Skip to content
AppDues
Log in
Contact salesSign up

SKAdNetwork and iOS 14 privacy changes

Apple's WWDC announcement reset how iOS apps can measure paid acquisition. The IDFA became permission-gated, fingerprinting was pushed behind the same wall, and SKAdNetwork became the sanctioned path for aggregate install attribution. This guide explains how the pieces fit, where the measurement gaps are, and how to plan spend while the ecosystem is still moving.

Why app businesses should care

Paid user acquisition is one of the main engines behind App Store growth. Search, editorial features, press, and word of mouth all feed on attention, and buying attention is the most direct way to feed that loop. Much of the tooling behind that loop was built on the IDFA, which Apple introduced with iOS 6 in 2012 and which let developers and their SDKs identify individual users for targeting, measurement, and optimization.

The scale of that spend is easy to underestimate. TikTok is one of the most viral apps of recent years, yet its parent company Bytedance reportedly spent over $1 billion a year on paid user acquisition and other marketing, according to a Bytedance insider cited at the time. When Apple made IDFA access conditional on an explicit user permission, it changed the inputs to that machinery. The knock-on effects reach beyond ad networks to any app whose growth plan depends on install ads. Apple introduced SKAdNetwork as a partial replacement, but its design trades precision for privacy. Understanding that trade is the difference between a plan that survives the transition and one that has to be rebuilt mid-quarter.

The IDFA opt-in math

The permission prompt is controlled by iOS itself. Developers can customize only a small piece of the surrounding text, which limits the kind of dark-pattern persuasion that grew up around GDPR consent screens.

A TapResearch survey cited at the time suggested opt-in would land well below 30%. Even a generous 50% opt-in rate shrinks further once you account for users who had Limit Ad Tracking enabled on iOS 13 — roughly 30% of users — since those users will not see the prompt at all on iOS 14. The arithmetic is multiplicative, not subtractive: 50% opt-in among the roughly 70% of users who see the prompt yields about 35% of all users in the optimistic case. For tracking to work end to end, the same user also has to opt in both in the app that showed the ad and in the app that was installed from it.

The remaining audience is not a neutral sample. LAT users on iOS are among the biggest spenders on the platform, so the trackable group skews toward lower spenders and is a biased subset of an already biased subset. Data from that group cannot be safely generalized to the rest of the platform. Treat any IDFA-derived signal on iOS 14 as directional at best.

Fingerprinting is not a workaround

When the IDFA is unavailable, some ad-tech vendors combine IP address, device type, local network data, clipboard contents, and other accessible signals into a device fingerprint. Apple has said fingerprinting will also require user permission, and iOS 14 makes the technique harder in practice.

Vendors have claimed fingerprinting accuracy of 70% or higher on iOS. Those claims are difficult to verify and likely overstated, and the accuracy should fall further once Apple's anti-fingerprinting changes take effect. Fingerprinting also only works at scale, which makes large ad networks, MMPs, and similar players the natural focus of Apple's monitoring. Do not build a measurement plan on a technique Apple has explicitly targeted.

How SKAdNetwork fits in

Apple introduced SKAdNetwork in May 2018 as an aggregate attribution channel that does not require a permission prompt. The system tracks clicks at the OS level and reports installs to the ad network. In exchange for losing granular measurement of roughly 70% of advertising — the IDFA opt-in rate at the time — developers get near-complete coverage of campaign-level install data.

The implementation is more involved than the summary suggests. Delivering attribution without exposing individuals requires deliberate constraints, and those constraints are where most of the practical difficulty lives. Still, the direction is sound: attribution on mobile was never as reliable as the ecosystem implied. Limit Ad Tracking, fraud, first-click versus last-click assumptions when a user sees several ads, and unreliable fingerprinting all created noise. A single, privacy-preserving source of truth is a real improvement, provided you plan around its limits.

What a SKAdNetwork postback can and cannot tell you

The practical question for most teams is not how SKAdNetwork works internally but what arrives in a postback and what it supports. Use this map when scoping dashboards or setting expectations with stakeholders:

  • Install attribution at campaign level: available, and coverage is broad because no permission prompt is involved.
  • Campaign count: each ad network is limited to 100 campaigns per app, with no ad groups or creative sets inside those buckets.
  • Who receives the data: Apple exchanges it with the ad network directly. The app and other third parties cannot collect, verify, or act on it.
  • Retargeting signals: a redownload field exists in the postback, but tracking limits make retargeting much harder in practice.
  • View-through attribution: not supported. If an ad impression leads to an install without a trackable click, Apple treats the install as organic.
  • Revenue and ROAS: severely constrained. Conversion Value and Source App ID can be withheld when Apple judges them a privacy risk, and postback timing is tightly limited.

A useful rule of thumb: SKAdNetwork gives you a clearer picture of installs and a much fuzzier picture of return on ad spend. Plan reporting around that split rather than forcing one number to do both jobs.

The five constraints that shape strategy

Campaign limits. The 100-campaign cap per app per network looks generous only if you ignore what it replaces. Ad networks must now do all targeting and optimization inside those buckets. Apple's reasoning is straightforward: unlimited campaign IDs would become a de facto tracking tool, since a network could create a unique campaign ID for each ad that is displayed. Even 1,000 campaigns would make individual identification easier.

Minimal developer control. Because data flows between Apple and the network, developers lose the ability to verify or act on most of the process. That is defensible from a privacy standpoint — fewer recipients means less chance of triangulation — but it concentrates power with ad networks, whose historical record on transparency is mixed.

No retargeting. A user who installs and quickly abandons an app has demonstrated real interest, and re-engaging that person is often the most efficient advertising available. The redownload field is not enough to restore that capability under the new tracking limits.

No view-through attribution. Picture someone who sees a video ad, then searches for the app in the App Store minutes later. Apple counts that install as organic. On some networks, view-through attribution can account for up to 50% of traffic, so this gap is material rather than cosmetic.

Limited measurement. The updateConversionValue mechanism and its timer-based postbacks are the biggest constraint. The postback will theoretically make it through 100% of the time when updateConversionValue is called, but two fields — Conversion Value and Source App ID — will be left off if Apple determines they could violate an individual user's privacy. The result is decent install visibility and poor ROAS visibility.

Fallout and unintended incentives

Ad-tech vendors will have to rebuild their tooling around these constraints. For app teams, the more relevant question is how the shift changes behavior.

Apps that depend on tightly targeted install campaigns are the most exposed. But attribution accuracy was overstated for years, so moving from deterministic to probabilistic measurement may be less disruptive than the headlines suggest. Large networks with deep data, such as Google and Facebook, may still target reasonably well within the new rules. Facebook itself warned that iOS 14 would make it harder for developers to grow using ads on its platform and elsewhere. Another plausible outcome is lower CPMs as targeting and measurement degrade, though that seems less likely while app advertisers keep spending. Facebook and Google are also working on ways around web measurement limits in iOS 14, and a significant share of their ads run inside their own apps, so DTC and other non-app advertising should be less affected than app install ads.

The incentive problem is more subtle. To recover lost signal, more apps may push email or phone logins and use those identifiers for tracking. Apple cannot easily tell whether such data is shared with an ad network for lookalike targeting. Expect continued pressure toward Sign In With Apple as a login option; today it is required only when a developer uses a third-party login such as Facebook Login or Google Sign-In.

A second incentive pushes the other way: call updateConversionValue as early as possible. The call is most reliable in the first session, and waiting for a subscription conversion risks the user never reopening the app in time. Fewer events also make it less likely a postback clears Apple's privacy threshold. That pressure conflicts with subscription businesses that monetize on value delivered over days rather than minutes.

One precedent is worth noting. Developers of kids' apps got a head start on a post-IDFA world after Apple announced stricter tracking limitations for kids' apps at WWDC 2019. Apps that deliver true value still convert and retain customers well, even when the marketing playbook has to change.

Improvements that would make SKAdNetwork usable

Several changes would reduce disruption without weakening privacy:

  • Separate app-open postbacks from Conversion Value postbacks, so installs per campaign arrive clearly and updateConversionValue can fire later when a meaningful event occurs.
  • For subscription apps, allow specific in-app purchases to be registered so a postback can be sent automatically when a free trial converts or is cancelled, even if the user does not open the app that day.
  • Add fields such as Ad Group to the postback. Apple's privacy threshold would still apply, so high-volume apps gain granularity while smaller apps simply see fewer postbacks.
  • Add an Ad Displayed Date field. This is the hardest ask on privacy grounds, but without it a postback cannot be tied to the day the ad ran, which blocks most ROAS estimation.
  • Bring Apple Search Ads into SKAdNetwork so attribution data is standardized in one place.
  • Add accountability mechanisms for ad networks, such as replicating postbacks to MMPs or offering a developer API for aggregate data. At minimum, networks should be required to share SKAdNetwork data with developers.

A practical plan for the transition

The direction of travel is clear even if the details are not: the tools used for paid acquisition on iOS will look different within a season. Two mistakes are worth avoiding. Investing heavily in tooling that is about to change wastes the effort. Building new infrastructure before best practices settle invites rework.

A short checklist for the transition period:

  1. Keep investing in the fundamentals that do not depend on Apple's rules: customer satisfaction, product quality, conversion optimization, ASO, retention, word of mouth, and PR.
  2. If you already have efficient acquisition tooling, consider pulling forward part of the marketing budget and spending while it remains cost-effective.
  3. Reserve capacity to react quickly as ad networks, MMPs, and other vendors ship SKAdNetwork support and SDK updates.
  4. Expect the first months of iOS 14 to feel chaotic as adoption climbs and everyone, Apple included, works out the impact. New strategies and vendors should emerge as the dust settles.
  5. Keep the install-versus-ROAS distinction visible in reporting so stakeholders do not read precision into numbers that cannot support it.

Frequently asked questions

What is SKAdNetwork in plain terms?

It is Apple's system-level attribution channel for app install ads. Clicks are tracked by iOS, and install reports go to the ad network without a permission prompt. You get campaign-level install data, but not user-level tracking.

How many users will opt in to IDFA tracking on iOS 14?

A TapResearch survey cited at the time suggested below 30%. Even a 50% opt-in rate among the roughly 70% of users who see the prompt yields about 35% of all users — 50% × 70% = 35%. The roughly 30% who had Limit Ad Tracking on in iOS 13 will not see the prompt at all. LAT users on iOS are among the biggest spenders, so the trackable group skews toward lower spenders and is a biased subset of an already biased subset.

Can SKAdNetwork measure view-through installs?

No. If a user sees an ad and later installs without a trackable click, Apple counts the install as organic. On some networks, view-through attribution can account for up to 50% of traffic, so this is a meaningful blind spot.

Why is ROAS so hard to measure with SKAdNetwork?

Postback timing is tightly limited, and two fields — Conversion Value and Source App ID — can be withheld when Apple judges them a privacy risk. Installs remain visible; revenue attribution becomes fuzzy. That is why the updateConversionValue timing question matters so much for subscription apps.

Why only 100 campaigns per app per network?

Unlimited campaign IDs would function as a tracking tool, because a network could create a unique campaign ID for each ad that is displayed. Even 1,000 campaigns would make individual identification easier. The cap forces targeting and optimization to happen inside a fixed set of buckets.