Google Analytics can tell you how people discover your website, what they do after arriving, and which actions contribute to revenue. The hard part is understanding the language used inside the platform.
A small wording difference can change the meaning of an entire report. “First user source” and “Session source” are both acquisition dimensions, but they answer different questions. Active users and total users are related, but they are not interchangeable. A key event can be important to your business without being used as an advertising conversion.
This Google Analytics glossary explains those distinctions in plain English. It covers the core data model, reporting fields, acquisition terminology, engagement metrics, privacy concepts, ecommerce reporting, attribution, data quality, integrations and the latest 2026 additions.
The article has been checked against Google Analytics terminology and product information available through July 11, 2026. The “2027 edition” label means the glossary is structured for continued updates. It does not claim knowledge of features that Google has not announced.
How to Use This Google Analytics Glossary
This guide works in two ways. You can read it from start to finish to understand how GA4 fits together, or jump directly to a definition when an unfamiliar term appears in a report.
Readers new to analytics should follow this order:
- Start with the GA4 data model, account structure, dimensions and metrics.
- Learn how users, sessions, events and parameters relate to one another.
- Move to acquisition, engagement, key events and attribution.
- Finish with privacy, data quality, integrations and advanced reporting.
Experienced users can use the A–Z section near the end as a quick reference.
Fast A–Z Lookup Versus the Guided GA4 Learning Path
A traditional glossary presents hundreds of isolated entries. That format works when you need a quick definition, but it does not explain the relationships between terms.
GA4 data follows a connected path. A person visits a website. Analytics identifies the browser or app instance. The visit becomes a session. Actions within that session generate events. Event parameters add detail. Dimensions describe the resulting data, while metrics quantify it. Reports arrange those dimensions and metrics so that analysts can answer business questions.
Understanding this chain prevents many reporting errors. It also makes the interface easier to navigate because you know whether you are looking for a user attribute, session source, event parameter or calculated metric.
How Every Definition Is Structured
The most useful Google Analytics definitions answer more than “What does this word mean?”
A complete definition should explain what the field measures, how it is populated, what scope it uses, where it appears and how it can be misread. A percentage should include its calculation. A traffic field should state whether it describes the user’s first visit, the current session or credit assigned to an event.
This guide also uses practical examples. A technical definition becomes much easier to understand when you can see how it affects a campaign, content report or ecommerce decision.
Current, Beta, Legacy and Deprecated Status Labels
Google Analytics changes regularly. A term can remain visible after its meaning changes. It can also appear in one part of the platform before reaching every eligible property.
“Current” means the term is part of the standard GA4 product or current documentation. “Beta” means Google is still developing or expanding the feature. “Limited availability” means some properties may not have access. “Legacy” refers to Universal Analytics terminology or an older GA4 label. “Deprecated” means the term or feature should no longer be used for new work.
This distinction matters when documenting a measurement plan. A business should not design a critical reporting workflow around a beta feature without a fallback.
What Is Google Analytics 4? Core GA4 Terms and Definitions
Google Analytics 4, commonly called GA4, is Google’s current analytics platform for websites and applications. It can collect web and app data within the same property and organize that activity around events.
Google designed GA4 as the successor to Universal Analytics. Standard Universal Analytics properties stopped processing new data on July 1, 2023. Universal Analytics 360 properties received an extension until July 1, 2024. New Google Analytics implementations should use GA4 terminology and collection methods.
The GA4 Event-Based Data Model
The GA4 event-based data model treats an interaction as an event. A page view is an event. A product view is an event. A completed purchase is an event. Opening an app, starting a session, scrolling, clicking an outbound link and submitting a form can all generate events.
This differs from Universal Analytics, which organized data into categories such as pageview hits, event hits, ecommerce hits and social interaction hits. GA4 uses one flexible event structure across websites and apps.
An event has a name and may include parameters. The name describes the action. Parameters provide context.
For example, a purchase event may include a transaction ID, currency, total value, shipping amount, tax and an array of purchased items. The event records that a purchase happened. Its parameters explain what was purchased and how much money was involved.
This event-first structure is the foundation for nearly all GA4 terms and definitions. Users generate sessions. Sessions contain events. Events carry parameters. Parameters populate dimensions and metrics. Reports organize the result.
Account, Property, Data Stream, Stream ID and Measurement ID
A Google Analytics account is the top administrative container. An organization may have one account or several accounts, depending on its structure, access requirements and reporting needs.
A property sits inside an account. The property is the main collection and reporting environment for a website, app or connected digital experience.
A data stream is a source of incoming data within a GA4 property. A property can contain web streams, Android app streams and iOS app streams. A company with a website and mobile application can therefore analyze both inside one property when that structure suits its reporting needs.
A Stream ID is the numerical identifier assigned to a data stream.
A Measurement ID identifies a web data stream and normally begins with G-. It tells the Google tag where website data should be sent. It is not the same as a property ID, Stream ID or Google Tag Manager container ID.
A clean account structure should match how the business needs to control access and analyze customer journeys. Creating separate properties for every domain without considering cross-domain behavior can make reporting harder.
Google Tag, gtag.js, Google Tag Manager and Tag ID
The Google tag is the code framework used to send data from a website to supported Google products. It can be installed directly on a website or deployed through Google Tag Manager.
gtag.js is the JavaScript library used by the Google tag. A direct installation places the code on the site and configures it with a destination such as a GA4 Measurement ID.
Google Tag Manager is a separate tag-management system. It lets teams deploy and manage tracking tags through a container rather than editing website code for every change.
A Tag ID identifies a destination or Google tag configuration. Depending on the product, it may begin with prefixes such as G-, GT-, AW- or DC-.
Google Tag Manager does not replace Google Analytics. It controls how tags are deployed. Google Analytics receives, processes and reports the data those tags collect.
GA4 Metrics and Dimensions Explained
Almost every GA4 report combines dimensions with metrics. Understanding the difference is one of the fastest ways to become comfortable with the interface.
A dimension describes something. A metric measures something.
Page title, country, device category and session source are dimensions. Users, sessions, event count and revenue are metrics.
Dimensions vs Metrics in Google Analytics
The difference between dimensions vs metrics in Google Analytics can be understood through a traffic report.
Suppose the report contains a row for Organic Search. “Session default channel group” is the dimension. It describes the type of traffic. “Sessions,” “engaged sessions” and “total revenue” are metrics. They quantify what happened for that channel.
Dimensions are commonly text values, although they can contain dates, numbers formatted as labels or other descriptive values. Metrics are numerical and support calculations such as sums, averages, percentages and ratios.
The same dimension can be paired with several metrics. Page title might be analyzed with views, active users, average engagement time and key events. Each combination answers a different question.
Not every dimension is compatible with every metric. A field can become unavailable in an exploration when its scope or processing method conflicts with the fields already selected. Many standard fields are populated from event parameters, while custom definitions are needed when the required information is not included in GA4’s standard reporting schema.
Standard, Primary, Secondary, Custom and Calculated Fields
Standard dimensions and metrics are built into Google Analytics. Examples include Event name, Country, Sessions, Active users and Total revenue.
A primary dimension is the main descriptive field used to organize a report table. A secondary dimension adds another layer of detail.
For example, Session default channel group may be the primary dimension in a traffic report. Device category could be added as a secondary dimension to show how each channel performs on desktop, mobile and tablet devices.
Custom dimensions and metrics GA4 configurations let a business report data that is collected but not available as a standard field.
A custom dimension is used for descriptive data. Examples include membership tier, article author, selected delivery method or logged-in status.
A custom metric is used for numeric data. Examples might include loyalty points earned, application score or the number of items included in a custom interaction.
A calculated metric combines existing metrics using a formula. For example, an analyst might create a margin-related metric from revenue and cost fields when the required inputs are available.
Custom definitions should be created carefully. A parameter with thousands of unique values can cause high-cardinality reporting problems. Unique user IDs, session IDs and exact timestamps should not be registered as ordinary custom dimensions.
User, Session, Event and Item Scope
Scope describes the level at which a dimension or metric applies.
| Scope | What it represents | Common examples | Main interpretation risk |
|---|---|---|---|
| User | Information associated with a person or identified user | First user source, user-scoped custom property | Treating first-acquisition data as the source of every later visit |
| Session | Information associated with one visit or activity period | Session source, sessions, engaged sessions | Assuming a session-level source identifies the user’s original acquisition |
| Event | Information attached to an individual action | Event name, event count, event-scoped campaign credit | Combining event-level attribution with session totals without checking compatibility |
| Item | Information attached to a product inside an ecommerce event | Item name, item category, item revenue | Confusing item totals with transaction-level totals |
A user can have several sessions. A session can contain many events. An ecommerce event can contain several items.
Scope explains why two fields that sound similar can produce different results. “First user source” is user-scoped. “Session source” is session-scoped. An event-scoped source can represent attribution credit for a key event.
When selecting a dimension, ask what object it describes. When selecting a metric, ask what is being counted. This habit prevents many incorrect comparisons.
GA4 Event Tracking and Event Parameters
GA4 event tracking is the process of measuring user actions as events. Some events are collected automatically. Others require configuration.
A useful event plan starts with business questions, not with every possible click. Tracking hundreds of interactions without a clear purpose produces a noisy implementation and makes reports harder to trust.
Automatically Collected, Enhanced Measurement, Recommended and Custom Events
Automatically collected events require no additional event-level setup after the appropriate tag or software development kit is installed. Examples include first_visit, session_start and user_engagement for web measurement.
Enhanced Measurement provides optional web events that can often be enabled from the web data stream settings. Depending on the site and configuration, it can measure page views, scrolls, outbound clicks, site search, video engagement, file downloads and form interactions.
Recommended events have predefined names and expected parameters for common business actions. Examples include login, sign_up, generate_lead, purchase, refund, add_to_cart and view_item.
Using recommended naming helps GA4 understand the purpose of an action and can unlock relevant dimensions, metrics and integrations.
A custom event is used when no automatically collected, enhanced or recommended event describes the interaction. Custom names should be clear, stable and documented.
A common mistake is creating custom versions of standard events. Sending product_bought instead of the recommended purchase event may require extra configuration and can prevent standard ecommerce reports from working correctly. Google’s event reference advises using recommended events when they match the behavior being measured.
Event Parameters, User Properties and Item Parameters
People searching for event parameters GA4 explanations often expect every parameter they send to appear immediately in standard reports. That is not how custom reporting works.
An event parameter describes an event. A select_content event might include a parameter that identifies the selected content type. A generate_lead event might include form type, service category or lead value.
A user property describes a user rather than one action. It may represent an account type, membership level or other long-lasting characteristic. User properties should not contain prohibited personally identifiable information.
An item parameter describes a product or service inside an ecommerce items array. Standard item parameters include item ID, item name, item brand, category, price, quantity and coupon.
Some standard parameters automatically populate existing dimensions and metrics. A custom event parameter often needs to be registered as a custom dimension or custom metric before it becomes available in normal reporting tools.
Custom Definitions and Parameter Registration
A custom definition connects collected custom data with GA4 reporting.
Suppose a website sends an event parameter named article_author with every article view. The parameter may be visible during debugging and available in raw exported event data. To use it as a normal dimension in reports or explorations, the property should create an event-scoped custom dimension tied to that parameter.
Registration is generally not retroactive. Data collected before the custom definition is created will not suddenly populate the new reportable field.
After a custom definition is created, reporting availability can take time. Google indicates that custom data may become available for reporting and advertising after roughly 24 to 48 hours.
Teams should keep a measurement dictionary containing the event name, parameter name, scope, expected values, trigger condition, owner and business purpose. This prevents two developers from sending different values under the same parameter.
GA4 Users: Active Users, Total Users, New Users and Returning Users
The word “user” sounds simple, but GA4 users are identified through available identifiers and the property’s reporting settings. Google Analytics does not automatically know the real-world identity of every visitor.
A person who visits on a laptop and later on a phone may appear as two users when GA4 cannot connect the devices. Clearing cookies or using a different browser can create another pseudonymous identifier.
Active Users vs Total Users
The distinction between active users vs total users is important because GA4 often foregrounds active users in its standard reporting interface.
Total users is the number of unique users who triggered any event during the selected date range.
Active users is the number of unique users who engaged with the website or app during the date range. GA4 also uses event criteria when identifying active users. A user can be considered active through an engaged session or the collection of specific events associated with a new or engaged user.
The metric called “Users” in many GA4 reports refers to active users, not total users.
For reach analysis, total users may be useful because it includes anyone who triggered an event. For engagement analysis, active users gives a more focused picture of people who met GA4’s activity criteria.
Google notes that new users can sometimes exceed active users because not every new user necessarily meets the active-user definition. Thresholding can also cause apparent discrepancies between user metrics.
New vs Returning Users
The new vs returning users comparison helps distinguish acquisition from repeat usage.
New users is the number of unique users who triggered first_visit on a website or first_open in an app during the selected period.
Returning users are users who initiated at least one previous session, whether or not that earlier session falls inside the selected reporting period.
A person can contribute to both categories during a long date range. Someone may first visit on Monday and return on Friday. The person generated new-user activity and later returning-user activity.
Cookie deletion, browser changes, consent choices and device switching can affect this classification. A known customer may appear new to GA4 when the platform receives a fresh device identifier and no User-ID links the activity.
Client ID, Device ID, User-ID and Reporting Identity
A Client ID is a pseudonymous identifier associated with a web browser. On websites, the device identity used by GA4 is derived from the Client ID stored through the _ga cookie when storage is available.
For apps, the device identifier is based on the app-instance ID.
User-ID is an identifier assigned by the business to a signed-in user and sent to GA4. It can connect activity across devices when the same person signs in and the implementation applies the identifier consistently.
Reporting identity GA4 settings determine which identity spaces are used to represent users in reports.
Blended reporting identity evaluates User-ID first, then device ID, then modeled data when eligible. Observed reporting identity uses User-ID and then device ID. Device-based reporting uses only the device identifier.
Changing the reporting identity does not rewrite the underlying collected data. It changes how GA4 represents users in affected reporting surfaces. Google states that the selection can be changed without permanent impact on data collection or processing.
GA4 Sessions, Engaged Sessions, Engagement Rate and Bounce Rate
A session groups interactions that occur during a period of activity. Sessions remain useful in GA4, even though the platform’s collection model is event-based.
GA4 Sessions, Session Start and Session Timeout
GA4 sessions usually begin when Analytics collects a session_start event.
GA4 assigns a session ID that can be associated with events in that activity period. A user can generate multiple sessions over time.
The default session timeout is based on inactivity. A property can adjust the timeout within the available settings. When the user remains inactive beyond the configured period and later returns, a new session can begin.
Unlike Universal Analytics, GA4 does not automatically start a new session merely because midnight passes or a new campaign source is detected during an existing session. This difference is one reason GA4 session totals should not be expected to match historical Universal Analytics totals.
Session counts in GA4 reports and BigQuery can also differ because they are calculated through different reporting and query processes. Directional trends may remain consistent even when the exact totals are not identical.
Engaged Sessions and Engagement Rate GA4
Engaged sessions meet at least one of three conditions. The session lasts longer than ten seconds, contains a key event, or includes at least two page views or screen views.
The default ten-second threshold can be adjusted in session settings for a web stream.
The calculation for engagement rate GA4 is:
Engagement rate = Engaged sessions ÷ Total sessions × 100
If a website records 800 sessions and 520 are engaged, its engagement rate is 65%.
This metric does not prove that users were satisfied. A visitor may remain on a confusing page for more than ten seconds and still qualify. Engagement rate is best used as a diagnostic signal alongside content context, event behavior and business outcomes.
Google excludes events such as first_visit, first_open and session_start from creating an engaged session simply because they were marked as key events.
Bounce Rate GA4 and Average Engagement Time
Bounce rate GA4 is the percentage of sessions that were not engaged.
Bounce rate = Non-engaged sessions ÷ Total sessions × 100
It is the inverse of engagement rate. An engagement rate of 65% corresponds to a bounce rate of 35%.
This differs from the traditional Universal Analytics bounce-rate concept, which focused on single-interaction sessions. A one-page GA4 session can be engaged if it lasts long enough or includes a key event.
Average engagement time measures the average time the website was in focus or the app was in the foreground. It is more restrictive than elapsed session time. If someone leaves a tab open while working in another application, GA4 does not treat the entire period as active engagement.
Engagement time is sent through the engagement_time_msec parameter with subsequent events when the page loses focus, the user navigates away, the app moves to the background or another qualifying event occurs.
Traffic Acquisition Definitions: Source, Medium, Campaign and Channel
Acquisition reporting explains how people discovered or returned to a website or app. The challenge is choosing the field that matches the question.
First User, Session and Event-Scoped Traffic Sources
First user traffic-source dimensions describe how the user was first acquired.
If a person first visits through Organic Search and returns through an email campaign one week later, First user source may remain associated with the original search source.
Session traffic-source dimensions describe how a session began. In the same example, the later visit could be attributed to the email source and medium at session scope.
Event-scoped acquisition dimensions are used when GA4 assigns attribution credit to key events or advertising conversions. They can use the property’s selected attribution model.
Use first-user fields when asking, “Which sources originally acquired our users?”
Use session fields when asking, “Which sources generated visits during this period?”
Use attribution-oriented event fields when asking, “Which marketing touchpoints received credit for important outcomes?”
Source, Medium, Campaign and UTM Parameters
The phrase source medium GA4 refers to two related dimensions.
Source identifies where traffic came from. Examples include google, bing, a newsletter name, a referring domain or an advertising platform.
Medium describes the general method of arrival. Examples include organic, cpc, email, referral and ai-assistant.
Campaign identifies a marketing initiative. A campaign could represent a seasonal promotion, product launch, paid media initiative or newsletter series.
UTM parameters add campaign information to destination URLs. Common parameters include utm_source, utm_medium, utm_campaign, utm_id, utm_term and utm_content. Google also supports utm_source_platform, while some other documented parameters are not currently reported in GA4 properties.
Naming conventions should be documented before campaigns launch. facebook, Facebook, fb and meta-facebook may represent the same platform to a human analyst but create fragmented values in source reports.
Default Channel Group, Custom Channel Group and Primary Channel Group
A default channel group places traffic into rule-based categories such as Organic Search, Paid Search, Referral, Email, Organic Social, Paid Social and Direct.
Google controls the Default Channel Group rules. They cannot be edited.
Custom channel groups let a property create its own channel definitions. A business could separate influencer traffic from general referrals or create a channel for strategic partners, provided the available dimensions and rules support the classification.
A primary channel group is the property’s main editable channel-grouping dimension when that functionality is available. It can provide a business-specific view while leaving Google’s Default Channel Group unchanged.
Channel definitions should be reviewed when campaign tagging changes. An email campaign with an inconsistent medium may fall into Unassigned instead of Email.
Google’s channel classifications can evolve as the market changes, so a channel should not be treated as a permanent raw-data field. It is the result of classification rules.
Source Group and AI Assistant Traffic in GA4
Two of the most important 2026 changes affect source normalization and traffic from generative AI services.
Source Group GA4 Versus Source and Source Platform
Source Group GA4 is a dimension concept introduced in June 2026 to consolidate variations of common platform sources.
A business may receive values such as facebook, fb, Facebook and Meta-facebook. Source Group can normalize related values into a cleaner reporting category.
Source remains the collected or assigned source value. Source Group provides a standardized grouping. Source Platform identifies the platform responsible for directing traffic, such as a buying platform or traffic-management platform.
This distinction improves cross-channel analysis because analysts no longer have to clean every spelling variation manually before comparing platform performance.
Source Group was announced with retroactive population, allowing eligible reporting to use the dimension for historical analysis rather than only future visits. It also includes grouping support for emerging sources such as ChatGPT and Perplexity.
Source Group does not remove the need for disciplined UTM tagging. Clean campaign parameters still improve granularity, governance and troubleshooting.
AI Assistant Traffic GA4, ai-assistant Medium and Channel
AI Assistant traffic GA4 identifies visits referred by recognized AI assistants.
Google announced a dedicated AI Assistant classification in May 2026. Matching traffic can receive the ai-assistant medium, an AI Assistant channel classification and the (ai-assistant) campaign value.
Current channel documentation uses the label “AI Assistants” and lists sources such as ChatGPT, Gemini, DeepSeek, Copilot and Grok. Google’s AI Overviews and AI Mode are excluded from this channel and remain part of Organic Search classification.
This creates a cleaner way to measure discovery through conversational systems. Analysts can compare AI-assistant visitors with organic search, referral and social users.
Important metrics may include sessions, engaged sessions, engagement rate, key-event rate, revenue and landing-page performance. Raw traffic volume alone says little about value.
Measuring ChatGPT, Gemini, Claude and Perplexity Referrals
GA4 relies on available referral and classification information. It cannot identify every interaction that takes place inside an AI platform.
A click may appear as Direct when the referrer is removed, hidden or unavailable. Some apps open links through browsers or redirect systems that affect source information. AI-generated brand awareness can also lead to a later direct visit rather than an identifiable referral.
For that reason, AI traffic reports measure recognized click-through visits. They do not measure every answer in which an AI system mentioned the brand.
Businesses that created custom AI referral groups before the native classification should compare the old and new rules. A custom group may still include niche services that are not yet covered, but it may also double-count or classify traffic differently from the official channel.
Key Events, Conversions and Attribution Models in GA4
GA4 uses events to measure actions. Important events can be marked as key events. Some events or key events can then support advertising conversions.
Key Events GA4 and Important Business Actions
Key events GA4 are events that measure actions considered particularly important to the business.
For an ecommerce company, purchase is an obvious key event. A lead-generation business might mark generate_lead, book_appointment or a qualified form completion. A subscription product might use sign_up, start_trial or subscribe.
Not every interaction should be a key event. Marking scrolls, minor clicks and every page view as key events makes performance reports harder to interpret.
A key event should represent progress toward a meaningful outcome. The event must also be implemented reliably. A form event that fires when someone opens the form is not a valid lead-completion measure.
Any collected event can be marked as a key event. When triggered, the action is recorded and surfaced through key-event metrics and reports.
Key Events vs Conversions in GA4
The relationship between key events vs conversions changed when Google aligned conversion terminology across Analytics and Google Ads.
The current flow is:
Event → Key event → Conversion
An event records an action. Marking it as a key event tells Analytics that the action matters to the business. A conversion can then be created for cross-channel advertising measurement and Google Ads use.
This distinction means conversions in GA4 should not always be used as a replacement term for all key events.
Key events appear in standard behavioral reporting. Google Ads conversions shared with Analytics appear in the Advertising area and do not appear as conversions in standard GA4 reports.
A linked Google Ads account is required when creating a new advertising conversion from an Analytics event or key event through the integrated workflow. Existing historical Analytics conversions were retained as key-event data when the terminology changed.
Attribution Model GA4, Data-Driven Attribution and Lookback Windows
An attribution model GA4 setting determines how credit for key events is distributed across eligible marketing touchpoints in affected reports.
Data-driven attribution uses available path data and modeling to distribute credit based on the estimated contribution of each interaction. It can assign fractional credit to several touchpoints.
Paid and organic last click gives credit to the last eligible non-direct touchpoint from paid or organic channels. Direct traffic generally does not override a known eligible source when another qualifying interaction exists within the lookback period.
A lookback window defines how far back an interaction can remain eligible for attribution credit. A shorter window may suit a quick purchase cycle. A longer window may suit expensive products or services with extended consideration.
Attribution does not prove that a channel caused a sale in the experimental sense. It applies a defined credit-assignment method to observed and modeled journeys. Incrementality testing is still needed when the business wants stronger evidence of causal impact.
Content, Page and Navigation Definitions
Content reports explain which pages or screens people viewed and how they moved through the digital experience.
Page View, Views, Pages and Screens and Views per User
A web page_view event records a page view. It is commonly sent when a page loads or when browser history changes in a properly configured single-page application.
The Views metric combines web page views and app screen views in relevant cross-platform reports. Repeated views are counted. One user can therefore generate many views of the same page.
Views per active user divides the number of views by active users. It can help compare consumption depth, but it should be interpreted with the type of content in mind. A support article that solves a problem in one view may be more useful than a slideshow that forces several page loads.
Pages and screens is a standard reporting area for content performance. It commonly uses page title and screen class or related page and screen dimensions.
Landing Page, Entrance, Exit and Exit Rate
A landing page is the first page viewed during a session. Landing-page reporting helps analysts understand where visits begin and how those entry points perform.
An entrance is the first page or screen view associated with a session.
An exit is the final page or screen associated with a session before activity ends.
Exit rate measures exits relative to views for a page or screen under the applicable calculation. A high exit rate is not automatically a problem. A confirmation page, contact page or article that answers a narrow question may naturally be the final page in a visit.
The business purpose of the page matters. A high exit rate on the first step of checkout may be concerning. The same rate on an order-confirmation page may be expected.
Page Path, Page Location, Page Title, Content Group and Site Search
Page location normally contains the full URL collected for a page. Page path focuses on the path portion and may include query-string variations depending on the selected dimension.
Page title is the title collected from the page. It is easier to read than a URL, but duplicate or dynamically changing titles can reduce reporting quality.
A content group is a business-defined category used to group related pages or screens. A publisher might group content by topic. A software company might group documentation, product, pricing and support pages.
Site-search reporting commonly uses the view_search_results event. The search term may populate from recognized query parameters or custom configuration.
Internal-search data is valuable because it shows what visitors expected to find. Search terms with poor engagement or no results often reveal navigation, content or product gaps.
Ecommerce, Revenue and Customer-Value Definitions
GA4 ecommerce measurement uses recommended events and an items array. A correct implementation records both transaction-level and product-level information.
Ecommerce Events From view_item to purchase
Common ecommerce events include view_item_list, select_item, view_item, add_to_cart, remove_from_cart, view_cart, begin_checkout, add_shipping_info, add_payment_info, purchase and refund.
These events represent stages in the shopping journey. The exact funnel differs by business, but event naming should follow the recommended schema when the action matches.
The purchase event should include a unique transaction ID. This helps identify transactions and supports duplicate handling.
Purchase value should represent the intended monetary amount consistently. Currency should be sent using a recognized currency code. Item information should be included in the items array.
A confirmation page that reloads and resends the purchase event without protection can inflate revenue. Testing should include reloads, payment redirects, browser back actions and failed payments.
Item Scope, Item Revenue, Quantity and Item Parameters
Item scope applies to individual products or services inside an ecommerce event.
A single purchase can contain three different products. Transaction value belongs to the event or transaction. Item name, item category, item brand, price and quantity describe each item.
Item revenue measures revenue connected with individual ecommerce items. It should not be confused with total revenue or purchase revenue.
The difference becomes important when orders include shipping, tax, discounts, refunds or multiple quantities. Adding item revenue may not reproduce every transaction-level amount unless the implementation and selected metrics account for each component in the same way.
Custom item-scoped dimensions can report product attributes not covered by standard fields, such as size, colour family, subscription interval or inventory classification.
Total Revenue, Purchase Revenue, ARPU, LTV, ROAS and ROI
Purchase revenue represents revenue from purchases, adjusted according to the relevant reporting definitions and refund data.
Total revenue can include more than ecommerce purchase revenue. Depending on the property and collected data, it may include purchase, subscription, advertising and other supported revenue categories, less refunds where applicable.
Average revenue per user, often called ARPU, divides revenue by the relevant user metric. Always confirm whether the selected GA4 metric uses active users, total users or another user definition.
Lifetime value, or LTV, estimates or reports the value generated by users over a broader relationship period. GA4’s user-lifetime analysis can combine acquisition and lifetime behavior, subject to identity, retention and reporting limitations.
ROAS means return on advertising spend:
ROAS = Revenue attributed to advertising ÷ Advertising cost
ROI means return on investment:
ROI = Net return ÷ Total investment
ROAS focuses on revenue relative to media cost. ROI considers profit or net return and may include wider costs. Reporting a strong ROAS does not guarantee that a campaign is profitable.
Reports, Explorations, Segments and Audiences
GA4 provides several reporting surfaces. They do not always use identical processing, field availability or data-retention rules.
Overview Reports, Detail Reports, Reports Snapshot and Collections
An overview report summarizes a topic through cards. It provides a high-level view and links to relevant detail reports.
A detail report contains charts and a data table organized by a primary dimension. Editors and administrators can customize available dimensions, metrics, filters and charts.
Reports Snapshot is a broad summary page designed to surface selected information from across reporting.
A collection is a group of related reports published in the Reports navigation. Collections can reflect business objectives or reporting themes.
The Library is the area where eligible users manage collections and reports. Publishing a report makes it available through the normal navigation for the intended users.
An overview report summarizes detail reports within a topic and uses summary cards rather than functioning like a full detail table.
Free-Form, Funnel, Path, Cohort and User Explorations
Explorations provide flexible analysis beyond standard reports.
Free-form exploration works like a configurable pivot table. It supports selected dimensions, metrics, segments, filters and visualizations.
Funnel exploration measures movement through defined steps. It can reveal where users abandon a process such as registration, checkout or onboarding.
Path exploration starts from or ends at an event or page and shows common paths through the experience.
Cohort exploration groups users based on a shared acquisition or behavior condition and tracks their activity over time.
User exploration displays activity associated with individual pseudonymous or User-ID-linked identities when the required data and permissions are available.
Explorations may be affected by sampling when queries exceed applicable limits. They are also subject to user-level data retention settings in ways that standard aggregate reports are not.
Comparison, Segment and Audience
A comparison applies a temporary subset to a standard report. For example, a report could compare mobile users with desktop users.
A segment is used in explorations to isolate users, sessions or events that meet defined conditions. Segments are analytical tools and can be applied to historical data within the exploration’s available period.
An audience is a persistent group of users who meet specified conditions. Audiences can be used for reporting and, when integrations and consent requirements are satisfied, advertising activation.
Audience membership begins as users meet the conditions. Audiences are not fully retroactive populations of every historical user who would have qualified before creation.
The practical difference is simple. Use a comparison for quick report filtering, a segment for deeper exploration analysis, and an audience when the group needs ongoing membership or activation.
Privacy, Consent Mode and Modeled Data Definitions
Modern analytics reporting is shaped by consent choices, browser restrictions, platform policies and regional requirements. Data gaps should be expected and understood rather than hidden.
Consent Mode GA4, Consent Signals and Consent Types
Consent Mode GA4 lets a website or app communicate a user’s consent choices to Google tags and software development kits.
Consent Mode does not collect consent from the user. A consent banner or consent-management platform handles that process. Consent Mode passes the resulting status to Google’s measurement systems.
Common signals include analytics_storage, which controls analytics-related storage, and ad_storage, which controls advertising-related storage. Additional signals such as ad_user_data and ad_personalization affect advertising measurement and personalization use cases.
Basic Consent Mode normally blocks relevant tags until consent is granted. Advanced Consent Mode allows tags to load with consent defaults and send limited cookieless signals when storage is denied, subject to the implementation and policy requirements.
A business must design consent behavior around the laws and rules that apply to it. GA4 configuration alone does not create legal compliance.
Google provides consent-settings checks inside Analytics and notes that updates to detected signal status may take 48 to 72 hours.
Behavioral Modeling, Conversion Modeling and Cookieless Pings
Behavioral modeling estimates the behavior of users who decline analytics identifiers by using patterns from similar consenting users in the same property.
It is used only when the property meets eligibility and data-quality requirements. Modeled data is not invented at random. It is an estimate based on observable relationships, but it still carries uncertainty.
Conversion modeling estimates conversions that cannot be directly observed because of consent choices, device restrictions or gaps in identifiers.
Cookieless pings are limited signals sent when storage consent is denied in an eligible advanced Consent Mode implementation. They can communicate information such as consent state and basic event context without using normal analytics cookies.
Modeled data should be identified clearly in decision-making. It is useful for reducing blind spots, but it should not be described as a direct record of every individual action.
Google Signals, Data Retention and Personally Identifiable Information
Google Signals can provide additional cross-device and demographic capabilities for eligible users who are signed in to Google and have the relevant advertising-personalization settings enabled.
Its use can affect reporting identity, advertising features and data thresholds. Teams should understand the privacy and reporting implications before enabling it.
Data retention controls how long certain user-level and event-level data remains available for explorations and some advanced reporting features. It does not mean every standard aggregated report disappears after the chosen period.
Personally identifiable information, often abbreviated as PII, includes data that directly identifies a person, such as a full email address or certain personal identifiers. Standard GA4 implementations must not send prohibited PII in URLs, event parameters, custom dimensions or other fields.
A form page can leak an email address through the URL if it appends submitted values as query parameters. Measurement quality reviews should therefore include privacy checks, not just event validation.
GA4 Data Thresholding, Sampling and Data Quality
A report can be technically valid and still be unsuitable for a decision. Analysts must check the report’s processing, privacy controls, scope and completeness.
GA4 Data Thresholding and Sampling
GA4 data thresholding and sampling are different concepts.
Thresholding removes or limits data in a report to reduce the risk of identifying individual users. It can occur when reports include Google Signals data, detailed demographics, search-query information or other sensitive combinations.
Sampling analyzes a subset of available data when an exploration query exceeds applicable processing limits. The result is an estimate based on the sampled records.
Standard aggregate reports are generally designed to remain unsampled, but they can still be affected by thresholding, cardinality, modeled data and other processing differences.
A data-quality indicator in the interface can show when a result is sampled, thresholded or based on complete available data.
When accuracy matters, use the least detailed fields needed to answer the question. Removing an unnecessary demographic dimension may reduce thresholding. Shortening a date range or simplifying an exploration may reduce sampling.
Cardinality, High-Cardinality Dimensions and the “Other” Row
Cardinality is the number of unique values in a dimension.
Device category has low cardinality because it normally contains a small set of values. Page location can have high cardinality when thousands of URLs contain unique query parameters.
When a report contains more unique dimension values than the reporting system can display or process individually, some values can be grouped into an “other” row.
Common causes include timestamps, unique IDs, long search terms, transaction-like custom values and URLs with uncontrolled parameters.
High-cardinality custom dimensions should be avoided. A unique customer identifier belongs in the User-ID feature, not in an ordinary reportable custom dimension.
Cardinality problems are often governance problems. Clean URLs, stable taxonomies and controlled parameter values improve both analysis and platform performance.
Not Set, Unassigned, Missing Data and Data Freshness
(not set) means GA4 did not receive a value for the selected dimension. The cause depends on the field. It may result from missing parameters, configuration issues or a dimension that does not apply to the event.
Unassigned is a channel value. It means available traffic information did not match the rules for a channel group.
(data not available) is different. It can appear when GA4 has received advertising or campaign information but has not finished processing the related attribution data. Some event-level values can update later.
Data freshness describes how recently collected data has been processed into a reporting surface. Realtime data appears quickly but supports a limited set of fields. Intraday data updates during the day. Daily processing is more complete.
Standard-property intraday processing commonly takes several hours, while full processing can take 24 to 48 hours. Reports may change during that period.
When a number looks wrong, check these causes before assuming the platform lost data:
- Confirm the date range, time zone and selected property.
- Check whether the field is user, session, event or item-scoped.
- Review thresholding, sampling, cardinality and modeled-data notices.
- Compare intraday data with the completed daily report.
- Test tags, consent behavior, campaign parameters and event values.
Implementation, Testing and Debugging Definitions
Reliable reporting begins with reliable collection. Reports cannot repair an event that fires at the wrong time or sends the wrong value.
DebugView, Realtime, Tag Assistant and Preview Mode
DebugView displays events from devices or browsers operating in debug mode. It shows event order and related parameters, making it useful during implementation.
Realtime reports show recent activity across users. They help confirm that data is reaching the property, but they are less detailed than DebugView for inspecting one test device.
Tag Assistant helps verify Google tags and diagnose configuration. Google Tag Manager Preview mode connects a browser session with the Tag Assistant interface so that implementers can inspect which tags fired, which triggers matched and what data was available.
Seeing an event in DebugView does not guarantee that every report will populate immediately. The event may require processing, a custom definition or additional ecommerce parameters. A debug test also does not prove that the event works across all browsers, consent states, templates and devices.
Testing should include successful actions, failures, duplicate attempts, page reloads and navigation edge cases.
Measurement Protocol and Data Manager API
Measurement Protocol sends events to GA4 through server-to-server HTTP requests. It is useful when an action occurs outside the browser or app, such as an offline transaction, backend status update or connected-device event.
Measurement Protocol should complement normal client-side collection rather than silently replace required session and user context. Poorly constructed requests can create disconnected events, missing acquisition data or inaccurate engagement reporting.
The Data Manager API is another route for sending recommended and custom server-to-server events to web and app streams. Google added support for this workflow in May 2026, describing it as an alternative to Measurement Protocol for eligible event collection.
Server events require strong deduplication and identity governance. A purchase should not be sent once from the browser and again from the backend unless the implementation has a reliable method for preventing double counting.
Internal Traffic, Developer Traffic, Hostname Filters and Unwanted Referrals
Internal-traffic rules identify activity from employees, agencies, testing locations or other known networks. A data filter can then exclude or test that traffic.
Developer traffic can separate events generated on debug-enabled devices.
Hostname filters were added in June 2026. They can exclude events when the hostname does not match approved domains. This helps protect a property from data sent through copied tags, staging sites or unauthorized domains.
Unwanted referrals are domains that should not start a new referral attribution path. Payment processors are a common example. Without correct referral handling, a customer may appear to return from the payment provider before completing a purchase.
Excluding a referral does not remove the visit. It changes how the session source is treated. Referral configuration should therefore be tested alongside cross-domain measurement and payment flows.
Integrations, Data Import and External Reporting Terms
GA4 becomes more useful when it is connected with advertising, search, reporting and data-warehouse systems. Each integration has its own scope and processing rules.
Google Ads, Search Console and Google Business Profile Integrations
Linking Google Ads with GA4 supports audience sharing, advertising reporting and conversion workflows when account permissions and consent conditions are met.
Search Console linking adds organic search information such as queries and landing-page performance to dedicated reports. Search Console and GA4 should not be expected to show identical click, session or user totals because they measure different stages and use different processing methods.
Google Business Profile reporting was expanded in June 2026. The integration can include interactions such as calls, bookings, direction requests, website clicks, messages and menu-related activity, depending on the profile and available data.
Google Business Profile metrics in Analytics use a rolling six-month availability window. They can help connect Search and Maps visibility with local intent, but they should not be treated as a complete record of physical visits.
BigQuery Export, Data API and Looker Studio
BigQuery export provides event-level GA4 data in Google Cloud. It gives analysts more control over SQL queries, joins, transformations and data retention.
BigQuery data does not reproduce every processed GA4 interface metric automatically. The interface may use reporting identity, modeling, attribution logic and calculated reporting methods that are not represented in the same way in raw event exports.
The Google Analytics Data API retrieves report data programmatically. It works with GA4’s reporting schema rather than providing the same raw event-level structure as BigQuery.
Looker Studio is a dashboard and visualization platform. It can connect with GA4 and other data sources. A dashboard does not improve the quality of the underlying implementation. It only presents the data it receives.
Use standard reports for regular monitoring, explorations for flexible analysis, the Data API for automated report retrieval and BigQuery when event-level control or advanced modeling is required.
Data Import, Cost Data, User Data, Item Data and Offline Events
Data Import adds external information to Analytics using defined join keys.
Cost-data import can add spend, clicks and impressions from non-Google advertising platforms. This supports cross-channel efficiency analysis when campaign identifiers and naming match.
Item-data import can enrich product information. User-data import can add permitted non-identifying attributes under supported use cases and policies.
Offline events can be sent through supported server-side collection methods. Examples include completed phone sales, approved applications or in-store actions.
Imported data must match the intended identifiers. A campaign name alone may not be stable enough for reliable joins. Campaign IDs and documented keys usually provide better control.
AI, Predictive Analytics and New 2026 GA4 Terms
GA4 now includes several machine-learning and assistant-driven features. Their availability can vary by property, language and rollout stage.
Generated Insights, Analytics Intelligence and Anomaly Detection
Analytics Intelligence refers to machine-learning features that help users understand changes and patterns in data.
Anomaly detection identifies values that differ from an expected range. It can surface unusual increases or decreases that deserve investigation.
Trend-change detection identifies longer-term changes in the direction or rate of a metric rather than only a one-day spike.
Generated insights were added to the GA4 Home page in February 2026. They summarize notable changes since the user’s previous visit, including configuration changes, anomalies and seasonal patterns.
A generated explanation is a starting point. Analysts should still verify implementation changes, campaign launches, consent updates, outages and business events before making decisions.
Predictive Metrics, Predictive Audiences and Eligibility
Predictive metrics use machine learning to estimate future user behavior.
Purchase probability estimates the likelihood that a user who was active recently will complete a specified purchase-related action within a future period.
Churn probability estimates the likelihood that a recently active user will not remain active in a future period.
Predicted revenue estimates expected purchase revenue from eligible users over a future window.
Predictive audiences use these metrics to group users, such as likely purchasers or users at risk of churn.
Not every property qualifies. GA4 needs sufficient event volume, eligible users and consistent recommended-event data to train useful models. A predictive metric may disappear when the property no longer meets the requirements.
Predictive outputs are probabilities, not promises. They should be tested against business outcomes.
Task Assistant, Ask Advisor, Cross-Channel Budgeting and Conversion Attribution Analysis
Task Assistant provides configuration recommendations organized around actions such as connecting accounts, improving reports and fixing data-quality issues. It was announced in April 2026.
Ask Advisor is a beta conversational feature powered by Gemini models. For eligible English-language properties, it can answer questions about property performance, generate visualizations, help navigate features and read configuration information. It works with information available within the selected property.
Cross-channel budgeting is a beta capability for tracking paid-channel investment and exploring budget plans. Projection plans estimate future performance against indicators such as spend, conversions and revenue. Scenario plans compare potential returns at different budget levels.
The Conversion attribution analysis report is another beta feature. It includes views for assisted conversions and data-driven funnel stages, helping advertisers examine early, middle and late journey contributions.
These features may not appear in every property. They should be labelled as beta or eligibility-dependent in internal training materials rather than presented as universal GA4 functionality.
Universal Analytics to GA4 Terminology Translation
Many people still search using Universal Analytics language. Translating those terms is useful, but the closest GA4 concept is not always an exact replacement.
| Universal Analytics term | Current GA4 term or closest concept | Important difference |
|---|---|---|
| Goal | Key event | GA4 marks collected events as important rather than configuring destination, duration or pages-per-session goals in the old model |
| Conversion | Key event in standard reports; conversion in current advertising workflows | Conversion now has a more specific cross-channel and Google Ads context |
| Pageview | page_view event and Views metric | GA4 records page views as events and can combine page and screen reporting |
| Unique pageviews | No exact direct replacement | Analysis may use users, sessions, views and page-related event logic depending on the question |
| Bounce rate | GA4 bounce rate | GA4 defines it as the percentage of sessions that were not engaged |
| Average session duration | Average engagement time or session-duration analysis | Engagement time focuses on active foreground or focused-page time |
| View | Property reporting structure, filters, subproperties or separate properties | GA4 does not reproduce the old account-property-view hierarchy |
| Event category, action and label | Event name and parameters | GA4 uses a flexible event-and-parameter schema |
| User | Often Active users in standard GA4 reporting | UA and GA4 user identification and default reporting metrics differ |
| Session | GA4 session | Midnight and campaign-change behavior differ from Universal Analytics |
Goals, Conversions and Key Events
A Universal Analytics goal represented a desired outcome. Goal types included destinations, duration, pages or screens per session and configured events.
GA4 does not use the old goal structure. A meaningful action is collected as an event and marked as a key event.
This approach is more flexible, but it places greater responsibility on event design. A destination-based thank-you-page goal may become a generate_lead event that fires only after a verified submission.
The word “conversion” still appears in current GA4, but its role has narrowed. Standard reports use key events for important business actions. Advertising conversions support cross-channel measurement and Google Ads workflows.
Pageviews, Unique Pageviews, Views and Users
Universal Analytics pageviews and GA4 views are related, but their collection context differs.
GA4 records page views as events. Its Views metric can combine web page and app screen consumption. Repeated views count.
Unique pageviews do not have a direct standard replacement. Analysts should define the question first. To measure how many users saw a page, use a suitable user metric with the page dimension. To measure sessions containing a page, use session logic. To measure total consumption, use Views.
Historical UA user totals should not be placed beside GA4 active-user totals without explanation. The systems use different models, identity methods and processing.
Views, Filters, Content Grouping and Session Differences
Universal Analytics views allowed separate filtered reporting environments within a property. GA4 replaced that structure with property-level reporting, data filters, custom reports and, for Analytics 360, features such as subproperties.
GA4 data filters affect incoming data and can be destructive when activated as exclusions. They should be tested before permanent use.
Content grouping still allows pages or screens to be categorized, but implementation and report access differ.
GA4 sessions do not restart automatically at midnight, and a new campaign source does not necessarily create another session. Any year-over-year comparison across UA and GA4 should therefore be labelled as directional rather than perfectly equivalent.
Commonly Confused Google Analytics Definitions
Many reporting mistakes come from choosing a familiar-sounding field without checking its exact meaning.
User vs Session vs Event vs View
A user represents a person as GA4 can identify them through the selected reporting identity.
A session is one period of activity from that user.
An event is an action during the session.
A view is a page or screen consumption event counted by the Views metric.
One user may generate three sessions. Those sessions may contain 40 events and 12 page views. There is no reason for the four totals to match.
Use users to discuss audience size, sessions to discuss visits, events to discuss interactions and views to discuss content consumption.
Source vs Medium vs Channel vs Source Group
Source identifies the origin, such as Google, a newsletter or a referring site.
Medium describes the traffic method, such as organic, email, referral, cpc or ai-assistant.
Channel applies classification rules to source, medium and other information. Organic Search and Paid Social are channels.
Source Group normalizes variations of platform-source values. It can consolidate several Facebook-related source spellings into one grouped value.
The right field depends on the required detail. Use channel for executive summaries, source and medium for campaign diagnosis, and Source Group for cleaner platform-level comparison.
Segment vs Audience vs Comparison
A segment is a flexible analytical subset used in explorations. It can be based on users, sessions or events.
An audience is a persistent group whose users qualify over time. It can support advertising activation when linked services and consent settings permit it.
A comparison is a temporary subset applied to standard reports.
Use a comparison for a quick side-by-side report, a segment for deeper historical analysis and an audience when the group needs to update and remain available for activation.
A–Z Google Analytics Definitions Quick-Reference Index
This condensed index covers common Google Analytics terminology. The longer sections above provide the calculation, scope and interpretation behind the most important entries.
A–F Google Analytics Terms
Account: The top administrative container that holds one or more properties.
Active user: A unique user who met GA4’s activity criteria during the selected period.
Advertising: The GA4 workspace for conversion, attribution and campaign-performance analysis.
AI Assistants: The channel for recognized visits from services such as ChatGPT, Gemini, DeepSeek, Copilot or Grok. It excludes Google AI Overviews and AI Mode.
Anomaly detection: Machine-learning analysis that identifies metric values outside an expected range.
Attribution: The process of assigning credit for a key event or conversion to eligible marketing interactions.
Audience: A group of users who meet defined conditions and can update as users qualify or leave.
Average engagement time: The average period during which a web page was in focus or an app was in the foreground.
Behavioral modeling: Estimation used to fill behavioral measurement gaps for users who decline analytics identifiers, when the property is eligible.
BigQuery: Google Cloud’s data warehouse, which can receive event-level exports from GA4.
Bounce rate: The percentage of sessions that were not engaged.
Calculated metric: A property-defined metric created from a formula using available metrics.
Campaign: A named marketing initiative represented through automatically collected or manually tagged acquisition data.
Cardinality: The number of unique values associated with a dimension.
Channel: A rule-based traffic category such as Organic Search, Email or Paid Social.
Client ID: A pseudonymous browser identifier used in web measurement.
Comparison: A temporary subset applied to a standard GA4 report.
Consent Mode: A method for communicating consent choices to Google tags and measurement systems.
Conversion: In current GA4 terminology, an advertising-focused action created from an Analytics event or key event for shared cross-channel measurement.
Cookie: A small piece of browser-stored data that can support pseudonymous identification and measurement when permitted.
Custom definition: A reportable custom dimension or metric created from collected custom data.
Data API: An interface used to retrieve GA4 reporting data programmatically.
Data import: A feature for adding compatible external information to Analytics.
Data retention: A property setting that controls how long certain user-level and event-level data remains available for advanced reporting.
Data stream: A web, Android or iOS source that sends data into a GA4 property.
Data thresholding: Privacy protection that limits data in a report when detailed results could risk revealing user information.
DebugView: A development report that shows events and parameters from devices using debug mode.
Device ID: A browser or app-instance identifier used to recognize a pseudonymous user.
Dimension: A descriptive attribute used to organize analytics data.
Direct: Traffic for which GA4 has no usable referring or campaign source, often represented as (direct) / (none).
Engaged session: A session lasting longer than ten seconds, containing a key event, or containing at least two page or screen views.
Engagement rate: Engaged sessions divided by total sessions.
Event: A recorded user interaction or system occurrence.
Event parameter: Additional information sent with an event.
Exploration: A flexible GA4 workspace for free-form, funnel, path, cohort and user analysis.
First user: A prefix indicating that an acquisition dimension describes how a user was originally acquired.
G–P Google Analytics Terms
Generated insights: Automated summaries of notable property changes, anomalies and trends.
Google Signals: Additional Google-based identity and advertising information available for eligible, consented users and supported features.
Google tag: The website tag framework that sends data to Google destinations.
Hostname filter: A data filter that excludes events from unapproved hostnames.
Item scope: The level of data associated with a product or service inside an ecommerce items array.
Key event: An event marked as particularly important to business success.
Landing page: The first page viewed in a session.
Lifetime value: The value associated with a user over a wider customer relationship period.
Lookback window: The period during which a marketing interaction remains eligible for attribution credit.
Looker Studio: Google’s dashboard and data-visualization product.
Measurement ID: A web-stream identifier that normally begins with G-.
Measurement Protocol: A server-to-server method for sending events to GA4.
Medium: The general traffic method, such as organic, referral, email, cpc or ai-assistant.
Metric: A quantitative measurement.
Modeled data: Estimated data used to reduce measurement gaps when direct observation is incomplete.
New user: A user who triggered first_visit or first_open during the selected period.
Not set: A placeholder used when no value was received for a dimension.
Organic Search: Traffic from unpaid search results, including Google AI Overviews and AI Mode under current channel definitions.
Page path: The path portion of a web page’s URL.
Page view: An event recording that a web page was viewed.
Parameter: A value that adds detail to an event or configuration command.
Predictive metric: A machine-learning estimate such as purchase probability, churn probability or predicted revenue.
Primary channel group: The property’s main editable channel-group dimension when the feature is available.
Property: The central GA4 collection and reporting environment inside an account.
Q–Z Google Analytics Terms
Realtime: A report showing recent activity, normally within minutes, using a limited set of fields.
Referral: Traffic from a non-ad link on another website or app.
Reporting identity: The method GA4 uses to represent users through User-ID, device ID and, when eligible, modeling.
Returning user: A user who initiated at least one previous session.
Sampling: Analysis performed on a subset of available records when a query exceeds applicable limits.
Segment: A user, session or event subset applied in Explorations.
Session: A group of interactions occurring during a period of user activity.
Session source: The source attributed to the start of a session.
Source: The specific origin of traffic.
Source Group: A normalized grouping of related source values for common platforms.
Source Platform: The platform responsible for directing traffic, such as a buying or traffic-management platform.
Stream ID: The numerical identifier assigned to a data stream.
Task Assistant: A configuration feature that organizes recommended property setup and data-quality actions.
Total users: Unique users who triggered any event during the selected period.
Traffic acquisition: Reporting focused on where sessions and visitors came from.
UTM parameter: A campaign-tracking value added to a destination URL.
Unassigned: Traffic that does not match the rules of an available channel group.
User engagement: Time during which a web page is in focus or an app screen is in the foreground.
User-ID: A business-assigned identifier for signed-in users, used to connect activity when correctly implemented.
User property: An attribute that describes a user and can persist across relevant events.
Views: The total number of web page views and app screen views in applicable reporting.
Google Analytics Glossary FAQ
Beginner Questions About Learning GA4 Vocabulary
What are the most important GA4 terms for a beginner?
Start with user, session, event, event parameter, dimension, metric, source, medium, channel, engaged session and key event.
These terms explain most standard reports. Once their relationships are clear, attribution, ecommerce and data-quality concepts become much easier.
Should I learn events or reports first?
Learn the basic event model before trying to memorize reports.
Reports are built from the data collected through users, sessions, events and parameters. Understanding collection explains why a field appears, why it might be missing and whether the report answers the intended question.
Is Google Analytics Academy free?
Analytics Academy provides free training through Skillshop. Current learning paths include beginner, intermediate and advanced material, along with Google Analytics certification.
Can I practise GA4 without access to a business account?
Google offers a demo account containing ecommerce and game data. It is useful for exploring reports and learning the interface, although some permissions, exports and advanced features are restricted.
Questions About Definitions and Reporting Discrepancies
Why do two GA4 reports show different user totals?
The reports may use different user metrics, reporting identities, dimensions, thresholds or scopes.
One report may show active users while another query uses total users. Adding a detailed demographic or acquisition dimension can also affect the displayed result.
Why do GA4 and Google Ads show different conversion totals?
Differences can come from attribution models, conversion windows, counting methods, time zones, invalid-traffic handling, consent modeling and the date assigned to a conversion.
The current integrated conversion workflow is designed to reduce discrepancies, but the two platforms can still answer different reporting questions.
Why does BigQuery not match the GA4 interface?
BigQuery contains exported event data. The GA4 interface can apply reporting identity, modeling, attribution and processed calculations.
An exact match requires reproducing the relevant interface logic, not simply counting raw event rows.
Is GA4 data accurate?
GA4 can provide reliable decision support when the implementation, consent configuration and reporting method are understood. It is not a complete census of every real-world person or interaction.
Cookie restrictions, consent choices, blocked scripts, device changes, modeling and processing rules affect the data. Accuracy should be evaluated against the intended use, implementation quality and known limitations.
Questions About Keeping the 2026–27 Glossary Current
How often does Google change Analytics terminology?
There is no fixed schedule. Google releases new fields, reports, integrations and classifications throughout the year.
Important changes in 2026 included Source Group, AI Assistant traffic measurement, hostname filters, Task Assistant, generated insights, cross-channel budgeting and expanded conversion reporting.
What is the latest release included in this glossary?
The latest dated item in the official 2026 Google Analytics release log available when this article was reviewed was June 11, 2026. It introduced Source Group changes and hostname filters.
Does the 2027 edition contain predictions?
No. “2026–27 edition” signals that the glossary is designed to remain useful through an update cycle. Unannounced product changes should not be presented as facts.
Which definition should be treated as authoritative?
For product behavior, use the current definition tied to the exact GA4 field or feature. Check the field’s scope, calculation and product status.
Internal business definitions can be stricter. For example, GA4 may count a form submission as a key event, while the company may define a qualified lead only after validation in its customer relationship management system.
Using Google Analytics Definitions to Make Better Decisions
Knowing Google Analytics terminology is not about memorizing interface labels. It is about asking better questions.
A marketing manager who selects First user source instead of Session source may credit the wrong campaign. A content team that treats bounce rate as a satisfaction score may misjudge a useful article. An ecommerce analyst who mixes item revenue with total revenue may publish an incorrect product report.
The safest workflow is to define the business question first. Then select the scope, dimension, metric, date range and reporting surface that match it.
Treat every number as the product of a collection rule and a reporting rule. Check whether the data is direct, modeled, sampled, thresholded or incomplete. Document custom events and calculations. Revisit definitions when Google changes the product.
That approach turns a glossary into a working measurement system. It also makes GA4 reports easier to explain, audit and trust.
