The 7 GA4 metrics that actually matter (and the 5 that don't)

Open GA4 and you can choose from roughly 200 metrics across the various reports. Some are essential. Some are vanity. Some are quietly misleading. The actual problem most small businesses face isn't 'we don't have enough data' — it's 'we don't know which 12 numbers to focus on and which 188 to ignore.'

This is the short, opinionated list. Seven metrics that genuinely matter for the average business, five that almost nobody should pay attention to, and a quick note on the ones that are conditionally useful. Each one is explained in plain English with what it actually tells you about your business — not what GA4's tooltip says.

The 7 metrics that matter

1. Active users

What it is: the count of unique humans (well, unique browsers) who interacted with your site in the time period you're looking at. "Active" in GA4 means they did at least one of: triggered a key event, viewed a second page, or stayed at least 10 seconds.

Why it matters: this is the closest thing GA4 has to "how many real people came to my site this week." It's the headline number for understanding whether your audience is growing.

The gotcha: GA4 also shows you "Total users" and "New users" and a few other flavors. They are not the same number. Pick "Active users" as your standard and stop comparing across the others — that's where almost everyone gets confused.

2. Sessions

What it is: a session is one continuous visit. One person who comes back three times in a week generates three sessions but one user.

Why it matters: when you compare sessions to users, the ratio tells you something. Sessions/Users above 1.3 means people are coming back; close to 1.0 means most visitors come once and leave. For a blog or content site that ratio should be 1.5+; for a one-time-purchase site, 1.1 is fine.

3. Engaged sessions and engagement rate

What it is: an "engaged session" is one where the visitor stayed at least 10 seconds, viewed a second page, or triggered a key event. Engagement rate is the percentage of sessions that meet that bar.

Why it matters: this is the metric that replaced "bounce rate" in GA4, and it's better. A 65% engagement rate means 65% of visits had at least minimal interest; below 40% suggests either bad traffic sources, slow load times, or wrong-fit content. Use this — not bounce rate — to evaluate whether your visitors care about your site.

4. Key events (the new name for "conversions")

What it is: any specific action you've told GA4 to count as important — form submissions, button clicks, file downloads, purchases, video plays. You configure these manually under Admin → Events.

Why it matters: this is the only metric on the page that has anything to do with whether your website is actually helping your business. Sessions and engagement are inputs; key events are outputs. If you have not set up at least one key event in GA4, the whole tool is just telling you about traffic that may or may not be doing anything for you.

Setup tip: at minimum, configure key events for "contact form submitted" and "primary call-to-action clicked" within your first week. Without these, GA4 is purely vanity tracking.

5. Sessions by Default Channel Group (in Acquisition)

What it is: a breakdown of where your sessions come from — Organic Search, Direct, Email, Social, Paid Search, Referral, Organic Social, etc. Found in Reports → Acquisition → Traffic acquisition.

Why it matters: this is the single most actionable view in GA4. If 80% of your traffic comes from Organic Search and Google ships an algorithm update that hurts you, you have a real problem. If your traffic is balanced across 5 channels, you're more resilient. The distribution tells you the health of your business model, not just your website.

6. Top landing pages (Reports → Engagement → Landing page)

What it is: the pages that are the first thing visitors see when they arrive at your site.

Why it matters: this is your site's actual "front door," and it is almost never your homepage. For most content-driven sites, 60-80% of traffic lands on a blog post or a specific service page. Optimizing those pages — the one with conversion CTAs, signup forms, and clear next steps — moves the needle more than optimizing your homepage.

7. Conversion rate (key events ÷ sessions)

What it is: of all the sessions, what percentage triggered your key event. GA4 doesn't show this directly as a single metric, but it's easy to calculate: divide key events by sessions for the same period.

Why it matters: conversion rate isolates whether your site converts visitors well, separate from how much traffic you're getting. A site that gets 100 visitors and converts 5 is healthier than a site that gets 10,000 visitors and converts 50. Same absolute number, vastly different business.

Benchmarks vary wildly by industry, but typical ranges: B2B lead-gen sites 2-5%, e-commerce 1-3%, content sites with newsletter signups 1-2%. If you're below those, fix the site before buying more traffic.

The 5 metrics that don't matter (much)

1. Bounce rate (the old one)

GA4 actually deprecated bounce rate from the default views, but you can still add it back. Don't. The GA4 bounce rate definition is "100% minus engagement rate" — meaning a high bounce rate on a blog post (where someone reads the whole article in one minute and leaves satisfied) is technically a "bounce" but isn't a problem at all.

Use engagement rate instead. It tells you the same story without the misleading framing.

2. Average engagement time

GA4 shows you the average engagement time per session, per page, etc. It sounds useful — "are people reading my content?" — but it's wildly distorted by outliers. One visitor who leaves a tab open for 4 hours will skew your average. The median would be informative; the mean isn't.

If you care about engagement, look at scroll depth (configure it as a key event) or specific page-completion events. Pure time-on-page is mostly noise.

3. Real-time

Real-time is fun to watch — there are 12 people on the site right now, three of them are in California, look at them go. But the sample is too small to make any decision from. Real-time is good for confirming your tag fires and that's it. Close the tab.

4. Total users vs. new users (separately)

GA4 lets you display these separately on dashboards. Don't. "New users" is roughly meaningless without context, and "total users" inflates over long periods in a way that doesn't reflect business reality. Stick to Active Users and you'll save yourself confusion.

5. Average session duration

Same problem as average engagement time, but worse — GA4's session-duration calculation has known edge cases where short sessions get measured as zero. A high "average session duration" can simply mean you have outlier long sessions, not that your content is engaging. Ignore.

Conditionally useful: metrics that matter if X

A few that aren't on either list because they only matter for specific business models:

• Ecommerce revenue and transactions. Critical if you sell online; ignore otherwise. Configure GA4 ecommerce tracking through GTM or your platform's plugin.
• Page load time / Core Web Vitals. Matter for SEO because Google ranks faster sites higher. Found in Search Console, not GA4. Worth a monthly check.
• Device category. Only matters if you discover something surprising. If mobile is 75% of your traffic and you've never tested the mobile experience, that's an emergency. Otherwise, low priority.
• Geography. Useful if you're spending on geo-targeted ads. Otherwise, just curiosity.
• Search queries (via Search Console integration). Massively useful if you care about SEO. Connect Search Console to GA4 and you can see which queries drove your organic sessions.

The Monday-morning version

If you only have 60 seconds a week, look at exactly four things, in this order:

• Active users (week-over-week change). Are people showing up at the same rate as last week?
• Key events (week-over-week change). Are they doing what you want them to do?
• Top channel performance. Did one specific source change significantly?
• Top landing page. Where are the people you got actually arriving?

Those four answers give you 90% of what you need to know about your website most weeks. Everything else is noise unless something specific catches your attention.

GA4 measures everything. Your business cares about four things. The skill is in knowing which four — and ignoring the other 196.

Try Plainly free — see your four numbers in plain English

What we just walked through is essentially the Plainly worldview: pull the few metrics that matter, compare them week-over-week, explain what the changes mean, suggest what to try next. The free demo on the homepage will show you exactly this kind of read-out for a sample dataset — no signup required, no card. If you like the format, you can connect your real GA in one click.