuration. A bonus? This strategy will also help you create more advanced prediction models in the future.

Ultimately, success on the SKaDNetwork is determined by how well you work with the data you receive in the first 24 hours. Engaging high volumes of users, as much as possible, within this window is the key to obtaining the data points that will be integral to clustering and segmenting users that bring the most LTV.

We believe that unpacking the behaviors found in these data points -- often those you wouldn't have normally monitored -- is essential for building conversion values that can accurately predict and assess the quality of the users you're acquiring.