5 Things Every Subscription App Marketer Needs to Do Today to Prepare for iOS 14.5
Technology — like life — moves pretty fast. And often, when you think you have a grasp on things, the rules change. That’s what many marketing professionals are feeling as they await an iOS update from Apple. iOS 14.5 has enormous implications for privacy around advertising and has even the most successful companies worried about their business model and technology strategies.
If you’re still trying to figure out what would be the main implications of iOS 14.5 on subscription apps, you can read all about it on my latest blog.
If you have already done your research and you’re well informed regarding the implications, the short recap should be sufficient, or you can move right on to the practical part.
Short recap — main post iOS 14 effects
User-level data availability will decrease tremendously and affect your UA strategy, ROAS, attribution and measurement
- User-level data will not be available for the majority of your iOS campaigns.
- Retargeting gets a lot more competitive due to higher competition for the smaller potential audience, and therefore will be more expensive and will yield a lower ROAS.
- Lookalikes are a new challenge as advertisers will have to get sharp on signal selection to build lookalike campaigns maintaining a quantity — quality balance
- Accurate attribution will become a big challenge as well, due to attribution cut to 7 days (for all users, not just opt-iners).
- There will be fewer signals to rely on due to the event limit (8 events), and event prioritisation system.
My top recommendations to stay ahead of the challenges:
Start by figuring out the best balance between quality and quantity. Shifting focus from user actions to user value is going to be critical.
Start consolidating campaigns and test, test, test…
For advertisers, there is a new need for consolidation via fewer campaigns with broader targeting. There will be campaign restrictions and data limits. Create your campaigns now and test them across key metrics as you can still test your theories without reporting to the SKAdNetwork.
The key to tackling this challenge is starting early: If your campaign is successful, chances are it will succeed once the changes are live.
Facebook has come out with several recommendations to restructure your campaigns to achieve the optimal consolidated campaign structure:
For AAA and BAU campaigns that target the same country and optimization, you should consolidate to either one BAU or one AAA campaign per country. Try to limit to no more than five campaigns per single optimization.
For low spending product optimization mixes, try to fold them into existing campaigns or increase the number of countries in the mix.
Calibrate outcomes for SKAdNetwork reporting
Once you examine and experiment with different consolidations, the next step should be to understand the expected performance with SKAdNetwork reporting limitations. While it’s pretty clear that SKAdNetwork reporting will result in different attributed results compared to the current reporting, what is not clear is how. You can test this yourself — by comparing a SKAdNetwork campaign with a none SKAdNetwork campaign. First, create two identical campaigns with the same setup, optimization, targeting, and creatives. The only difference should be the SKAdNetwork effect.
Allow the campaigns to run for at least seven days, and note the differences between conversion and attribution reporting.
Right now you still have the very unique opportunity to receive two streams of data: IDFA-based deterministic data, as well as what you’ll receive from iOS 14.5 onwards. Very soon, as more and more users update to iOS 14.5 this window of opportunity will vanish.
Shift from user actions to user value to drive optimization
Advertisers will be able to measure indirect conversions with third-party tools like Google Analytics 4 or use their BI teams or a 3rd party vendor (like Voyantis) to build data models of their first-party data. Taking that further means integrating their own predictive models for internal use. Focusing on improving internal measurement capabilities will be critical, from existing tools to your team’s skills. If you have a robust measurement and analysis team, they’ll have to pin down “soft conversions” that indicate the later value and happen frequently. They’ll need to cross-check those conversions over different channels and focus on the most reliable metrics.
Use predictive models to drive optimization
Of course, leveraging AI for prediction-based signals and predictive modeling will take those capabilities to the next level. Think about it. How strong and how accurate can your signal be when optimizing your campaigns on a single event? When using proxies, the signal’s strength and accuracy are limited because they will usually fail to represent the user’s actual value. Throwing historical data into the mix means you’re not even sure to get a representation of changes in the product or user’s behavior.
On the other hand, using predictive signals allows marketers, using the same single signal, to truly embody the user’s value based on the complete set of actions and behaviors. This will enable marketers to understand better which users are predicted to be high-value and the real campaign performance. Predictive signals will put the concept of user scoring into focus, which is an evolution from user action-based decisions.
Build models that help send the right signals to the ad network
If you’re not using server-side conversions at the moment, you should definitely begin to develop technological capabilities and marketing know-how to support sending signals to indicate the user’s value or score. Server-side signal capabilities will become key to allow brands to send those more complex signals to the platform rather than a single event.
FB, Google, and Snapchat already support Server Side conversions, Ironsource has this in beta, and we expect that other networks will develop this as well.
Whether or not they’re ready for iOS 14.5 today, marketers can step up to the challenge with the right tools. Companies that can make every user count, and make small amounts of information go further, like making the right decisions around upper-funnel events, will make their spend go further. Predictive analytics brings home the sweet spot between understanding the user value and the quantity of signals. After all, companies that can maintain a certain amount of events and signals at high quality will see the difference between surviving a highly competitive environment with spiraling costs — and thriving among new opportunities.
Have any questions on iOS 14.5? Feel free to reach out .