Table of Contents
SKAdNetwork conversion values are the small, privacy-limited signals an iOS app encodes after an install to tell an ad network what a new user did, without revealing who that user is. In practical terms, a conversion value is the only post-install data point Admiral Media gets back from a SKAdNetwork campaign, so the schema you choose to encode behind it decides whether your bidding algorithm learns to find paying users or wastes budget chasing installs. This guide explains how SKAdNetwork conversion values work under SKAN 4, how to design a schema that actually optimizes, and what Admiral Media has learned designing measurement for app campaigns since the arrival of App Tracking Transparency.
SKAdNetwork (often shortened to SKAN) is Apple’s privacy-preserving attribution framework. It replaces device-level tracking with aggregated, delayed, and deliberately coarse postbacks. The conversion value sits at the center of that trade: it is the one field you control, the one piece of behavioral signal that survives Apple’s privacy filtering. Treating it as an afterthought is the most common and most expensive mistake the Admiral Media team sees in iOS measurement.
What are SKAdNetwork conversion values?
A SKAdNetwork conversion value is a coded signal that an app records after install to describe early user behavior, which Apple then returns to the ad network in an aggregated postback. There are two kinds. A fine conversion value is a 6-bit integer from 0 to 63, giving 64 possible states. A coarse conversion value is one of three buckets, low, medium, or high, introduced in SKAN 4 for situations where install volume is too small to safely return the fine value. An advertiser receives one or the other, never both, and only when Apple’s privacy threshold is met.
The reason this matters is mechanical, not philosophical. App campaign bidding on Google and Meta optimizes toward the value you feed it. Under the deterministic tracking that existed before App Tracking Transparency, that value was a real in-app event tied to a real device. Under SKAdNetwork, the conversion value is a compressed proxy for that event, filtered through Apple’s privacy layer. The encoding choice, what each number from 0 to 63 means, is the entire information channel. Encode it badly and the algorithm optimizes toward noise.
Apple documents the coarse value states directly in its SKAdNetwork.CoarseConversionValue reference. Admiral Media treats the conversion value as the single most important configuration decision in any iOS user acquisition setup, ahead of bid strategy and audience structure, because every downstream optimization inherits its quality.
How do conversion values work under SKAN 4?
Under SKAN 4, conversion values work across three separate postback windows, each with its own value and its own delay, which together extend measurement to 35 days after install. This was the headline change from earlier versions, which offered a single postback and a much shorter measurement horizon.
The three windows are fixed by Apple: postback one covers days 0 to 2 after install, postback two covers days 3 to 7, and postback three covers days 8 to 35. Apple describes this multi-window behavior in its documentation on receiving postbacks in multiple conversion windows. The critical constraint: the fine 0 to 63 value is only available in the first postback, and only when crowd anonymity allows it. Windows two and three return coarse values at best. So the richest signal you will ever get arrives in the first 48 hours, which forces a hard design question about how much of a user’s value can be predicted that early.
Crowd anonymity is Apple’s privacy throttle. It defines four tiers, numbered 0 to 3, based on how many installs a campaign drives. Low volume sits in the lowest tier, where Apple may strip the fine value, the coarse value, the source identifier, or the source app identifier. As volume grows and a single user’s behavior blends into a larger crowd, Apple releases more granular data. The practical consequence is that fragmented campaign structures, many small campaigns each below the anonymity threshold, systematically lose signal. Consolidation is therefore a measurement requirement under SKAN, not merely an efficiency tactic.
One more lever shapes timing. The lockWindow parameter, documented by Apple in updatePostbackConversionValue, lets an app declare that it is finished updating a value and trigger the postback timer early. Locking sooner gets data back faster for quicker optimization decisions, at the cost of observing less of the user’s behavior. That is a genuine trade, and the right answer depends on how fast your app reveals user value.
The Admiral Media Signal Encoding Ladder
Designing a conversion value schema is a structured decision, not a guess. The Admiral Media Signal Encoding Ladder is the framework the Admiral Media team uses to turn 64 abstract integers into a signal that trains bidding algorithms toward revenue. It moves from business goal down to implementation, so the schema serves the economics rather than the other way around.
- Anchor on the value you actually monetize. Before touching a single bit, define the outcome the app makes money from: a subscription start, a deposit, a verified lead, a first purchase. The conversion value exists to predict that outcome early. If the schema is not anchored to revenue, every later step optimizes toward a vanity event.
- Map what is observable inside the first 48 hours. Because the fine value lives only in postback one, list the events a real user can plausibly complete in days 0 to 2: onboarding milestones, a trial start, a tutorial completion, an add-to-cart. The ceiling on early signal is whatever happens in that window, so the product funnel, not the marketer, sets the limit.
- Choose an encoding model: revenue buckets, event milestones, or a predicted-value proxy. The 0 to 63 range can represent banded early revenue, an ordered sequence of funnel events, or a modeled prediction of longer-term value compressed into the available bits. Each model suits a different app economy, and picking the wrong one is the difference between a schema that scales and one that plateaus.
- Reserve coarse-value fallbacks for low-volume campaigns. Map low, medium, and high to meaningful value tiers so that campaigns sitting below the crowd anonymity threshold still return usable signal. A schema that only works when the fine value is available is a schema that fails on exactly the new campaigns that need direction most.
- Set the lock window to your value-revelation speed. If most monetization signal lands by day one, lock early and feed the algorithm faster. If value reveals itself over a week, hold the window open. The Admiral Media team tunes this per app rather than applying a single default across a portfolio.
- Validate the schema against incrementality, then iterate. A conversion value schema can look healthy in-platform while measuring the wrong thing. Admiral Media pressure-tests schemas against incrementality and cohort outcomes, then revises the mapping, because the first schema is a hypothesis, not a finished system.
The ladder is deliberately ordered top down. Most failed SKAN setups invert it, starting from whatever events the SDK happens to fire and reverse-engineering a meaning afterward. That produces a schema that is easy to implement and useless to optimize against.
Fine or coarse: which conversion value should you optimize for?
You optimize for the fine value where you can earn it and design the coarse value as a deliberate fallback, never as an accident. The two are not interchangeable, and the choice is dictated by volume and crowd anonymity rather than preference. The table below summarizes how Admiral Media weighs them.
| Dimension | Fine value (0 to 63) | Coarse value (low / medium / high) |
|---|---|---|
| Granularity | 64 states, enough for revenue bands or detailed funnels | 3 states, broad tiers only |
| When it is returned | Postback 1 only, and only above the crowd anonymity threshold | Available when volume is too low for the fine value |
| Volume dependency | High: needs campaigns large enough to clear anonymity tiers | Low: designed for smaller campaigns |
| Best use | Scaled campaigns where early revenue signal trains bidding | New, niche, or low-volume campaigns that still need direction |
| Main risk | Signal collapses if campaigns are fragmented below threshold | Too coarse to separate high-value users from average ones |
The strategic read is that fine and coarse values are two ends of one schema, not two options. A schema built only for the fine value goes blind on every campaign that has not yet scaled. A schema built only on coarse values throws away the resolution that makes bidding sharp on the campaigns that matter most. Admiral Media designs both layers from the start so signal degrades gracefully instead of disappearing.
How do you design a conversion value schema that actually optimizes?
You design a schema by choosing an encoding model that matches how your app makes money, then mapping it so the highest values mark the users worth paying most to acquire. There are three workable models, and the right one depends on the app economy.
Revenue bucketing assigns ranges of the 0 to 63 scale to bands of early revenue. It suits apps where users spend real money in the first days, such as ecommerce or games with early in-app purchases. The strength is a direct line from conversion value to cash. The weakness is that apps with delayed monetization, most subscription products, see little day-zero revenue to bucket.
Event milestone encoding assigns values to an ordered sequence of funnel steps: registration, onboarding complete, trial start, paywall view, subscription. It suits subscription and lead-generation apps where the path to revenue is a chain of predictive actions rather than an immediate purchase. The Clark case study is instructive here. Working with Clark, the Admiral Media team found that optimizing toward carefully chosen mid-funnel events, rather than only the obvious lower-funnel goal, lifted performance, because users explore an app before committing. That same logic governs a good milestone schema: the values that predict revenue are often upstream of revenue itself.
Predicted-value proxy encoding compresses a model’s estimate of a user’s longer-term value into the available bits, so the conversion value carries a forecast rather than a raw event. It suits mature apps with enough historical cohort data to model lifetime value credibly. It is the most powerful model and the most demanding, because a weak prediction encodes noise with false confidence.
Web-to-app is the quiet fourth option, and it sidesteps SKAdNetwork entirely. When acquisition runs through a web landing step before the App Store, conversions can be measured server-side without depending on a SKAN postback at all. In Admiral Media’s work with Miles Mobility, a web-to-app structure on Google paired with a mobile measurement partner produced cleaner attribution than in-app SKAN signal alone could.
What do Admiral Media’s campaigns reveal about measuring under signal loss?
Admiral Media’s campaigns show that disciplined measurement design, not just more spend, is what scales iOS apps after the loss of deterministic tracking. The conversion value schema is one piece of a larger measurement architecture that has to hold together: clean event definitions, consolidated campaign structure, and an independent read on incrementality.
When App Tracking Transparency reshaped attribution, the immediate effect on live accounts was a flood of conversions with unidentified sources. In Admiral Media’s work with TIER, the mobility brand, the team adjusted strategy and reallocated budget so the account still understood where value came from, then scaled user acquisition: TIER reached a 297% increase in new customers, expanded into two new channels beyond its original platform, and scaled its acquisition budget fivefold in under three months. The client specifically credited Admiral Media’s support on tracking and incrementality, which is the discipline that conversion value design depends on.
The Clark engagement shows the same principle applied to event strategy. By treating which events to optimize toward as a deliberate choice rather than a default, the Admiral Media team improved upper, middle, and lower funnel performance at once. Comparing month three to month one, Clark saw cost per lead fall by 50%, cost per install drop by 29%, conversion rate rise by 41%, and installs grow by 18%. Those gains came from choosing the right signal to chase, the exact judgment a SKAN conversion value schema forces you to make.
Miles Mobility points to the structural option of stepping outside SKAdNetwork where the funnel allows it. By moving acquisition through a web-to-app structure on Google with a mobile measurement partner for attribution, the Admiral Media team helped Miles Mobility reach 260% more conversions at a 25% lower cost per acquisition. The lesson for conversion value strategy is that SKAN is one measurement path, not the only one, and a strong iOS plan decides deliberately when to encode signal through SKAdNetwork and when to route around it.
Across these accounts the pattern is consistent. The brands that kept scaling after signal loss did so because measurement was designed first, with the conversion value schema treated as core infrastructure. This sits alongside Admiral Media’s broader iOS measurement work, including its AdAttributionKit measurement playbook and its approach to incrementality testing for mobile apps, which together form the truth layer that validates whatever a SKAN postback reports. For the re-engagement side of the same problem, the team’s post-ATT app retargeting framework applies the same addressability-first logic.
Apple’s direction of travel reinforces the point. AdAttributionKit, documented in Apple’s developer reference, is the successor framework that carries conversion values forward while adding capabilities such as re-engagement attribution. The schema thinking in this guide transfers directly, which is why Admiral Media designs conversion value logic to be framework-portable rather than locked to a single SKAN version. Managing more than 500 million euros in ad spend across 150-plus mobile brands, the Admiral Media team has had to make these decisions at scale rather than in theory.
Frequently Asked Questions
What is a SKAdNetwork conversion value in simple terms?
A SKAdNetwork conversion value is a short coded signal an iOS app records after install to describe what a new user did, which Apple returns to the ad network in a privacy-protected postback. It can be a fine value, a 6-bit number from 0 to 63, or a coarse value of low, medium, or high. Because Apple strips out device-level identifiers, this value is usually the only behavioral signal an advertiser receives from an iOS campaign. How you define what each value means determines how well bidding algorithms can optimize.
What is the difference between fine and coarse conversion values?
The fine conversion value is a 6-bit integer from 0 to 63, offering 64 possible states for detailed signal such as revenue bands or funnel stages. The coarse conversion value, added in SKAN 4, has only three states, low, medium, and high, and is returned when a campaign’s install volume is too low to meet Apple’s crowd anonymity threshold for the fine value. Advertisers receive one or the other, not both. The fine value is also limited to the first postback window.
How many conversion windows does SKAdNetwork 4 have?
SKAdNetwork 4 has three conversion windows that together measure up to 35 days after install. The first window covers days 0 to 2, the second covers days 3 to 7, and the third covers days 8 to 35. Each window can produce its own postback with its own conversion value. The detailed fine value of 0 to 63 is only available in the first window, so the richest signal arrives in the first 48 hours.
What is crowd anonymity in SKAdNetwork?
Crowd anonymity is Apple’s privacy mechanism that decides how much data a SKAdNetwork postback can contain, based on install volume. It defines four tiers, from 0 to 3, where low-volume campaigns sit in the lowest tier and may have the fine value, coarse value, source identifier, or source app identifier removed. As volume rises and an individual user blends into a larger group, Apple releases more granular data. This is why consolidating small campaigns often recovers lost signal.
How do you choose a conversion value schema for a subscription app?
For a subscription app, event milestone encoding usually works best, because most subscription revenue arrives after the SKAN measurement window rather than on day zero. The schema assigns conversion values to predictive early actions such as registration, onboarding completion, trial start, and paywall engagement. The goal is to encode the early behaviors that correlate with later paid conversion, since the conversion value has to predict revenue that the postback window will never directly see. Validating the schema against cohort outcomes and incrementality keeps it honest.
Does AdAttributionKit replace SKAdNetwork conversion values?
AdAttributionKit is Apple’s successor framework to SKAdNetwork, and it carries the conversion value concept forward while adding capabilities such as re-engagement attribution and interoperability with existing SKAN. The schema design principles, anchoring values to revenue, encoding observable early behavior, and planning for coarse fallbacks, transfer directly. Admiral Media designs conversion value logic to be portable across both frameworks so measurement does not break as Apple’s attribution stack evolves.
Can you measure iOS campaigns without relying on conversion values?
Yes, in some cases. Web-to-app campaigns route users through a web step before the App Store, which lets conversions be measured server-side without depending on a SKAdNetwork postback. Incrementality testing provides an independent read on whether ad spend is driving real lift, regardless of what any single postback reports. In Admiral Media’s experience, the strongest iOS measurement combines a well-designed conversion value schema with these complementary methods rather than treating SKAN as the only source of truth.
