GUIDE

iOS 14.5+: From day one until now

Introduction

It’s been two and a half years since Apple first announced the release of iOS 14 at its Worldwide Developers Conference (WWDC) in June 2020, and more than two years since its release to the public on September 16, 2020. From the initial announcement until the rollout of iOS 14.5 in April 2021, the mobile marketing industry was sent into overdrive and pushed to reassess the way it handles user privacy and its approach to the mobile advertising ecosystem in general.

At Adjust, we’ve remained at the forefront of these changes, working closely with Apple, our partners, and clients to make the transition as seamless as possible and to provide guidelines and solutions that serve the needs of marketers, developers, and advertisers. Mobile Measurement Partners are still essential to mobile marketers. The definition of ‘measurement’ and how we deliver value has simply evolved. It’s no longer just about a strict focus on attribution and deterministic data but a broader, cohesive analysis of aggregated data.

With the official release of SKAdNetwork 4.0 (SKAN 4) and iOS 16 in October of 2022, further change was brought to the space. The way SKAN works was significantly altered, leading the industry to once again reassess its approach to mobile marketing and measurement on iOS—something many businesses had only recently figured out. At Adjust, we’ve continued to test, engineer, and solve for the updates and changes, which is why we have already released our SKAN 4-ready SDK. As with all things iOS, we are constantly reiterating our commitment to investing in next-generation solutions that align with the privacy-centric evolution of the market—moving away from solutions that focus on individual data points and towards optimizing insights for aggregated data. As part of this, we also launched Conversion Hub in 2022, our all-in-one solution for smart, easy, and effective conversion value mapping.

From understanding SKAN up until and including SKAN 4 and AppTrackingTransparency (ATT) to best practices for getting the opt-in, building and mapping conversion value setups, working with Adjust’s Conversion Hub, and creating marketing strategies that perform best in the post-IDFA world, there are a lot of topics and themes to cover.

The industry is becoming increasingly privacy-centric, and Adjust embraces these measures to ensure user privacy and data protection. In this guide, we put a lens on the changes from iOS 14.5+ to iOS 16.1, the significance for the industry, and how Adjust is adapting to deliver value to clients. We examine the ins and outs of working with SKAN and provide a holistic breakdown to ensure that marketers are up to speed with everything they need to know.

Part 1

iOS 14 and the initial release: What was all the fuss about?

Up until the release of iOS 14.5, Apple had allowed each device to have a unique and resettable identifier, which was accessible to all apps downloaded on that device. The option to limit ad measurement was available in the settings of an Apple device, but most users didn’t know, or didn’t bother. This ID, the Identifier for Advertisers (IDFA) could be used by marketers to measure clicks and compare them to installs, which mobile measurement partners (MMPs) were then able to attribute. Iterations of this, starting with the Unique Device Identifier (UDID) had been in place since 2008—the switch to the IDFA took place in 2012.

The mobile ecosystem has since developed and evolved, however, along with user knowledge of it. Concerns about data privacy and how data is accessed and managed was and is a  growing theme among users (and legislators). This has led to regulations like the EU’s General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA), which brought in new rules about how data must be respected and processed in those jurisdictions.

While the app industry continued to thrive alongside these changes, Apple’s announcement at WWDC 2020 that access to the IDFA would be contingent on gaining user consent via a pop-up garnered a much more severe reaction from the mobile marketing industry. While the impact has not been as catastrophic as some early predictions indicated, and despite user privacy having already been top-of-mind in the industry, it did cause a seismic shift in how we’re fundamentally able to approach attribution and measurement.

Important: Although ATT was initially announced in 2020, it wasn’t until the release of iOS 14.5 in April 2021 that it was actually enforced.

This change meant that attribution as it was understood for apps and app marketers would be significantly impacted and that the way they had been measuring user acquisition could no longer be guaranteed. Essentially, there are two attribution and ad measurement approaches that can now be leveraged on iOS: The ATT framework that manages access to the IDFA with user consent, and SKAN (starting with iOS 14.5’s SKAN 2.2). If consent is acquired, those users can be measured and attributed in the same way as before the iOS 14.5 rollout, but for those who don’t, working with SKAN is crucial, and requires a completely different mindset and approach to how we gather and process information about users.

Part 2

What do these privacy changes mean for marketers and the industry as a whole?

From an industry perspective, these changes represent an ongoing shift toward an ecosystem centered around user privacy. Apple states that privacy is a fundamental human right, and they design their products based on this principle. Adjust wholeheartedly supports this framework and is working together with Apple, our clients, our partners, and the industry at large to ensure complete user privacy and respect for people’s decisions regarding their data.

What this means for marketers, however, is a potential lack of precision and accuracy in user acquisition (UA), attribution, and campaign performance. This is simply because advertisers have less visibility into user level data and campaign metrics. UA with the IDFA relies on precise campaign data that gives marketers visibility over performance, empowering decisions around which channels to invest their budget in, and how to optimize performance. There are usually specific KPIs like Day 0/Day 1 retention, user LTV, ROI, and ROAS. Marketers working with this model often operate within thin margins regarding where they invest budgets, And how to scale may come down to a matter of a few percentage points. Before SKAN 4, this level of insight was prevented simply because the aforementioned KPIs weren’t supported. This makes understanding which channels to focus on complex in an ecosystem where mobile performance marketing had previously made it relatively automated and straightforward, and instead pivots attention onto how essential strategically-minded UA managers are post-IDFA. UA has always been a central pillar in mobile marketing but the user-privacy driven shakeup has meant that old ways of doing things needed to be completely revisited, and that remaining agile and flexible is more essential than ever. We can no longer solely rely on automation to shift the percentage points in our favor.

Advertisers who get high opt-in rates are still able to have a good understanding of their KPIs per campaign and can optimize efficiently, as all tools and means of measurement they’re familiar with can be leveraged. Also, the better your opt-in rates, the more data you have to work with to more comprehensively understand and strategize regarding non-consented data. On that note, we’re happy to report that consent rates are consistently climbing as marketers are able to communicate the benefits of opting-in.

The rest of the inventory that is tracked via SKAN 2 made things more complicated to measure and manage. Let’s take a look at how it impacted the three most common monetization models: in-app advertising, in-app purchases, and subscriptions.

  • In-app advertising

    There are two main types of ads, contextual and targeted. Contextual, or non-targeted ads, are less specific, while targeted ads are served to specific users — based on their IDFA. They’re more expensive, but they offer higher conversion rates, making them the more popular choice generally speaking. Naturally, on iOS 14.5+, if a user hasn’t opted in, their IDFA is not available, and they can’t be served targeted ads. This is why some users have noted seeing more ads since opting out. Many publishers are serving an increased number of ads to make up for revenue lost from lack of ability to serve targeted advertising.

  • In-app purchases & subscriptions

    Users pay for in-app goods or services in a wide variety of apps, such as when buying gold or coins in a gaming app, unlocking premium features in a health and fitness or productivity app, or signing up for a subscription to a service. Here, iOS 14.5+ doesn’t have a direct impact on revenue, but the lack of attribution for opted-out users makes it difficult for advertisers to measure the success of campaigns, or to understand where high LTV and top-performing users are coming from.

The last two and a half years have proven that advertising has not stopped on iOS, and it won’t stop. There will always be users to acquire. This is why we have built support for ATT and SKAN, and we are happy to offer solutions that focus on campaign optimization, actionable metrics, and which empower our clients to continue to focus on growth.

Privacy and data protection are central to the mobile advertising industry, and Adjust recommends that advertisers pursue a holistic approach to acquisition, attribution, and measurement on iOS 14.5+, leveraging the ATT framework to obtain as much user-level, consent-based data as possible to use in conjunction with aggregated SKAN data.

Part 3

AppTracking Transparency, user consent, and the opt-in

As we covered above, it's the lack of access to the IDFA that causes the headache when it comes to attribution and campaign measurement on iOS 14.5+. Apple’s ATT is the framework that allows ad targeting and measurement once a user has opted-in or granted consent to tracking. The more users you have consenting, the bigger your pool of efficiently measurable, granular data will be. Moreover, these users can receive targeted advertising and the data can be used to inform and continually improve strategies for conversion values, predictive analytics, and SKAN campaigns.

The most accurate data in a post-iOS 14 world starts with a robust opt-in strategy—prioritizing user consent is not the only pillar of success, but it is key. It not only minimizes disruption, it delivers significant competitive advantages. While the more users you have consented the better, even a relatively low opt-in rate or percentage of opted-in users can prove pivotal to success.

With a higher opt-in rate, marketers are able to receive more deterministic data points—and it’s these data points that then inform the broader strategy for both opted-in and non opted-in, aggregated data sets. They also help to inform pLTV. This is essential for teams wanting to work with predictive KPIs like pLTV, which helps anticipate growth opportunities within aggregated collections of SKAdNetwork and consented user data. The more deterministic data insights that can be gained, the more robust the pLTV results will be.

When working with our clients to help develop top opt-in strategies and flows and to incorporate optimizing for the opt-in as part of an overall UX strategy, there are a few top themes we identified. These are the variables that have the biggest impact on influencing a user’s response to an ATT pop-up.

  • Location: Pinpointing the precise moment to serve the opt-in request is the top factor to define. We’ve identified that it is usually best to display the prompt during the onboarding flow.
  • Messaging: Pre-permission prompts that cue the ATT framework prompt mean the Apple pop-up will feel more natural. Use 2-3 short sentences to emphasize the benefits of opting-in. Also, although the first sentence in the ATT prompt itself can’t be customized, we highly recommend adapting the second string. This is another opportunity to address any concerns about data privacy, and to highlight the benefits of opting in.
  • Size: Users typically respond to pre-permission prompts that are full screen, as opposed to modals. This is because the former offers a more seamless user experience, while the latter can feel like an interruption.
  • Button placement: The placement of your call-to-action (CTA) button is an easy lift. Buttons with simple text, placed next to each other horizontally, with the positive acceptance on the right hand side, yield the best results.

We’re also happy to announce that opt-in rates are slowly but steadily increasing overall. Around the initial rollout of iOS 14.5 the global average sat at around 24%, as of Q4 2022, this number had increased to 28%.

In addition to building a fully optimized opt-in strategy, it’s also crucial to define an overall UX strategy that the opt-in is simply part of, and to continually test. Once a UX approach regarding when, where, and how you prompt users to give consent is designed and integrated, the work isn’t quite done. That’s why we advocate for continued, rigorous testing, including A/B testing and randomized controlled experiments.

Part 4

iOS 14.5+ and working with SKAdNetwork

For users who don’t opt-in, Apple’s SKAdNetwork, a combination of SDK functions and API calls, is the solution that marketers can turn to for attribution of app installs and reinstalls. Free for advertisers to use, Apple’s goal with SKAN is to provide basic, privacy-focused attribution, and it doesn’t require user consent because Apple designed it within their privacy guidelines. The attribution information from SKAN is relayed from the device, to Apple, and then to ad networks, developers, and mobile measurement partners (MMPs) like Adjust. Clients can implement SKAN on their own, or with Adjust’s help. It’s important to note that even when integrated with Adjust, it’s Apple that performs the attribution in this scenario, while Adjust takes care of aggregating the data.

Up until SKAN 3, SKAN provided space for 6-bits of downstream metrics, a number between 0 and 63 (or between 000000 and 111111 in binary), with an initial 24-hour timer. Otherwise known as a conversion value (and considered the fine grain conversion value on SKAN 4) this can be assigned to any value that can be expressed in binary, and it’s up to apps to decide which events they want to include. Every time the conversion value is updated to a fresh six-bit code defined within the app (up until SKAN 3), the timer is extended by an additional 24 hours. Once this conversion value-window expires, a second 24-hour window is triggered for attribution. The idea behind this is to obfuscate the time of install, making it impossible to link event triggers to individual users. This data is then shared by the SKAN 3 in the aggregate, with no granular, user-level data accessible.

Put a little more simply, conversion values are a number between 0-63, used for measuring up to six events using the bit logic. Each conversion value is linked to specific conditions, which can then be unpacked into meaningful KPIs for reporting.

To make the most of this system, advertisers and marketers need(ed) to work thoroughly within the first 24 hours, leveraging all data possible to paint a clear picture of user behavior, from which projections and segments can be identified. It’s not just about acquisition, it’s about understanding what user behaviors in that first 24-hour window can tell us about what they might do later. For many apps, this meant a complete strategic rewriting regarding events tracking as they worked to identify and predict later-stage events. For example, a Day 7 subscription event could be tied to an earlier event by calculating the likelihood via predictive analysis.

How a conversion value strategy is set up is key to success on SKAN. We’ll get into how Adjust supports clients in understanding, building, and mapping conversion values (both event-based and value-based) below, and how marketers and developers can identify which solution is suitable for their app.

Part 5

A few more changes with iOS 15

The huge changes to attribution and tracking of post-install events on iOS came with the announcement of iOS 14 and the release of iOS 14.5 and ATT. And while it was not nearly as industry-defining as its predecessor, iOS 15 (released in September 2021) presented a few key changes and further developments in Apple’s steps toward further user privacy.

As of release, copies of “winning” SKAN postbacks (the attribution call when an app install occurs) can also now be sent to developers, not only to ad networks. This change was welcomed by the industry and understood to be providing developers with more transparency over their data, giving access from install to post-install.

Private Relay was one of the most interesting announcements from a mobile advertising perspective. Designed to prevent the tracking of iOS users on Safari, Private Relay in its current form redirects all network traffic through its servers and hides the user IP address, meaning they can’t be tracked and that user profiles can’t be built. Some rumors indicated that Apple would switch Private Relay on by default from iOS 16 onward, meaning that all SDK traffic will be routed through it, a move that would further complicate the process of gaining insight into users and campaign performance. So far, however, this hasn’t happened. Also, with SKAN 4, Apple now offers web-to-app attribution for Safari users.

Mail Privacy Protection was also rolled out, a feature that allows Apple Mail users to hide their IP address and location, and to anonymize the tracking of email opens. If these options are toggled on, access to information like email opens is no longer available to marketers. Hide My Email is another privacy-related feature that will allow users to mask their email with a generated one when completing any online form.

Aimed at helping to improve UA strategies and campaign performance, another interesting iOS 15 feature is Custom Product Pages and Product Page Optimization. Marketers and developers create custom pages on the app store that can be targeted and optimized for up to 35 individual user segments and then linked to relevant UA campaigns.

Part 6

SKAN 4, iOS 16.1.1, and how it works now

Released in October 2022, SKAN 4 introduces a bunch of new features and options that will change the iOS measurement and attribution space considerably throughout 2023. These updates provide significant opportunity to ad networks and app developers but also come with new complexity and mean that strategies and setups for conversion values and SKAN campaigns need to be reassessed.

In short, the key SKAN 4 feature updates are as follows:

  • Three postbacks instead of one: This is the huge one for mobile marketers. Instead of one postback with a 24-hour measurement window, there are now three. Postback 1 covers a 0-2 day post-install measurement window (with a 24-48 hour postback timer), Postback 2 covers day 3-7, and Postback 3 covers day 8-35, both of which have 24-144 hour postback timers.

    Img

  • Conversion value grains and coarse conversion values: There are now two types of conversion values: fine grain (the familiar 0-63 values from SKAN 3) and coarse grain conversion values, which assign a user as low, medium, or high. Fine grain values can only be received in Postback 1, while coarse grain values can be received in any of the postbacks.

  • Crowd anonymity tiers: Thanks to the new tiers, it’s now possible that campaigns with lower volume might receive some attribution information/a coarse grain value, potentially minimizing the number of “null” conversion values received (which has been a persistent problem on SKAN 3). Installs assigned to tier 1, 2, or 3 crowd anonymity will receive all three postbacks (with varying degrees of granularity), while tier 0 will only receive Postback 1 with a bare minimum of information. 
    Img

  • Source identifier: Referred to as campaign ID on SKAN 3 and only representing 2 digits, SKAN 4 allows up to 4 digits and has renamed this to source ID.

  • lockWindow: This allows you to apply a lock, the lockWindow at any point in each of the measurement windows to determine when the random timer for the postback will start, potentially reducing the waiting time to receive it. See below, where the lock is applied in the second measurement window at around day 5, meaning the random timer will start then instead of at day 7. This does mean, however, that no data following the lock, within the same window, will be measured.
    Img

  • Web-to-app attribution: It’s now possible to attribute from Safari advertising that directs to App Store product pages.

While SKAN 4 is already out, industry-adoption will take some time, meaning there’s still ample opportunity to design and plan your strategy to fully leverage the new capabilities later in 2023. For now, we recommend thinking about how you will assign coarse grain conversion values and how you can use them to gain later-stage signals that will help you build predictions for LTV and early-stage insights.

Part 7

Adjust’s solutions: How we’re supporting these changes

Adjust’s goal in working with clients on iOS 14.5+ though to iOS 16.1.1 and beyond is to empower data-driven decision-making. We want our clients to comprehensively understand and have confidence in their marketing strategies on SKAN and working on iOS in general. We enable our clients and partners to continue leveraging powerful device-level data without violating end-users’ privacy with our three-pronged approach to the iOS 14.5+ ecosystem.

Since the initial announcement of the privacy changes in June 2020, Adjust has continually taken steps to ensure minimal disruption for our clients, who we empower to thrive in their iOS measurement efforts via our three-pillared measurement framework:

  1. Convert: Users to consented users
    Your most accurate data starts with a robust opt-in strategy. Maximize user consent with our research-driven opt-in best practices to minimize disruption and gain a competitive advantage by continuing to receive user-level data with the AppTrackingTransparency framework.
  2. Collect: Insights with transparency
    You can collect free, aggregated SKAN data with our extensive network of integrated partners, complemented by flexible and transparent conversion value mapping with Conversion Hub.
  3. Compute: Data points into actionable insights
    By leveraging both deterministic and aggregated data, our conversion model computes those data points to accurately project performance for non-consented data and KPIs.

Prioritizing a strategy to maximize user consent is key to post-iOS 14.5+ success because it not only minimizes disruptions, it makes long-term strategizing and allocation of budget more straightforward. With your consented data, you can continue to make marketing decisions with confidence. And while the more consented users you have the better, even a relatively low opt-in rate can prove crucial to success. This is why we stress the importance of a strong opt-in strategy—and the data points you can gather are key to delivering performance results from your non-consented users.

Any model used for forecasting or projection is only as powerful and precise as the data you feed it. If you feed your model only a few aggregated, anonymized data points, your model will be inherently less accurate than if you fed it defined user-level data.

For full SKAN 4 support, make sure you have our latest SDK (SDK 4.33.0, available here.) in your app. Our SDK has always been open-source for ultimate transparency.

Part 8

Adjust’s solutions: Conversion Hub, SKAN 4, and beyond

As we’ve covered extensively in this guide, conversion value strategies that make the most of SKAN are essential for success on iOS. This is why we launched Conversion Hub in September 2022, our smart, easy-to-use, conversion value solution that empowers clients, via machine learning, to map and configure app-specific conversion value setups. Conversion Hub is an in-house SKAN expert for clients of all sizes. We are currently working on updating Conversion Hub to maximize coarse grain values and lockWindow. By gathering insights and working closely with clients and partners, we will offer a robust update in-line with our mission to bring next-generation solutions that reflect industry needs to the market.

You can already make the most of your SKAN and iOS campaigns with Conversion Hub, which functions with two conversion value modes and setups designed to make fine conversion value modeling and mapping as easy and effective as possible.

1. Conversion Hub’s fine conversion value modes:

  • 63 CVs mode: Makes use of all 63 possible conversion values by mapping a revenue range to each value.
  • 6-bit mode: Provides maximum transparency on post-install engagements and empowers user journey optimization. Up to six events can be chosen and mapped to the six bits available.

2. Conversion Hub’s setup flows:

  • Smart setup: Designed for marketing teams seeking inspiration and guidance on the conversion value model that will work best for their app, vertical, and business model. Let Adjust's machine-learning algorithm suggest ready-to-go conversion value mapping based on your monetization efforts and vertical-based insights.
  • Advanced setup: Perfect for experienced teams who already have a fine conversion value model in mind. The most customizability and granularity is empowered by setting up the 63 CVs or 6 bit mode within the advanced setup flow.

Conclusion

The changes brought about by Apple’s privacy updates and rollout of iOS 14.5+ have been industry defining, but they have not been as severe for user acquisition or attribution as many people expected. As an industry leader, we understand that iOS and Apple (as well as our entire ecosystem) are constantly evolving. We work hard to stay well-connected with our partners and clients so we can remain at the forefront of changes and developments.

From day one, our approach has been to ensure that we serve our clients with robust iOS 14.5+ and SKAdNetwork solutions, that we empower marketers to make data-driven decisions with confidence, and that we uphold the principles of user privacy and data transparency. This philosophy focuses on continuing to leverage powerful user-level data while ensuring end-users’ privacy.

The future of measurement will leverage aggregated data driven by machine learning, just as Conversion Hub is already demonstrating. We understand that privacy underpins our industry, and as innovation progresses and evolves, so too will the specifics and technical requirements of user privacy. We place the best interests of our users and our customers at the center of the products and solutions that we develop, and will continue to do so as we optimize for SKAN 4, iOS 16.1.1, and beyond. You can already start planning for how you will assign coarse conversion values, and moving further into 2023, you’ll be able to plan how you’ll predict LTV as we fully leverage SKAN 4’s capabilities together. Learn more about what we’re calling the 4.0 milestones of SKAN 4.0 here.

To find out more about iOS 14.5+, to keep up-to-date with the latest information and developments, or to receive advice tailored to your specific app or business needs, you can request a demo to see how the Adjust solution works in practice, or browse our iOS 14.5+ resource center for an overview of all our guides and information.

Want to get the latest from Adjust?

Keep reading

Trusted by over 100,000 apps worldwide