The impact of Apple’s AppTrackingTransparency policy and how to prepare
Apple’s iOS 14.5 is due to be released within the next fortnight and will include the mandatory adherence to the AppTrackingTransparency framework. Thereafter all iPhone, iPad, and Apple TV app developers will need to receive a user’s permission to track their activity across other apps and websites via their device’s advertising identifier (IDFA), for targeted advertising and to measure campaign effectiveness.
IAB Tech Lab quickly provided a tool to manage and maintain the SKAdNetwork ID list.
We have also checked-in with a few of our expert members on this topic and asked them some questions related to the potential impact and what recommendations they have for the industry to prepare. We hope that you find it useful…
Richard O’Sullivan, Vice President Australia & New Zealand – InMobi
If any app developers have not yet looked at this forthcoming change, what are the first three things that they should do?
- Implement SKAdNetwork
- Publishers will have to upgrade SDKs to become compatible with Apple’s SKAdNetwork (SKAN) framework, without which they will not be able to access any performance advertising spending. Performance ad campaigns (i.e., app install and remarketing) are a significant component of mobile app publishers’ advertising revenue – up to 60% for some publishers on InMobi Exchange. These campaigns, having always relied on IDFA-based third-party attribution solutions for measurement and optimization, will henceforth only use SKAdNetwork as the default attribution method for iOS 14.5 and above. And all old measurement and attribution workflows will stop working on all LAT supply from 14.5 onwards
- To enable app install attribution via SKAdNetwork, publishers need to collect the SKAdNetwork IDs (SKAN IDs) for every demand source they work with and add them to their app’s info.plist file. The info.plist file, or information property list file, is an XML file that is a part of the app’s code, containing various properties of the application
- Publishers must ensure that their app’s “info.plist” files are always up to date with the full list of SKAN IDs of all bidding DSPs, ad networks or advertisers. InMobi’s SKAN IDs are available both in XML and JSON format.
- Please ensure that the SKAN IDs are entered in lower case. Publishers should upgrade to the SKAN-compatible versions of their monetization partners’ SDKs. InMobi iOS SDK 910+ supports attribution via SKAN. You can download the latest SDK on the InMobi support portal, website or CocoaPods. For more info, please read InMobi’s guide on iOS 14.5 changes.
- Implement ATT and Run Tests to Maximize User Opt-In via ATT
- Starting from iOS 14.5, publishers will have to implement Apple’s App Tracking Transparency (ATT) framework to request consent for tracking any user data that can be shared with other companies. Note that according to the latest App Store policy, publishers should compile a list of all data types they or their third-party partners are collecting and note how the data is used (whether they are linked to them or used to track them). Publishers should ensure this list is complete and is always up to date with every partner they work with, even on an ongoing basis. Publishers should test their consent strategy to maximize opt-ins.
- Optimize Yield Across Inventory by:
- Integrating with Universal ID vendors
- Optimizing LAT and non-LAT waterfall for mediated waterfall traffic
- Activating higher CPM ad formats
- Activating first-party deals
How much flexibility and opportunity do app developers have in communicating the choices and resulting impacts of this new requirements to their users?
Starting from iOS 14.5, publishers will have to implement Apple’s App Tracking Transparency (ATT) framework to request consent for tracking any user data that can be shared with other companies. Note that according to the latest App Store policy, publishers should compile a list of all data types they or their third-party partners are collecting and note how the data is used (whether they are linked to them or used to track them). Publishers should ensure this list is complete and is always up to date with every partner they work with, even on an ongoing basis. However, publishers should test their consent strategy to maximize opt-ins
Prime users about the opt-in prompt: Publishers can look at priming the users about tracking and privacy before displaying the request tracking prompt. Doing this will help users make a more informed choice and likely increase opt-in rates. Publishers can use their own app real estate to do this. They can explore opportunities outside the app too. Some options that publishers could consider are running push notifications, displaying sticky banners with illustrations, modifying their current GDPR/CCPA consent dialog to act as an explainer before the opt-in prompt shows up, or even launching related email campaigns. Even Apple has clarified that it’s okay for developers to provide additional information to educate users before showing the ATT opt-in pop-up “as long as you are transparent to users about your use of the data in your explanation.”
Test timing, messaging of the opt-in prompt: Publishers can show the opt-in prompt to users only once in their lifecycle, but they are free to choose when. Users are most likely to grant their consent when they are most engaged with the app and have already seen proof of value. Apple will also allow publishers to customize some of the messaging in their opt-in prompt. Publishers should leverage both points to test (across apps, user cohorts, geographies) the timing and messaging of their prompts to maximize opt-in rates.
How can the industry keep across any changes to the versions of the SKAdNetwork API and any related requirements?
With the release of iOS 14.5, performance advertisers will have no other way to measure and to attribute installs other than Apple’s SKAdNetwork (SKAN) framework for their LAT or non-consented traffic. MMPs that used to provide third-party measurement and attribution solutions will also stop supporting probabilistic attribution on LAT traffic because of Apple’s restrictions on using device or user network data for any tracking purposes if users have not provided their explicit consent.
Some more points for advertisers to keep in mind as they implement SKAN workflows at their end.
Testing: We are encouraging our demand partners to perform an end-to-end postback validation testing with us and confirm if they can record correct data. We have also created and made available a test app that reduces the usual postback turnaround time from 24-48 hours to 5-10 minutes just for testing purposes. Buyers can access InMobi’s test app, and steps to use our test app can be found here.
Measuring the right events: Buyers should note that Apple’s SKAN API allows for a limited number of post-install events to be tracked compared to before, where advertisers could track many more events and then narrow down to the most significant ones. Furthermore, there is also a 24-hour time window between the in-app action and the postback activity. In this new world, buyers should be more mindful and intentional about the events they choose to track and optimize for in their campaigns.
An Evolving Framework: Because SKAN is a framework that’s relatively new to the entire ecosystem, demand partners could expect tech fixes and upgrades along the way as the ecosystem sees broader SKAN adoption.
Please refer the InMobi Exchange whitepaper for more detailed responses and answer to other similar questions.
Manuela Cadd, Country Manager AU/NZ – Verve Group
From a consumer privacy perspective, do you believe that this is the right type of product evolution for the industry?
Over the years, consumers have transitioned from browsing the web anonymously to adapting to authenticated websites and services such as messengers and social media platforms, becoming accustomed to sharing information openly with the digital world. The advertising industry began to get used to tracking each click and impression, gathering as much information as it could. Now that users are focusing on taking back ownership over their data, the industry’s engines have started turning again. International governments have started focusing on protecting user privacy with the introduction of regulations such as GDPR and CCPA. Now that Google has followed Apple’s strategy and announced to phase out third-party cookies without working on alternative identifiers, the industry starts to look at data differently again.
As with all disruptive changes, there is a sudden need for innovation that will bring new discussions and technologies to the table. Instead of holding on to identifying unique consumers, advertisers and publishers will adapt to market-cohorted and contextual segments. Google and Apple, however, are using the buzz around consumer privacy to draw up their guards, strengthening their positions of power behind the walled gardens they built. And their latest decisions have not been well-received by other big players in the market.
Instead of incentivizing and giving in to the duopoly, we are driven to tune into an increased collaboration in the market. While some of our industry efforts should lie in educating users on the benefits of advertising and addressability, our main sources of energy should be directed towards developing alternative solutions for the open internet. There are many different platforms to consider when developing new measurability and addressability solutions, including desktop web, mobile web, or mobile apps. While we might pursue different approaches to solve for consumer addressability and measurability, cohorts and context will play a great role in all of them. With the help of IAB’s taxonomy, the industry can unify the language used across solutions to help the whole open ecosystem trade within it.
Overall, the trend towards “privacy-first” advertising is obvious. We should keep in mind that the challenges we are facing today hold opportunities to rethink not only our industry values but also innovate our dusted advertising practices. The fundamental changes have rippled the industry waters and will form new currents in which our data flows.
Peter Barry, Regional Director, ANZ & Head of Audience, APAC – PubMatic
How significant do you anticipate the initial reduction on reach and campaign volumes will be? What about in the long-term?
It is difficult to say exactly how many users will opt in to being tracked, a lot depends on how publishers frame the message to their users. We do know that historically it takes roughly 2 months for iOS updates to reach 50% uptake (source: statcounter global stats Jan-feb 2021) – Apple currently reports that 80% of all devices are running iOS 14.
The changes may mean that performance campaign eCPMs will drop and brand eCPMs may become more competitive. IDFA enabled traffic may command a higher eCPM and we may see a shift in budgets to Android and MWeb/Web. Contextual signals will also become more relevant.
If any app developers have not yet looked at this forthcoming change, what are the first three or four things that they should do?
– 1. Consider the use of pre-prompt. So, get on the front foot to convey the upcoming changes to consumers and what it may mean for investment in content. You need to let them know the value exchange. You should A/B test these messages to see what works best. This runs ahead of the actual iOS consent screen.
– 2. Update a plist file with SKAdnetwork IDs – similar to app-ads.txt. SKAdNetwork framework allows app advertisers to measure the effectiveness of app install campaigns without the ability to track individual users.
– 3. Update all of your app monetization SDKs to recent versions that support iOS IDFA changes.
– 4. Check to ensure important parameters are populated in the ad requests (App bundle, app store URL, user gender, age, keywords, location (country, postal code, and GPS-based lat/long), and device info (OS, make, and model).
Could policies such as this result in an ever-increasing reliance on first-party data, thereby ultimately limiting advertising’s reach across the open internet and incentivising walled-garden strategies in the longer term?
As an industry we are all working together to build out a long term, opt-in solution that works cross-device and across the open web. This solution should lead to a greater degree of consumer trust as well as a set of scaled options outside of the walled gardens.
Ultimately the opportunity is for publishers and developers to get a greater share of media spend to re-invest in content.
Yun Yip, Managing Director APAC – HYP
How significant do you anticipate the initial reduction on reach and campaign volumes will be? What about in the long-term?
What we (HYP as well as our publisher and app partners) have evidenced in this short window is that the impact / drop off will vary largely as a result of the value exchange data sharing is between the publisher/app and the consumer. So far, our partners have seen a slight decrease (around 15%, not as significant as initially thought) because their users view the service provided by the app as highly beneficial and worthwhile. As an example, I willingly share my location data with my running, ride sharing and weather apps. The opportunity cost of withholding my data from those apps is too high. The percentage of drop offs, therefore, will vary highly across the types of apps. With regards to the longer term view, maintaining reach and users’ data sharing will very much depend on 2 key factors.
First, the publisher’s ability to articulate the value behind data sharing. If the publisher/app is able to provide a narrative that enables the user to understand the enhancement and relevance of content/experience provided through data provided, as well as enable trust around privacy and governance, the reduction in reach and opt ins will be skewed towards inferior apps and pubs.
Second, the publisher continues to deliver and personalise data sharing benefits with the user.
Could the regression of personalised advertising within iOS have a negative impact on consumers at all?
Yes it absolutely can – who wants to be shown an ad (or many ads) that hold no relevance to them? We can all relate to it. Advertising is only useful and meaningful when the consumer deems it beneficial or aligned to their needs/wants etc. I’ll address the question from a slightly different angle. “What is the extent of regression of personalised advertising within iOS? What is our solution?”
iOS 14 has actually opened up the opportunity for marketers to rethink being over reliant on a particular type of data and / or network. At the core of this is working with publishers and users directly to create a transparent and clear value exchange for data sharing. Effective personalisation is still possible and there are a few easy ways to get there:
- cheap is cheap for a reason. Choose to work with quality publishers and apps where the value exchange will always be clear and valuable between them and the consumer – and opt-ins become a no brainer.
- consider ID solutions which have been built on consumer consent and control.
- contextual relevance is very much alive. Tailor your audience segments based on context (hint: HYP’s ID-less targeting allows personalisation based on demographic/behaviour/income/geo, ranked on context).
- as the cookie crumbles, and another route to personalisation comes to a dead end, ensure that your data partner is 100% cookie free.
- beyond targeting and personalisation, consider using attribution and incrementality to inform how your consumers want to be communicated with.