Changelog

Report Review · August 4, 2026

A finished report can now ask one or two questions only the site’s owner or agency can answer — and record what they know beside the evidence.

• Anchored questions. Each question is tied to a specific finding. Answers take more than one selection where responsibility is genuinely shared, carry an optional note, and stay on the record with the report.
• Notes without an account. An anonymous reviewer can write notes, and they upload with the report if it is later saved.
• Saved reports. Save a report to an account to keep it past its normal expiry. The new Reports page lists everything by URL, with Proof Watch as a per-URL switch — saved reports and account settings each get their own page.
• Check this yourself. Every check that raises a finding ships steps you can run in your own browser.
• Try it on the homepage. The report excerpt includes a working example: the answer options respond, and “Add a note” shows what a note looks like. Nothing you click there is stored.

August 3, 2026

• Sharper pixel identification. Scan inventories name Meta, Microsoft UET, and Reddit pixel identifiers, and catch the first-party Google Ads conversion request that gtag routes through www.google.com.

Alpha 1 · July, 2026

MeasureProof opened for use on June 29 with the free scan, evidence-first reports, and the weekly Proof Watch. The releases that followed built out accuracy, coverage, and the public proof surface.

• Alpha opens. Free scan, evidence-first reports, and the weekly Proof Watch, live June 29. Blocked reports state only what they could actually verify.
• Accuracy on consent-gated sites. The scan behaves like a normal visitor’s browser so consent platforms serve it real tracking, waits for late, consent-gated beacons before concluding, and classifies first-party and server-side GA4 collection endpoints instead of misreading them.
• Outcome-first reports. Reports open with the outcome, a short synopsis of what was checked, and the next step under it.
• Tracker coverage. Adobe (Launch, Target, Audience Manager, Edge Network), PostHog, Google Ads, Floodlight, LinkedIn Insight, Reddit, Klaviyo, and fifteen further vendors added to scan inventories.
• Consent baseline. Reports gained a conservative consent section built from what the scan observed, and the method page became a per-check catalog.
• Proof studies. Published the first two — a duplicate GA4 pageview observed on a live store, and a setup verified clean — plus check explainers on duplicate pageviews and tags firing before consent.
• Blocked-scan recovery. You submit one URL and get one report; when a first attempt is blocked, the scan continues automatically where recovery is possible.
• Correct per-row counts. Each inventory row shows the count for its own tracker rather than the page total.
• Real evidence on the homepage. The homepage example became an excerpt of a real production scan, anonymized — with two full anonymized reports published behind it, and an FAQ.
• Contact form. On the website; the plain email address still works.