Last updated
Labs
Search by name or topic, or filter by category and level. Every lab opens free in this browser.
Why you'd want to
The labs exist so you can try a measurement or personalization choice in this browser, not only read about it. Each page names what it reads and what it sends.
How to use them
Each lab includes a code walkthrough, with Google and Adobe tabs.
Search titles, descriptions, and tags, then narrow by category and level. The full list is on this page.
Category
Level
113 of 113 labs
A short path before the rest of the labs.
How a visit becomes a hit, a path, or an event.
Consent, storage, and what a page can read.
Rules, tests, and the page a visitor actually sees.
A goal, a rule, and a simulated visitor. No live model.
How a client calls a read-only tool server.
Sixteen labs for tagging, consent, models, and proofs.
What the phone can measure, kept in this browser.
Campaign links, platforms, and which touch gets credit.
How search and assistants find and cite the site.
Funnels, cohorts, and charts for reading a count.
Speed, the edge, and how a request leaves this origin.
Three panels, three sources. The browser column is local. The network column is Resource Timing. The back-end column is the beacon this page posts to its own API, plus the fields the server adds. Personalization uses the same consent switch: when it is off, the site falls back to the default page and this page’s own beacon is suppressed. Switching that control from on to off stores one consent_change with consent denied. Local storage personalization lists the keys behind that memory. Web pathing shows the order of those actions. Location maps the city on the request and the hops that follow it. Segments turn those signals into on/off rules, and Recommendations use the pathing counts to suggest a next page. dataLayer is the same activity as GA4 events. Live is the third-party view of those visits in Google Analytics. API is a request and a response, using a public list of data centers.
Measurement consent
Off. Events are still built in this browser so you can see them, then dropped. A first Decline stores nothing. Switching from on to off stores one consent_change whose consent field is denied, and nothing after that. The Google tag is not loaded until you accept. After a withdrawal, analytics storage is set back to denied, and the rest of the site stays on the default experience.
Scroll this page to cross 25%, 50%, 75%, and 100%. Each threshold becomes its own event.
Source · this browser
Read from navigator, screen, document, and the Performance API. This whole panel stays on the device. A beacon carries the smaller profile, not this list.
Reading this browser…
Source · Resource Timing
Each row is a request this document made. The list is not uploaded. Posts to /api/collect show up here as first-party fetches after they finish. googletagmanager.com and google-analytics.com are labeled third-party Google Analytics.
Waiting for resource entries…
Source · POST /api/collect
The raw body leaves the browser. The route adds receipt time, parsed user agent, and edge headers, then stores the record in the database. The stream and a two-second poll both read new rows, so a stored event still shows up when the request lands on another serverless instance. This panel drops its own events while consent is off. Switching measurement from on to off still stores one consent_change with consent denied.
Stream: connecting · 0 stored
Waiting for a web vital from this page. If consent is off, the suppressed payload appears here and this panel does not post it. Page views, clicks, and scroll depth are sent by the site-wide tracker. They raise the stored count, and they are not this panel’s raw payload.
0 pushes this tab
Consent is off, so this tab does not push events to dataLayer. Open the dataLayer lab for the queue, the consent gate, and the collect request.
Reading path transitions…