GA4 Metrics Reference: What to Track, How to Define It, and When Benchmarks Matter

Overview

The fastest way to get lost in GA4 is to start with the interface instead of the measurement plan. GA4 is event-based, not session-based in the old Universal Analytics sense, so familiar terms can behave differently and some teams carry over outdated expectations. That is why a useful GA4 KPI guide starts with definitions before charts.

At a high level, a metric in GA4 is a quantitative value: users, sessions, views, conversions, purchase revenue, engagement rate, and many others. A dimension adds descriptive context: source, medium, landing page, device category, page path, country, or campaign. Good reporting combines both. “Conversions” alone is not yet insight; “form_submit conversions by landing page and source/medium” is closer to something operational.

Since July 1, 2023, standard Universal Analytics properties stopped processing data, and UA access for 360 continued until July 1, 2024. GA4 is now the active Google Analytics standard, which means current reporting, implementation choices, and stakeholder training all need to reflect GA4’s model rather than UA habits. The safest evergreen interpretation is simple: do not ask whether a GA4 metric matches an old UA number exactly. Ask whether the metric is well defined, consistently implemented, and fit for the decision you need to make.

If you only remember one principle from this article, let it be this: in GA4, the best metrics are not the most available metrics. They are the metrics tied to a clear action. For each metric you track, define four things:

• What it measures: the plain-language meaning.
• How GA4 calculates it: the reporting logic or event dependency.
• What can change the number: implementation edits, filters, consent choices, attribution settings, or cross-domain setup.
• What decision it supports: budget, content, UX, funnel, or channel optimization.

That framework keeps dashboards stable even when platform features evolve.

Core framework

Use this section as a reference for choosing what to track in GA4. Rather than listing every available field, it focuses on the metrics most teams return to repeatedly and the caveats that matter in reporting.

1. User metrics: who is showing up

Users are usually one of the first GA4 metrics stakeholders ask for, but they need careful framing. In practice, users answer reach questions: how many people your site or app reached during a period. This is useful for trend monitoring, campaign review, and comparing channels at a high level.

How to define it: a count of users active within the reporting period, based on GA4’s identity and reporting settings.

What can change it: identity spaces, consent behavior, device switching, cross-domain configuration, and changes to tagging coverage.

Best use: directional audience trend analysis, not precise people counting across every environment.

Related user metrics like new users and returning users are valuable when you care about acquisition quality and loyalty. But they are only as trustworthy as your identity continuity. If cross-domain tracking is broken or consent suppresses identifiers, these splits can become less stable.

2. Session metrics: when visits still matter

GA4 moved to an event-based model, but sessions still matter. Sessions help answer visit-level questions: how often people start journeys, how traffic patterns shift, and how efficiently landing pages or campaigns generate site visits.

How to define it: a group of user interactions within a session window, as counted by GA4.

What can change it: session timeout rules, campaign parameter changes, broken referrals, cross-domain setup errors, and implementation gaps on landing pages.

Best use: acquisition reporting, landing page analysis, and top-level traffic diagnostics.

Do not force sessions to answer user-quality questions on their own. A campaign can generate more sessions and still perform worse if those visits do not engage or convert.

3. Views and page-level metrics: what content gets seen

Views remain one of the simplest google analytics metrics to understand. They answer visibility questions: which pages, screens, or content groups are being consumed.

How to define it: the total number of page or screen views recorded by GA4.

What can change it: duplicate page_view firing, single-page application routing issues, missing virtual pageviews, and GTM configuration mistakes.

Best use: content performance, information architecture review, and validating site changes.

Views are especially useful when paired with landing page, page path, source/medium, and conversion metrics. On their own, they often over-reward pages that attract attention without moving users forward.

4. Engagement metrics: the GA4 replacement for shallow visit quality

GA4 emphasizes engaged sessions, engagement rate, and average engagement time. These are often more informative than older bounce-centric habits because they try to capture meaningful interaction, not just short exits.

How to define it: engagement metrics reflect whether users meaningfully interacted during sessions and how long they were actively engaged.

What can change it: automatic event behavior, custom event firing, content format differences, and site UX changes.

Best use: evaluating content depth, landing page quality, and audience fit.

Still, engagement metrics should not be treated as universally comparable benchmarks. A documentation site, a lead-gen homepage, and a checkout flow have different ideal engagement patterns. Longer is not always better. In some flows, a shorter path to conversion is healthier.

5. Event counts: the foundation of GA4 reporting

Because GA4 is event-based, event count is not just a technical metric. It is the backbone of custom measurement. Scrolls, video starts, file downloads, sign-up attempts, add_to_cart, generate_lead, purchase, and custom interactions all live here.

How to define it: the number of times a tracked event fired.

What can change it: duplicate tagging, trigger logic errors in Google Tag Manager, naming inconsistencies, missing parameters, and partial deployment across templates or environments.

Best use: validating behavior tracking and building custom funnel analysis.

When teams ask “what to track in GA4,” event design is usually the real question. If your events are sloppy, every downstream metric becomes harder to trust. For a strong baseline, align event naming, parameter naming, and conversion selection before you build executive dashboards. If you need a broader implementation template, see GA4 Events Checklist: What to Track on Every Website.

6. Conversion metrics: what success actually looks like

Conversions are among the most important GA4 metrics because they turn activity into outcomes. But in practice, “conversion” is often under-defined. Some teams mark too many events as conversions. Others lump micro and macro outcomes together and lose reporting clarity.

How to define it: a designated key event that represents a business-relevant action.

What can change it: event configuration, duplicate fires, thank-you-page logic, consent settings, server-side routing, and attribution model settings.

Best use: channel performance, funnel reporting, campaign review, and executive KPI dashboards.

A helpful operating model is to split conversions into tiers:

• Primary conversions: purchases, qualified leads, booked demos, completed applications.
• Secondary conversions: newsletter signups, account creations, pricing page CTA clicks when they have predictive value.
• Diagnostic events: start_checkout, form_start, video_progress, error states.

This keeps reporting readable and prevents inflated success rates.

7. Ecommerce metrics: where implementation quality becomes obvious

For ecommerce analytics, the core metrics are typically purchase revenue, items purchased, add-to-cart rate, begin_checkout, and purchase conversion metrics. GA4 ecommerce tracking can be powerful, but only if item-level parameters and transaction IDs are implemented consistently.

How to define it: revenue and commerce actions based on recommended ecommerce events and parameters.

What can change it: missing currency values, duplicate purchase events, broken transaction deduplication, refund handling, and incomplete item arrays.

Best use: merchandising, funnel analysis, campaign ROI review, and site UX improvement.

If purchase data looks unstable, audit implementation before debating benchmarks. A revenue chart with duplicate transactions is not a performance signal; it is a tagging issue.

8. Acquisition metrics: where attribution starts, not ends

Source, medium, campaign, landing page sessions, users, and conversions are still core reporting tools. They are essential for understanding channel mix and evaluating campaign tracking quality.

How to define it: metrics segmented by traffic source dimensions and GA4 attribution logic.

What can change it: UTM errors, auto-tagging conflicts, missing referral exclusions, cross-domain tracking gaps, and changing attribution settings.

Best use: channel optimization, campaign QA, and comparing acquisition efficiency.

Many attribution disputes are really taxonomy problems. If UTMs are inconsistent, GA4 cannot rescue the analysis. Keep naming conventions documented, and audit them regularly.

9. Retention and lifecycle metrics: useful, but context-heavy

Retention-style metrics matter most for SaaS, media, membership, and product-led businesses. They help answer whether users come back and whether acquisition creates durable value.

How to define it: metrics that measure repeat activity over time.

What can change it: identity instability, app/web stitching, login state, and incomplete lifecycle event tracking.

Best use: product adoption and audience loyalty reporting.

For many marketing teams, retention is worth monitoring but should not replace conversion and revenue metrics in the main dashboard.

10. Benchmarks: when they help and when they mislead

GA4 benchmark metrics are most useful as prompts, not verdicts. They can tell you a number looks high or low relative to a peer set, past period, or business model, but they rarely explain why.

Use benchmarks when:

• You compare your site against its own historical baseline.
• You compare similar page templates or funnels.
• You compare businesses with similar models, traffic mix, and conversion definitions.

Use caution when:

• The implementation changed recently.
• Consent settings changed measurement coverage.
• Traffic mix shifted heavily because of one campaign.
• The benchmark ignores business type, sales cycle, or site purpose.

For more benchmark-oriented KPI ideas, see GA4 Metrics Benchmark List: The KPIs Marketers Track Most and GA4 Dashboard Metrics by Business Type: SaaS, Ecommerce, Lead Gen, and Content Sites.

Practical examples

These examples show how to turn raw GA4 metrics into reporting that supports decisions.

Lead generation site

A B2B lead-gen site should not stop at users and sessions. A stronger scorecard might include:

• Users and sessions by source/medium
• Landing page engagement rate
• Form start event count
• Generate_lead conversions
• Lead conversion rate by landing page and campaign

Decision supported: which channels and pages create qualified intent, not just traffic.

What to watch: form events firing on validation errors, duplicate submissions, or thank-you-page reloads.

Ecommerce store

A retail site needs funnel visibility, not only revenue totals. A practical dashboard could include:

• Sessions by channel
• Product detail views
• Add_to_cart rate
• Begin_checkout count
• Purchase conversions
• Purchase revenue
• Revenue by source/medium and device category

Decision supported: whether problems sit in acquisition, product interest, checkout UX, or device-specific friction.

What to watch: missing item parameters, duplicate purchase events, and checkout steps not firing consistently.

Content or publishing site

A content site often over-focuses on views. A better reporting set includes:

• Users and views by landing page
• Engaged sessions per article group
• Average engagement time
• Scroll or article completion events
• Newsletter signup conversions

Decision supported: which topics attract search traffic and which topics create deeper audience value.

What to watch: scroll events that fire too easily and make every article look equally successful.

SaaS or product-led site

A SaaS site usually needs a bridge between marketing and product behavior. Useful metrics include:

• Users and sessions by acquisition source
• Account creation conversions
• Trial start events
• Key activation events
• Return user trends

Decision supported: whether marketing channels attract users who actually activate, not just sign up.

What to watch: poor event naming across web and app properties, which can break lifecycle reporting.

If your team is still sorting out tool boundaries, Google Tag Manager vs GA4: What Each Tool Does and When You Need Both is a useful companion piece.

Common mistakes

Most GA4 reporting problems are not caused by the dashboard. They come from unclear definitions, weak implementation, or expectations borrowed from older analytics models.

Tracking everything as a KPI

Not every measurable event deserves executive attention. Keep top-level dashboards focused on business outcomes and a few supporting diagnostics.

Assuming GA4 metrics are direct replacements for UA metrics

GA4 uses a different model. Comparing old and new numbers as if nothing changed usually creates confusion. Compare trends within GA4 once the setup is stable.

Ignoring implementation dependencies

Metrics such as conversions, revenue, and engagement all depend on clean event tracking. If GTM logic is inconsistent, reports will drift. Build a QA routine, not just a launch checklist.

Using benchmarks without checking definitions

A conversion rate benchmark means little if one team counts newsletter signups and another counts qualified pipeline starts. Always benchmark like with like.

Blending diagnostic and success metrics

Events like form_start or add_to_cart matter, but they are not equivalent to final outcomes. Keep them distinct so teams can see where drop-offs happen.

Under-documenting custom events

GA4 custom events are powerful, but undocumented naming leads to long-term reporting debt. Maintain an event dictionary with event names, parameters, trigger logic, owners, and reporting usage.

Forgetting privacy and consent effects

Consent choices can materially affect observable traffic and conversions. Treat measurement coverage as part of dashboard interpretation, not a separate legal footnote.

A useful discipline here is to review analytics output the way a critical reviewer would: question definitions, edge cases, and unexplained jumps before sharing conclusions. The article Critique for analytics: borrow Microsoft’s reviewer model to harden measurement outputs offers a good mindset for this kind of QA.

When to revisit

This metrics reference should be revisited whenever the measurement environment changes. In GA4, metric definitions may stay familiar while the inputs behind them shift. That is why stable reporting depends on periodic review.

Revisit your GA4 KPI guide when any of the following happens:

• Your implementation changes: new GTM container logic, server-side tracking, cross-domain updates, or ecommerce tagging edits.
• Your site or app changes materially: redesigns, new templates, SPA routing changes, new checkout flows, or form vendor replacements.
• Your consent or privacy setup changes: banner logic, Consent Mode updates, or changed data collection defaults.
• Your attribution inputs change: new UTM rules, auto-tagging updates, paid media expansion, or CRM integration changes.
• Your business model changes: new subscription plans, new lead qualification rules, new conversion definitions, or product-line expansion.
• GA4 reporting features change: interface updates, new calculated metrics, or revised configuration options.

To make this practical, run a simple quarterly metrics audit:

1. List your top 10 reported metrics.
2. Write the current definition of each in one sentence.
3. Identify the event or configuration dependency behind each metric.
4. Check whether any implementation, site, or consent changes happened since the last review.
5. Compare current trends to recent history and flag unexplained breaks.
6. Confirm that each metric still supports an active business decision.
7. Remove or demote metrics that are noisy, redundant, or poorly trusted.

That process keeps dashboards lean and credible.

In the end, the best answer to “what to track in GA4” is not a giant universal list. It is a maintained set of definitions that your team understands, trusts, and can act on. Start with users, sessions, views, engagement, events, conversions, revenue, and acquisition. Then tailor the final KPI set to your business model, implementation maturity, and reporting decisions. If you treat metric definitions as living documentation rather than one-time setup, your GA4 reporting will stay useful long after the interface changes again.