SKAN 4 Conversion Schema: How to Design One That Actually Works (Step-by-Step Guide)

Apple’s SKAdNetwork 4 made attribution more powerful but also more complex. Marketers are asking questions like “What events should I map to SKAN 4?”, “Should I use fine or coarse values?”, and “What if my campaign volume is too low?”.

This guide answers those questions step by step, with examples and insights you won’t find in the standard documentation.

Step 1: Audit Your Business Metrics

What exactly is a conversion schema and why does it matter?

A conversion schema is the rulebook that maps in-app events to Apple’s SKAN postback windows. It matters because it’s the only way your raw user behavior becomes actionable signal. Without it, you’re stuck with suppressed postbacks or meaningless null values¹.

Before you even open a tool, define what success looks like for your app. For a game, that may be tutorial completions and in-app purchases. For a subscription app, it could be free trial starts or early retention. Use historical data to find the early events most predictive of long-term value.

👉 If install volume is low, simplify. Coarse values are safer than splitting into fine buckets. Fine values only return if Apple’s privacy thresholds are met, which we’ll cover in detail later in this guide.

Step 2: Understand the SKAN 4 Windows

How do the three postback windows actually work?

SKAN 4 sends up to three postbacks per install:

• Day 0–2: fine or coarse values (best for onboarding and high-volume early events).
• Day 3–7: coarse only (ideal for first purchase or trial starts).
• Day 8–35: coarse only (good for long-term signals like revenue bands or renewals).
  

There is also a lock window option, which ends measurement early to deliver results faster. The trade-off? You may miss late events. It’s best if your key milestones happen within the first few days².

Step 3: Map Events to Windows

Which events should I map to each window?

Think about what matters at each stage of the user journey:

👉 Google stresses mapping events that happen at meaningful volume. A clever schema balances predictive early actions with long-term confirmation³.

Step 4: Fine vs Coarse Values

What are fine and coarse values in SKAN 4?

• Fine values: 64 possible buckets (0–63) available only in the first postback window (Day 0–2). These allow detailed event or revenue mapping, but only work if you have enough installs to fill buckets consistently.
• Coarse values: Three possible tiers (low, medium, high) available across all three windows. These are simpler, more resilient to low volume, and less likely to return nulls.

Should I use fine values, coarse values, or both?

• Fine values give granularity but need scale.
• Coarse values are safer across windows and reduce risk of null postbacks.
• Always normalize so that “high” consistently represents your most valuable users, regardless of which window it appears in.

Step 5: Test, Validate, Iterate

What if my campaign volume is too low?

Campaigns under ~100 installs per day often fail Apple’s privacy thresholds, leading to null values. Small “test campaigns” can create a false sense of efficiency. Budget enough to cross thresholds so the data you collect is actually usable².

How do I know if my schema is working?

Run tests, track null rates, and compare SKAN signals to internal analytics. If fine values are noisy, simplify. If certain early actions (like day 3 return) strongly predict long-term ROAS, prioritize them in your schema. Hybrid measurement combines SKAN with probabilistic modeling and MMP dashboards and is the safest way to validate your setup^1,3. Learn more about iOS attribution solutions here.

Our Take: Qualifiers Beat Purchases Early

Many guides focus on revenue events, but in practice early engagement qualifiers can be more predictive. For example, a game may find that “tutorial complete + day 3 return” users monetize better than those who only make a tiny early purchase. Mapping those qualifiers in Day 0–2 fine values can give you stronger signals faster.

Final Checklist

1. Identify KPIs and predictive events.
2. Map events to SKAN windows based on volume and timing.
3. Balance fine and coarse values strategically.
4. Budget to cross thresholds.
5. Validate SKAN signals against other measurement frameworks.
6. Iterate as user behavior and Apple’s rules evolve.
   

The Bottom Line

A SKAN 4 conversion schema is not about filling 64 buckets. It’s about building a predictive map of your user journey that clears privacy thresholds and ties back to business KPIs. Treat schema design as a living strategy and you’ll be ready not just for SKAN 4, but also for SKAN 5 (expected in 2026). Marketers who stress-test schemas now will be ready when SKAN 5 brings faster postbacks and built-in incrementality.

Future-proof your iOS growth before SKAN 5 arrives. Book a free strategy session today!

1. AppsFlyer, SKAN Conversion Studio (2024) ↩︎
2. Adjust, Set Up SKAN 4 Mapping (2024) ↩︎
3. Google, SKAN Schema Setup (2024) ↩︎