Attribution & Measurement

Walled Garden

Also known asWalled GardensClosed Ecosystem

A closed advertising ecosystem (Meta, Google, TikTok, Amazon, Apple) that controls its own inventory, data, and measurement — and reports on its own performance rather than exposing raw data to advertisers.

Key takeaways

  1. 01A walled garden is a self-contained ad ecosystem where the platform owns the audience, the inventory, and the measurement — and grades its own homework.
  2. 02The major walled gardens — Meta, Google, TikTok, Amazon, Apple — are Self-Attributing Networks (SANs) that report installs/conversions to MMPs rather than sharing click-level data.
  3. 03Their advantage is rich first-party data and logged-in audiences; their cost to advertisers is measurement opacity and inability to fully verify or de-duplicate.
  4. 04Post-ATT, walled gardens grew stronger because their first-party logged-in data is resilient where open-web device-ID tracking broke.

A walled garden is a closed advertising ecosystem in which a single platform controls the inventory, the audience data, and the measurement — and reports performance back to advertisers rather than exposing the underlying raw data. Meta, Google, TikTok, Amazon, and Apple are the dominant mobile walled gardens.

Why walled gardens self-attribute

The big platforms operate as Self-Attributing Networks ([[self-attributing-network]]): instead of passing click-level data to your [[mmp]], they run attribution internally and send back a claimed-install postback. The MMP de-duplicates these claims across networks, but it can't see inside the garden. This is why two platforms can both "claim" the same install and why advertisers lean on [[incrementality]] testing to validate walled-garden numbers.

The trade-off for advertisers. The upside is enormous reach and rich first-party, logged-in audience data that powers targeting and optimization. The downside is opacity: limited verification, limited de-duplication across gardens, and a structural conflict of interest in self-reported results. The shift to privacy (ATT, [[skadnetwork]], Privacy Sandbox) strengthened walled gardens further, because their logged-in first-party data held up where open-web device-ID tracking collapsed.

Quick answers

What is a walled garden in advertising?

A walled garden is a closed ad ecosystem — Meta, Google, TikTok, Amazon, Apple — where the platform controls its inventory, audience data, and measurement, and reports performance to advertisers rather than sharing raw data. Advertisers get reach and rich targeting but limited independent verification.

Why are walled gardens a problem for attribution?

Because they self-attribute. The platform runs attribution internally and sends your MMP a claimed install instead of click-level data, so you can't fully verify or de-duplicate across networks — multiple gardens may claim the same conversion. Advertisers compensate with MMP de-duplication rules and incrementality testing to measure the platform's true causal contribution.

Did ATT make walled gardens stronger?

Yes. Apple's ATT and the broader privacy shift broke device-ID tracking on the open web and in smaller networks, but the walled gardens run on logged-in first-party data that doesn't depend on the IDFA. Their relative measurement and targeting advantage grew as open-ecosystem signals degraded.

Back to glossary