What Crowd Anonymity Is
Crowd anonymity is the privacy principle at the center of Apple's SKAdNetwork 4 and AdAttributionKit. The idea is simple: one install from a small campaign in a small country could, combined with other data, point to one person. One install among thousands with the same campaign, publisher and country cannot. So Apple releases more detail in a postback when the crowd behind it is larger.
Apple turns that idea into four postback data tiers, numbered 0 to 3. Each install gets a tier, and the tier decides which fields appear in its SKAdNetwork postback.
The Four Tiers
- Tier 0. A two-digit source identifier and no conversion value. The postback confirms that an install happened and little else.
- Tier 1. A two-digit source identifier plus the coarse conversion value of low, medium or high.
- Tier 2. More source identifier digits, the fine conversion value from 0 to 63 in the first postback, and coarse values in all postbacks.
- Tier 3. The full four-digit source identifier, fine and coarse values, the publisher app ID or source domain, and the country code when that crowd is large enough.
The source identifier is hierarchical. Networks usually encode campaign in the first two digits and use the third and fourth digits for ad group, creative or placement. At low tiers only the first two digits survive, so a network's creative-level reporting depends on reaching tier 3.
What Decides the Tier
Apple does not publish exact thresholds, but it does describe the inputs:
- the crowd size of the publisher app or website that showed the ad,
- the crowd size of the advertised app,
- the country where the install happened, and
- the number of source identifier digits the network registered.
Large publishers, large advertised apps and large countries reach high tiers easily. Small apps running niche campaigns in small markets often see most installs at tier 0 or 1.
How to Reach Higher Tiers
The lever you control is concentration. Every split of the same budget into more campaigns, ad groups, creatives or countries reduces the crowd behind each combination Apple evaluates.
- Consolidate campaigns on iOS. Fewer, larger campaigns per network receive more fine values than many small ones.
- Keep targeting broad where the network allows it, so installs share the same source identifier.
- Agree on digit usage with each network. If you do not need creative-level SKAN reporting, fewer digits mean more installs per combination.
- Lock values early for high-value users. This does not change the tier, but it means the data you do get arrives sooner.
Reading Tiered Data
Always read SKAN results with the tier mix in view. A campaign where 70% of installs are tier 0 will look like it drives no conversions, even when its users convert well. Comparing that campaign to one at tier 3 on conversion value alone is comparing two different levels of visibility.
The SKAN dashboard in Linkrunner includes a privacy tier breakdown, so a campaign review shows how many installs reached each tier before anyone reads the conversion values. For campaigns stuck at low tiers, the coarse value mix and modeled estimates from your own data are a better guide than the fine value average.
Crowd anonymity is also why SKAN data should sit next to device-level mobile attribution for users who granted ATT consent, rather than replace it. Each covers what the other cannot.