Glossary
Measurement, defined.
Short, plain definitions of the terms that decide whether your ad data can be trusted. Each one links to the full guide.
Tracking
- Consent Mode v2
- Google's framework for passing a visitor's consent choices (ad_storage, analytics_storage, ad_user_data, ad_personalization) to its tags. Required for EEA traffic in Google Ads; with it, Google models conversions from visitors who decline. Read the guide →
- Enhanced conversions
- Sending hashed first-party customer data, such as an email address, with a conversion so Google Ads can match it to a signed-in Google user. It recovers conversions that cookies alone miss. Read the guide →
- GCLID
- The GCLID (Google Click Identifier) is a unique ID that Google Ads adds to your landing page URL when someone clicks an ad with auto-tagging on. It lets Google tie a later conversion back to the exact click, campaign and keyword. Full definition →
- Google Tag Gateway
- Google Tag Gateway for advertisers is a free Google feature that loads your Google tag and sends its measurement requests through your own domain instead of Google's domains. Your CDN or load balancer forwards the traffic to Google, so fewer signals get lost to browser and ad blocker restrictions. Full definition →
- Server-side tagging (server-side GTM)
- Running your tags in a server container on your own subdomain instead of only in the browser. The server receives events and forwards them to GA4, Google Ads and Meta, which gets past ad blockers and extends cookie lifetime. Read the guide →
Attribution & incrementality
- Attribution
- Assigning credit for a conversion to the touchpoints before it, by a rule (last click) or a model (data-driven). Useful for daily optimization, but it also credits conversions that would have happened anyway. Read the guide →
- Geo lift test
- An incrementality test that changes ad spend in a set of regions and compares their sales with matched regions where nothing changed. It needs no user-level tracking. Read the guide →
- Google Meridian
- Google's open-source, Bayesian marketing mix model, generally available since January 2025. Free to use, but running it well takes years of varied weekly data and statistical skill.
- Incrementality
- The share of conversions that would not have happened without an ad. You measure it by comparing a group that saw the ads with a comparable holdout group that did not. Read the guide →
- Marketing mix modeling (MMM)
- A statistical model that estimates each channel's contribution to sales from aggregate weekly spend and outcomes over time, without user-level tracking. It needs two to three years of varied data to be reliable.
App measurement
- gBraid and wbraid
- Google's privacy-preserving click parameters for iOS traffic since App Tracking Transparency: wbraid for web-to-web journeys, gbraid for web-to-app. If an app never hands the opening URL to Firebase, gbraid is lost and iOS conversions go unattributed. Read the guide →
- MMP (mobile measurement partner)
- A third-party attribution platform such as Adjust, AppsFlyer, Branch or Airbridge. It attributes every app install and in-app event to its source, across paid, organic and owned channels, and reports conversions back to each ad network. Read the guide →
- SKAdNetwork and AdAttributionKit
- Apple's privacy-preserving attribution for app installs driven by ads: aggregated, delayed postbacks with a limited conversion value. AdAttributionKit is its successor. An MMP runs on top of it rather than replacing it.
Bidding & metrics
- AI Max for Search
- Google Ads' set of AI features for Search campaigns: search term matching beyond your keywords, text customization and final URL expansion. Dynamic Search Ads are being upgraded to it. Read the guide →
- Meta Andromeda
- The AI retrieval system Meta has used since late 2024 to decide which ads are eligible for each person. It reads your creative to match ads to people and learns from the conversion events you send back. Read the guide →
- Value-based bidding
- Bidding strategies such as Maximize conversion value and Target ROAS that optimize for the value of conversions, not their number. They only work as well as the values you send. Read the guide →

