Edited to reflect the postponement of the “permission to track” requirement: 11AM PDT
There are still numerous unknowns and conjecture-based analyses when it comes to user attribution on iOS after Apple turns on the requirement for the “permission to track” prompt. This change, planned initially for the launch of iOS 14, was recently pushed to the beginning of 2021, which gave the industry more time to prepare but didn’t clarify the situation.
Before we dive into our recommendations in light of this expected change, it’s worth noting several things.
First of all, we don’t know how users will react to App Tracking Transparency dialogues. Until the permission request is mandatory, we won’t see which attempts to explain the benefits of IDFA sharing through onboarding flows and explanation prompts are effective, if any.
Secondly, the collection of IDFA identifiers by advertisers will not constitute a guarantee of successful cross-app tracking, as all sides in the attribution chain have to collect the IDFA in order to cross-reference them. With Facebook already claiming it itself will not collect IDFA — together with most self-attributing networks—there’s not going to be a clear Return on Investment for collection efforts.
Thirdly, while the ability to attribute and measure may be hindered, the effectiveness of the ad itself is still the most crucial factor to consider—more about that in our Creative Center explainer.
Right now, we have a few recommendations for our users who are currently crafting their strategies for campaigns that target iOS users.
Don’t Assume A Gradual Rollout
The main question, which Apple is yet to answer, is whether the permission prompt will be required after an OS update or a specific date. An OS update will create a grace period for advertisers, with the inventory unaffected by the policy reduced gradually as more and more devices are updated to the new OS version. If the requirement will not be based on an OS update, most unaffected inventory might disappear overnight.
How Should I prepare?
The scope of changes to targeting means that many of the marketing SDKs you work with will be updated to support SKAdNetwork. Getting ready to work with SKAdNetwork through updated SDKs should be a priority, especially since clashes between different implementations from different SDKs are possible. You should also try to learn as much as possible from the data you will receive. Yes, it’ll be a single conversion value, and there will be a delay in reporting, but with a robust analytics infrastructure, you will still get valuable insights.
Go over your funnel. Which marketing data SDKs do you have? All of them should be up to date. You should also not neglect your measurement platform and other sources of data. At the end of the process, you should be able to clearly map your new data stream, and decide on the KPIs which you’ll track.
What should I test?
You should begin to experiment with upper-funnel strategies as soon as possible. Gradually increase your investment in optimization towards KPIs such as installs or events, as well as other early KPIs for which the data is going to be more timely and accurate. The idea is to understand the relationship between the metrics you have even without IDFA, and your KPIs. Even if it’s not very clear in the beginning, with enough time patterns will emerge.
In addition – think of the impact on the users as well. Due to these changes, users will be seeing ads that may be less relevant for them. This might require a different approach to your targeting strategy and an increased emphasis on the quality of your ad creative. After all, it is the ad itself, not the targeting, that’s the main driver of clicks.
Will I be able to continue working with my cohorted views, such as D7 ROAS?
The immediate implication of the permission requirement is the loss of ability to directly tie revenue to media spend. In other words, you won’t be able to group and compare D1 ROAS vs D7 ROAS etc..
When you implement updated SDKs for your entire funnel, you’ll be able to track a single value. Obviously, the further down the funnel this value is, the longer your campaign optimization delay will be.
Much effort is being invested in getting the most out of this situation. New analytics tools and metrics might help you regain some of the lost data.
At Bidalgo, we’re also not sitting still. We are currently working with our customers on finding innovative ways to extract meaningful insights from their BI systems, and creating heuristic models to enrich and contextualize existing data. Our Measurement Platform partners are also exploring new ways of helping advertisers.
Should I stop trying to collect IDFA?
While the industry at large anticipates high opt-out rates, the more user-level data you have, the better your campaign optimization potential.
We do want to point out that you should think of the IDFA collection prompt as a UX/Product flow. Don’t just spring it on your users, think it through. Determine the best possible place to ask them for their IDFA, and explain how it will benefit them.
What else should I try to do?
Rethink your conversion funnel to ensure that you can measure – or at least approximate – the value of the user within a single reported period, using proxy metrics for your KPIs if needed. Try to find correlations between existing valuable KPIs and events that occur within the first 24-48 hours.
Bidalgo is here for you
We continue to work with our marketing partners and measurement platforms to make your transition as frictionless as possible. Please don’t hesitate to approach your Bidalgo contacts for any question, suggestion, or discussion you want to have around the shift.