The release of iOS 14.5 (rolled out on April 26) is set to disrupt the mobile industry, with Apple introducing its App Tracking Transparency feature and the ability for users to decide whether they want to share their IDFA (identifier for advertisers) with developers and therefore be tracked.
Throughout the coming weeks, the GamesIndustry.biz Academy will explore the impact of this change for mobile game developers through a series of guest articles from experts. If you're interested in contributing, you can email firstname.lastname@example.org. You can read our first article on the topic, introducing iOS 14.5 and its impact, on this page.
The gaming app ecosystem is facing a shake-up now Apple's iOS 14.5 privacy rules have come into force. Adapting your mobile operations for these changes is a complex topic that will have serious ramifications for how you conduct business in iOS games going forward. While the changes have been long anticipated, it will become clearer over the next few weeks what the long-term effect will be.
Over the past year, we have been talking to a number of advertisers and ad networks about how they will be impacted by the changes. Something that is obvious from these conversations is that advertising will remain vital post-iOS 14.5.
But the consensus is that performance marketing will be less efficient due to a decrease in the ratio of deterministic data. With this in mind, here are some best practices to minimize the impact of iOS 14.5 on your gaming app strategy.
For extremely data-driven verticals like hypercasual games, user acquisition relies on precise campaign data so that marketers know which channels to invest their budget and how to optimize the performance of their campaigns. They're usually using specific KPIs like d0/d1 retention, user lifetime value (LTV) or ROAS to drill down into which campaigns are driving the best performance.
All of these metrics need exact data, as marketers operate with very thin ROI margins. Knowing where to invest and where to scale down is a matter of just a few percentage points.
Understanding which channels to focus on to get new users will bring back the guesswork that performance marketing on mobile had evolved beyond
With less deterministic data to rely on, understanding which channels to focus on to get new users will bring back the guesswork that performance marketing on mobile had evolved beyond in recent years. However, because advertising will remain a key piece of the mobile ecosystem, measurement and performance marketing will remain as prevalent as ever, just how they are deployed to aid apps' growth will shift slightly. Marketers will rely more on models and context, with less deterministic data to rely on.
A word of caution about Mixed Media Modeling -- we have found it has limited success in the mobile space. Mixed Media Modeling attempts to find the correlation between changes in input (for instance spending per channel) and see how it impacts output (usually installs). However, spending across channels is highly correlated -- usually, apps that spend a lot on Facebook also spend a lot on Google. This means there is only a faint signal to look for in the data making the models fragile and unstable. There is also no way to factor in organic installs. For this reason, it is necessary to add extra steps to 'decorrelate' the data.
Areas of modeling that you should consider when creating your post-iOS 14.5 data strategy include:
- Relative Channel Importance (RCI) - In general, the more you spend more on ads, the more installs you get. But which channels drive the biggest increases? RCI looks at the variances within your data to indirectly map the impact of each marketing channel on a bottom-line metric (like installs). This is similar to Mixed Media Modelling but contains the extra 'decorrelating' steps to ensure the models are more robust.
- Extrapolation - Extrapolation is where you compare data across networks with similar attribution share and infer the total number of installs.
- Smarter measurement - Machine learning algorithms can be used to probabilistically associate behavior across networks with a particular install using device entropy and patterns.
- Behavioral classification - By deploying machine learning algorithms to look for patterns in user events from different networks, you can then use them to classify unattributed installs.
Press your advantage
Since leveraging deterministic data is the one way to keep your operations as close to the status quo as possible, having a higher user consent rate is set to be a key competitive advantage. It allows for more accurate predictive modeling for optimization and ensures you will still have cohorted analytics data.
Once the automatic IDFA access is gone, Cost per Mille (CPM) will be heavily impacted. In the absence of identifiers, the deterministic link vanishes and DSPs/ad networks will need to take into account multiple parameters that can only indirectly tell if the user is a high spender or not (time of the day and basic device info). This uncertainty will be priced in, as the lack of deterministic data imposes a greater risk, which will be reflected in the price.
On the flip side, inventory that still has IDFA access will likely become more valuable -- and many industry players will compete for access to this known quantity at a premium price point.
There are a number of strategies that can be used to ensure opt-in rates are high. Developers relying on advertising for revenue will need to iterate their prompt strategy fast to get high opt-in rates, as their CPM and revenue will be at stake from day one after the App Tracking Transparency (ATT) rules begin to be enforced.
Keep it prompt
Beginning with iOS 14.5, any user that does not opt-in will not be served targeted ads, which means that the publisher will see a drop in revenue per ad. This is why some publishers may warn their users that opting out will mean seeing more ads -- since they need to make up for that loss in revenue with volume if they're missing out on quality.
In research carried out by Adjust in 2020, we found that hypercasual games can display more ads than gameplay within a minute and still generate profit. But there are diminishing returns on displaying ads. In the same report, we found hypercasuals that show more than four ads per minute hit a ceiling at around $35,000 per month in revenue.
The sweet spot appears to be between two and three ads per minute, a total that enables hypercasual game companies to boost their revenues by as much as 10%. This demonstrates the careful balancing act that will need to be considered when deciding on an ad monetization strategy for iOS 14.5 -- publishers will need to show more ads to compensate for lost revenue, but not so many that they cause users to churn.
What kind of ads developers display is also an important consideration. If we consider how the inventory auction works, it's a balancing act between what is relevant for the user and what is giving the best return on investment. Networks are often checking how many impressions lead to an install, or Installs Per Mille (IPM), and leveraging qualitative metrics like engagement, or quality of engagement, to consider prices.
But what changes post-IDFA? It becomes harder to tie a user with certainty to an individual click or impression
Social platforms, for instance, are heavily optimizing towards these qualitative metrics. If your creative gets a lot of likes or comments, these platforms are going to value this highly. Similarly, if your ad gets hidden a lot, that will be penalized by their internal metrics. This is a common optimization strategy across networks.
But what changes post-IDFA? It becomes harder to tie a user with certainty to an individual click or impression. But we still have good information about what kind of ads get higher engagement. So for ad inventory like banners and mid rectangles, which are on the lower end of the engagement spectrum, we are likely to see a fall in prices now that the IDFA is restricted.
However, it is likely that the drop in price for banners and mid rectangles will be more than for video, rewarded and rich media. Those formats are highly engaging and retain their performance levels even without the IDFA. Networks know that users like to engage with and click on these types of ad inventory, meaning they hold their value. The fact they are likely to be more attractive to the average user is still valuable, even if you can't specifically target an individual user or audience like you could before.
Set your KPIs
For gaming verticals that monetize via ads, with hypercasuals the best example, estimating the opportunity for revenue generation within an app can help to determine a healthy Cost Per Action (CPA) -- which is key for their success. In iOS 14.5, If you are optimizing on SKAdNetwork towards CPA, we highly recommend using a bucketing conversion value strategy to track "ad revenue" conditions.
SKAdNetwork provides space for 6-bits of post-install information, a number between 0 and 63 (or between 000000 and 111111 in binary), with an initial 24-hour timer. This 'conversion value' can be assigned any value that can be expressed in binary. Every time the conversion value is updated, to a fresh six-bit code defined within the app, this extends the timer window by a further 24 hours.
Once this conversion value-window expires, a second 24-hour timer for attribution starts counting down. Within this 24-hour window, randomly, SKAdNetwork returns the attribution data. The idea behind this random timer is to obfuscate the time of install, so that event triggers cannot be linked to individual users. The SKAdNetwork system shares this data in the aggregate, with no granular data accessible at the user level.
How you set up your conversion value strategy will be key to success on SKAdNetwork. For example, what's vital for hypercasual game clients is to measure ad impressions by range count. To do this they create 'buckets' that users are categorized into depending on how many ads they view.
For example, if you want to measure how many users have made 20-30 ad impressions or generated USD $1-2 in ad revenue, you can define a conversion value schema that supports count or value ranges.
With a bucketing strategy, you could map a user racking up 20-30 ad impressions to conversion value 3, or a user generating $4-5 in ad revenue to conversion value 5. This provides a granular look into how that user is performing in the first hours of their journey inside your app, while providing the flexibility to define the ranges as you see fit.
The iOS 14 changes will be a big adjustment for how app marketers operate on Apple devices, but with the right strategies, you can protect your growth trajectory and ensure you are primed for success in the post-IDFA world.
As chief product officer, Katie Madding leads Adjust's global product vision and development. Madding is responsible for Adjust's roadmap, pushing for features that meet at the intersection of the company's vision and its clients' goals.