iOS 14.5: Thoughts from Lomit Patel @ Together Labs
As the release of Apple’s long-awaited iOS 14.5 update nears, we catch up with Mobile Heroes to learn how they are preparing for the post-IDFA world. Today we talk with Lomit Patel, VP of Growth at Together Labs (formerly IMVU).
How are you feeling about iOS 14.5?
I’m excited to see Apple taking the lead and shaking up the industry to ensure that consumers know the value of sharing their personal data. I feel good because this will impact everyone across the industry equally and force us to be better stewards of our customers’ data.
To prepare, since last August, we have been actively looking into ways to drive the iOS business, via Probabilistic Modeling as well as Media Mix Modeling. find the best way to move forward because iOS still drives our best quality LTV customers. We are very grateful that Apple gave the industry the extension which allowed both the demand and supply sources to properly configure their systems. I’m confident that iOS will continue to drive our best quality LTV customers moving forward.
What are your biggest concerns?
Managing our budget blindly is never a good strategy and that’s concerning. Not being able to attribute spend to specific campaigns will drastically affect how we run our business on iOS and how we measure and optimize for ROAS. Especially when it comes to retargeting.
We are most concerned with the disruptions to attributions at the MMP level and the changes in optimization methods available for different channels. We also feel that the lack of lookalike audiences and losing the ability to retarget large segments of users will increase CPMs and impact performance and user LTV.
What have you done to prepare?
We have been busy preparing since Apple made the announcement. We began by closely collaborating with our partners: we tested non-personalized (LAT) traffic with partners like Liftoff and since September we’ve been beta testing AppsFlyer’s SDK so we can learn and prepare together. We then baselined our pre-iOS14.5 campaign performance, prioritized the events for SKAdNetwork (SKAN) capture, and shifted significantly more budget to Android. We froze our ML so that we can still use it for user acquisition post-iOS 14.5. We started proactively testing the iOS 14.5 IDFA opt-in to get real data on the impact instead of waiting until Apple forces the change, while also increasing our CRM efforts to capture users’ push and email opt-in permissions, so we can also continue to effectively message our users outside of our app.
Are you planning to re-allocate your mobile budget? If so, how?
We are currently continuing to spend on iOS to capture users prior to the upcoming changes and the risks they entail. We are also baselining performance to accommodate for changes in performance due to SKAN attribution and optimization methods. Our AI Intelligent machine makes all the key decisions on budget allocations, so we will continue to trust the system to proactively relocate budgets from iOS to Android channels as needed, without human interference.
We have also launched a big TV & OTT test to open up a new growth channel that has the potential to become a major new UA source for us.
What is your ad creative strategy for iOS 14.5 and beyond?
We will pay more attention to detailed ad creative performance metrics on Android campaigns, then pivot those learning over to iOS. This is similar to what we currently do for App Store Optimization (ASO) by testing in Google Play Store and applying the key learnings into our iOS App Store listing.
How will you evaluate performance?
Since SKAN may limit what we can collect, we will start focusing on early data events such as registrations and using our baseline average revenue per user (ARPU) by platform to determine how the projected revenues on iOS14.5 campaigns are performing.
How do you plan to optimize for user LTV?
We plan to use existing Customer Relationship Management (CRM) means such as notifications, email, and our social channels to continue to engage existing users to maximize engagement, retention, and revenue.
How are you mapping conversion values and what is the thinking behind your plan?
A user’s first purchase continues to be our main KPI. Given that the measurement windows may not capture all the purchases, we will be focusing on higher-funnel events, such as registrations, and using their correlation to purchases and revenue as a proxy.
Are you testing a user consent pop-up? If so, what opt-in rate are you seeing?
We have been proactively testing a user consent pop-up since mid-March and the opt-in rates are right in line with our projections. I will not share our specific opt-in rates due to confidentiality but I will say, it’s much better than most of the industry’s estimates.