Wiki · Concept · Last reviewed June 25, 2026

Private Click Measurement

Private Click Measurement is Safari and WebKit's browser-mediated system for measuring ad click conversions with limited attribution data instead of ordinary cross-site tracking identifiers.

Category: Concept Published: June 25, 2026 Modified: June 25, 2026 Last reviewed: June 25, 2026 Tags: browser privacy, ad measurement, Safari, WebKit, surveillance

Definition

Private Click Measurement, often shortened to PCM, is a WebKit and Safari ad-attribution design for measuring whether a click on one site leads to a conversion on another site without using user or device identifying cookies. The Privacy Community Group draft defines it as a privacy-preserving way to measure clicks that result in cross-site navigations, such as purchases or sign-ups after an ad click.

WebKit introduced PCM publicly as a proposed web standard in 2021, saying it was enabled by default in iOS and iPadOS 14.5 betas. A 2022 WebKit update says PCM shipped in iOS and iPadOS 14.5 and Safari 14.1 on macOS. Apple's Safari privacy notice names the user-facing setting Privacy Preserving Ad Measurement and says it can be disabled in Safari's advanced settings.

PCM should be distinguished from Google's Attribution Reporting API. Google Privacy Sandbox documentation says Attribution Reporting supports view-through measurement, event-level reports, richer summary reports, and third-party report recipients. PCM is narrower and centered on click-through attribution.

Mechanism

The basic PCM story has four parties: user, browser, click source website, and attribution destination website. A user clicks or taps on a source site, lands on a destination site, and later performs a triggering action such as a purchase. The browser records enough state to produce a delayed attribution report, but the report is designed not to carry ordinary personal identifiers.

WebKit's 2021 description set tight data budgets: an 8-bit identifier on the click source side and a 4-bit identifier on the conversion side. The source can distinguish 256 campaign-like values and the destination can distinguish 16 conversion-like values. Apple's Safari privacy notice says reports use limited data, IP address protection, no data storage, no cookies, and random delay between 24 and 48 hours.

WebKit's 2022 update added conversion-fraud-prevention work and a replacement path for third-party tracking pixels. The same update says the combined attribution report is sent to both click source and click destination and contains no user or device identifiers. The browser becomes the measurement broker.

Agent Context

PCM matters for AI Browsers and Computer Use because agents can click, compare, purchase, test, or navigate through ad-funded pages. A report can arise from a delegated task even when the human did not inspect every ad, redirect, or conversion page. The attribution record is not consent, attention, persuasion, or purchase authority.

Agent runtimes should keep a separate record of automated navigation, delegated purchase scope, and ad-measurement exposure. If a shopping agent or browser test harness generates PCM-eligible clicks, the audit trail should mark them as machine-assisted.

Governance Use

A serious PCM review should record the click source, attribution destination, source identifier policy, trigger policy, reporting endpoints, user setting state, browser version, app-to-web path if any, report delay, anti-fraud token behavior, and whether the flow replaces a tracking pixel. It should also record what the report changes: optimization, billing, fraud review, payout, budget, or analytics.

PCM belongs in the same governance cluster as Attribution Reporting API, Private Aggregation API, Aggregation Service, Real-Time Bidding, and Data Minimization. These pages separate measurement, aggregation, auction markets, and collection limits.

Limits

PCM is not proof that ad measurement is fair, necessary, consented to, causal, or harmless. It narrows a reporting channel; it does not answer whether a user wanted advertising measurement or whether the resulting decisions are acceptable. The PrivacyCG draft says it is neither a WHATWG Living Standard nor on the W3C standards track, so standards status should not be overstated.

The social limit is simple: a system can reduce cross-site tracking while preserving the assumption that advertising systems deserve behavioral feedback. That can be better than identifier-based tracking and still be a governance problem when measurement is on by default, poorly explained, or used to optimize manipulative funnels.

Review Record

Source Discipline

Claims about PCM purpose, bit limits, delay, handling, and Safari/WebKit history should cite WebKit and Apple sources. Claims about specification status and terminology should cite the PrivacyCG draft and repository. Claims comparing PCM to Attribution Reporting should cite Google Privacy Sandbox's comparison. Claims about actual deployment need operator evidence.

Spiralist Reading

Spiralism reads Private Click Measurement as a governance hinge: the browser can refuse the old tracking bargain while still carrying a smaller measurement bargain. That is progress only if the bargain remains visible. The question is who gets to turn a person's click, or an agent's delegated click, into market feedback.

Sources

Return to Wiki