Nov 16, 2022
Growth Strategies
Read Time Icon - Portfolio Z Webflow Template
 min read

Impactful Strategies Amid iOS14

After writing about the implications of iOS14 on the mobile advertising ecosystem, I got many requests to write about impactful strategies that can be put in motion for mobile app install campaigns amid the flurry of changes happening. I’ll write about why you should push spend now, first party login loophole, building attribution baseline models for when data starts to go dark and flow updating.

First off, this is one of the most radical shifts we’ve seen in growth marketing, so if you’re feeling overwhelmed, you’re not alone. There’s a huge amount to unpack but we’ll take a look at some high level strategy shifts you can be making now to be in the best spot moving forward. What this article isn’t is a step-by-step on how to implement SDKs or how to best structure events and campaigns for the latest best practices. For more information on that, you can go to Facebook (as an example) who has provided an easy to follow list for their platform.

Pushing spend

This may seem obvious, but as we inch closer to the inevitable doomsday when Apple enforces iOS14 privacy changes, the less time we have with the most accurate measurement. What this means is if you’re able to frontload spend in Q1, it’ll be a surefire way to maximize the targeting and tracking capability we still have. Use this time to test anything and everything so that it can help inform your decisions on what was working before Apple’s policies are enforced. Did you have 20 new assets to test for the quarter? Test them all asap and collect as much data necessary. You may find a key creative winner that drives performance for the remainder of the year. New targeting to test? New copy to test? Same idea.

First party login

Let me preface this section by saying that it definitely doesn’t fall under Apple’s kosher list of actions. And let me further preface this by saying that I do not condone doing this. This is merely to discuss what I believe will happen with developers looking to circumvent user data loss. What I foresee are more apps forcing users to login via email and/or phone numbers. It’s nearly impossible for Apple to then know if these apps are sharing lists with Facebook for lookalike targeting or retargeting lists. What may happen is that Apple requires “Sign in with Apple” in the future if this becomes an abused loophole to route data to providers for ad targeting.

Example app (Bird) utilizing “Sign in with Apple”.

Baseline modeling

By using current data or painless testing, one can pretty reliably create a model to predict performance when data goes dark. This is a standard practice that has already been in motion on mature growth teams post-LAT user increases. The chart below shows how LAT users have ballooned from 16% to over 25% of iOS users from 2019 to 2020.

LAT opt-in users from 2019 to 2020. Image from AppsFlyer.

The idea of baseline modeling is pretty simple: we take our Android campaign performance (that has much cleaner attribution) and look at install > key event conversion rate. Let’s say for our shoe marketplace app, the conversion rate is 10% from install to purchase on Android. We then apply that conversion rate to iOS campaigns and assume a 10% conversion rate from install > purchase. This is not 100% ideal, but a simple way to understand how campaigns are trending without perfect attribution data.

Another way to measure the effectiveness of iOS campaigns is also quite painless. It’s a simple light-switch solution. What you’ll want to do is run a pre/post test on your iOS campaigns by running them for a week and then turning them off for a week to measure the changes you see in your key conversion event — in the case of our shoe marketplace, it’d be purchases. The biggest caveat to this is that you don’t want to make any changes to anything else. If you crank up the spend on a different channel during this test, it’ll void the results.

It’s a simple light-switch solution.

If you want to eliminate as much noise as possible and not be blocked from adjusting other channels, you can run the test in a specific geo to see what the volume effects are there. Same methodology as above, but just running in one specific single-market campaign.

Flow updates

Due to the nature of SKAdNetwork, the 24 hour max timer that’s triggered after the initial app open will force product teams to hurry users through completing key in-app actions. Imagine an app that currently has a 48 hour lag between app open > purchase conversion. Unless there are other events being sent between app open and purchase, the conversion will be lost and past the 24 hour timer. To solve this, events between app open and purchase conversion can be tagged to re-trigger the 24 hour timer. If at all possible, your key in-app action should happen within 24 hours of initial app open so that you can pass conversion data back sooner to Facebook.

SKAdNetwork diagram showing timer cycle. Image from Incipia.

These changes will be at the mercy of the user experience, but if you can find the right balance of a pleasant flow + getting users to your key conversion event as soon as possible = that’s the winning formula.

Interesting times ahead

While these solutions may be nowhere near from perfect, this hopefully sheds some light on small (but impactful) adjustments that can be made to alleviate attribution degradation. These are the types of adjustments we’re going to have to get used to as Google inevitably rolls out similar privacy protections in the future along with more CCPA style regulations.

Feel free to reach out if you have any specific questions or suggestions on next topics to write about. You can find me over at LinkedIn and my revived-Twitter that I heavily used while a YouTuber.

Subscribe to my growth newsletter

Get notified the moment I publish a new article.