justtrack: SKAdNetwork (SKAN 4) is now table-stakes iOS attribution, so treat conversion values as product instrumentation

Original article (source): justtrack - “SKAdNetwork and SKAN 4, Explained for Gaming UA” (published July 29, 2026)

• https://justtrack.io/blog/skadnetwork-explained/

The punchline

If you are buying users on iOS, SKAdNetwork is no longer “an MMP thing”, it is part of your day job.

The only lever you truly control inside SKAN is what you encode as the conversion value, so designing that schema (and keeping it stable) becomes measurement work, not paperwork.

What they think changed with SKAN 4 (the practical bits)

Their overview is short, but it captures the shape of SKAN 4:

• Multiple postbacks (up to three): you can get an early signal, then later signals over a longer window.
• Coarse + fine conversion values: more detail when volume is high enough, a simpler low/medium/high style signal when it is not.
• Crowd anonymity tiers: Apple only releases granular detail when enough installs protect anonymity.
• More room for campaign structure (source IDs), plus web-to-app support.

The operational takeaway

They frame SKAN as one signal in a blended system:

• SKAN (aggregated and delayed)
• deterministic measurement from consenting users
• modelling to fill gaps

Whether you agree with their product framing or not, the useful mindset is: stop treating SKAN configuration as “set and forget”. If your conversion value stops reflecting early value, your optimisation loops drift.

Tiny win

Pick one paid channel and do a 30-minute SKAN hygiene check:

1. Write down your current conversion-value mapping (what does each value mean?).
2. Verify the mapping still matches your real activation and early value events.
3. Add a change-log note in your analytics docs so you can annotate performance shifts.

Read the original: https://justtrack.io/blog/skadnetwork-explained/