Attribution Reporting API
Attribution Reporting API 是什么?
Attribution Reporting APIA Privacy Sandbox API in Chrome and Android that lets advertisers measure ad conversions across sites without cross-site identifiers, using browser-mediated, noise-injected event-level or aggregated reports.
The Attribution Reporting API (ARA) is the Privacy Sandbox mechanism for measuring ad-driven conversions without exposing cross-site identifiers. Sources (ad clicks or views) are registered on the publisher side; triggers (conversions) are registered on the advertiser side. The browser maintains the join between them locally and, at delayed and randomized intervals, sends one of two kinds of report. Event-level reports return a noisy, low-resolution mapping from a source event to a conversion event, suitable for click attribution. Aggregatable reports go through a separate Aggregation Service running in a trusted execution environment, which adds calibrated noise (differential privacy) before returning aggregate dashboards across many users. Both report types are designed to bound the cross-site information any caller can learn. ARA is one of the central pieces of the post-third-party-cookie measurement story and a focus of regulator attention (CMA, French CNIL, German DPAs). For ad-tech and measurement vendors, integrating ARA requires substantial backend changes for noise budgets, report aggregation, and trust-establishment with the Aggregation Service.
● 示例
- 01
An advertiser registers click-trigger and conversion-trigger callbacks; Chrome delivers event-level reports days later, with noise to limit user-level inference.
- 02
A measurement vendor stands up an Aggregation Service in a TEE so that ARA aggregatable reports can be processed into campaign-level dashboards.
● 常见问题
Attribution Reporting API 是什么?
A Privacy Sandbox API in Chrome and Android that lets advertisers measure ad conversions across sites without cross-site identifiers, using browser-mediated, noise-injected event-level or aggregated reports. 它属于网络安全的 隐私与数据保护 分类。
Attribution Reporting API 是什么意思?
A Privacy Sandbox API in Chrome and Android that lets advertisers measure ad conversions across sites without cross-site identifiers, using browser-mediated, noise-injected event-level or aggregated reports.
Attribution Reporting API 是如何工作的?
The Attribution Reporting API (ARA) is the Privacy Sandbox mechanism for measuring ad-driven conversions without exposing cross-site identifiers. Sources (ad clicks or views) are registered on the publisher side; triggers (conversions) are registered on the advertiser side. The browser maintains the join between them locally and, at delayed and randomized intervals, sends one of two kinds of report. Event-level reports return a noisy, low-resolution mapping from a source event to a conversion event, suitable for click attribution. Aggregatable reports go through a separate Aggregation Service running in a trusted execution environment, which adds calibrated noise (differential privacy) before returning aggregate dashboards across many users. Both report types are designed to bound the cross-site information any caller can learn. ARA is one of the central pieces of the post-third-party-cookie measurement story and a focus of regulator attention (CMA, French CNIL, German DPAs). For ad-tech and measurement vendors, integrating ARA requires substantial backend changes for noise budgets, report aggregation, and trust-establishment with the Aggregation Service.
如何防御 Attribution Reporting API?
针对 Attribution Reporting API 的防御通常结合技术控制与运营实践,详见上方完整定义。
Attribution Reporting API 还有哪些其他名称?
常见的别称包括: ARA, Attribution Reporting。
● 相关术语
- privacy№ 960
Privacy Sandbox
Google's umbrella initiative for replacing third-party cookies and cross-site identifiers with privacy-preserving alternatives — Topics, Protected Audience (FLEDGE), Attribution Reporting, and on-device APIs — under heavy regulatory and competitor scrutiny.
- privacy№ 1286
Topics API
A Privacy Sandbox API in Chrome and Android that derives a small set of high-level interest topics from the user's recent browsing locally on the device, exposing them to participating sites instead of cross-site tracking identifiers.
- privacy№ 351
差分隐私
一种数学框架,用于在发布统计或训练模型时量化隐私损失,通过加入经过校准的噪声使任何单个个体的影响在可证明的范围内。
- privacy№ 1263
第三方 Cookie
由浏览器地址栏域名之外的另一域名设置的 Cookie,历史上常用于跨站点跟踪用户。
- privacy№ 266
跨站点跟踪
将同一用户在多个不相关网站上的活动关联起来,构建长期行为画像的做法。
- cloud-security№ 1300
可信执行环境 (TEE)
处理器中一种安全、隔离的执行环境,代码和数据在其中获得机密性与完整性保护,主机操作系统和虚拟化层也无法访问。