Reference

Analytics glossary

By the editors of Analytics1.comLast reviewed 19 October 2026

The short answer

Short, quotable definitions of the analytics terms small businesses and agencies meet most, with GA4's current meanings where they differ from older tools. Each entry links to the guide or calculator that goes further.

Attribution model: The rule that decides which marketing touchpoints get credit for a conversion: first click, last click, linear, time decay, position-based or data-driven. GA4 uses data-driven attribution by default. See attribution models explained.
Attribution window (lookback window): How far back a platform looks for an ad click or touchpoint to credit with a conversion. In GA4 the default is 30 days for acquisition key events (first visits) and 90 days for all other key events; Google Ads sets a conversion window per conversion action.
Bounce rate (GA4): In GA4, the share of sessions that were not engaged: 100% minus the engagement rate. It is not the same as Universal Analytics' bounce rate, which counted single-page sessions.
Break-even ROAS: The return on ad spend at which ads pay for themselves on the first order: 1 divided by the margin before ad spend. At a 40% margin it is 2.5. Try the break-even ROAS calculator.
CAC (customer acquisition cost): Sales and marketing spend in a period divided by the new customers won in it. Report it per channel as well as blended. See the LTV to CAC calculator.
Call tracking: Assigning phone numbers to marketing sources, or to individual website visitors, so each inbound call can be traced to the ad, keyword or page that produced it. See call tracking explained.
Churn rate: The share of customers (logo churn) or recurring revenue (revenue churn) lost in a period, usually a month: customers lost divided by customers at the start of the period.
Consent Mode: Google's API for telling its tags whether a visitor agreed to analytics and advertising storage. Version 2 uses four signals: analytics_storage, ad_storage, ad_user_data and ad_personalization. See Consent Mode v2 explained.
Conversion (Google Ads): An action Google Ads counts and can optimise bidding toward, such as a purchase, lead or qualified call. It can be imported from a GA4 key event or recorded by the Google Ads tag.
Cookieless analytics: Web analytics that counts visits without setting cookies or storing identifiers on the device, such as Plausible or Fathom. Whether it needs consent depends on the jurisdiction and exactly what it processes. See GA4 alternatives.
Data-driven attribution: An attribution model that uses machine learning on your account's own conversion paths to share credit among touchpoints, rather than a fixed rule. It is the default in GA4 and Google Ads.
Data thresholding: GA4 withholding report rows with small user counts when Google signals is on, to stop individuals being identified. Switching reporting identity to device-based usually reduces it.
Dynamic number insertion (DNI): A script that swaps the phone number on a website for a tracking number from a pool, per visitor, so a call can be matched to that visitor's session, source and keyword.
Engaged session: A GA4 session that lasted longer than 10 seconds, included a key event, or had at least two page or screen views. Engagement rate is engaged sessions divided by all sessions.
Enhanced measurement: GA4 events collected without extra code once switched on in the data stream: page views, scrolls, outbound clicks, site search, video engagement, file downloads and form interactions. See GA4 event tracking.
Event (GA4): Any interaction GA4 records, from a page view to a purchase. Events carry parameters (details such as value or page) and are automatically collected, enhanced measurement, recommended or custom.
First-party data: Data a business collects directly from its own customers and visitors: orders, CRM records, on-site behaviour and survey answers.
GCLID: The Google Click Identifier that Google Ads adds to landing-page URLs when auto-tagging is on. Storing it with a lead lets you send offline conversions, such as qualified calls or closed deals, back to Google Ads. On some iOS traffic Google uses gbraid or wbraid instead.
Key event: A GA4 event marked as important to the business, such as a lead or purchase. Google renamed GA4 "conversions" to key events in March 2024; importing a key event into Google Ads makes it a conversion.
LTV (customer lifetime value): The gross profit a customer brings over their relationship with you. For subscriptions: monthly revenue per account times gross margin, divided by monthly churn. See SaaS metrics.
Margin of error: The range around a survey or test result within which the true value plausibly lies, at a stated confidence level. About 385 responses give a 5% margin at 95% confidence for a large population. Try the survey sample size calculator.
Net Promoter Score (NPS): Percentage of promoters (9 to 10 on a 0 to 10 likelihood-to-recommend scale) minus percentage of detractors (0 to 6), from -100 to 100. Try the NPS calculator.
p-value: In an A/B test, the probability of seeing a difference at least as large as the one observed if the versions actually performed the same. Below 0.05 is the usual threshold for significance. See the A/B test calculator.
ROAS (return on ad spend): Revenue attributed to ads divided by ad spend. A ROAS of 4 means $4 of revenue per $1 spent. It ignores product costs, so compare it with your break-even ROAS.
Server-side tagging: Running tags on a server you control (for example Google Tag Manager's server container) instead of in the visitor's browser, so you decide what each vendor receives. It does not remove the need for consent. See server-side tagging.
Session (GA4): A group of interactions by one user, started by the session_start event and ended after 30 minutes of inactivity by default. GA4 does not start a new session at midnight or when the campaign changes.
Statistical significance: A result unlikely to be due to chance at a chosen threshold, usually 95% confidence (p below 0.05). Decide the sample size before a test and check once; peeking inflates false wins.
UTM parameters: Tags added to a link (utm_source, utm_medium, utm_campaign, utm_term, utm_content) that tell analytics where a visit came from. Build them consistently with the UTM builder.
Visitor identification (reverse IP lookup): Matching the IP address of a website visit to the company that owns the network, to show B2B teams which organisations are visiting. It identifies companies, not people. See how to identify companies visiting your website.
Zero-party data: Information a customer deliberately gives you, such as quiz answers, preferences or survey responses, as opposed to data inferred from behaviour. See interactive content.