Exits and Entrances: The GA4 Metrics You Didn't Know You Were Missing

I was chatting with Talia Wolf on her Heart Before Cart podcast recently — we were talking about GA4 metrics that people miss — and I mentioned exits and entrances. Her reaction: “Wait, where is that?”

That felt completely fair to me. I am quite frankly annoyed that exits and entrances are hidden in GA4’s Explorations section. In Universal Analytics, exits were right there in the standard reports — one of the basics you could pull without thinking about it. When GA4 rolled out, Google moved them to Explorations-only. There’s no pre-built report waiting for you, no widget to add to your standard dashboard, no way to add it to Google Data Studio. You have to know to go looking, and you have to build the exploration yourself.

Because it’s not surfaced in the standard interface, a lot of people just don’t know this data exists. Which is a shame, because exits and entrances together tell you something critical: it isn’t just that people are leaving, but enough context to start understanding why the pattern looks the way it does.

Let me show you where to find them and what to do with the data.

I want to start here because “exit rate” has a reputation it doesn’t always deserve.

An exit in GA4 means the page where a session ended. Not that the person closed their browser in frustration, not that they had a bad experience — it means this was the last page GA4 tracked before the session stopped being recorded.

What you should know is that GA4’s default session timeout is 30 minutes of inactivity. So if someone opens a page, reads it carefully, then leaves the browser tab open on their desk for an hour, that session will time out and register as an exit on that page, even if the person came back later and kept reading. Keep that detail in mind as you interpret certain patterns in the data, and I’ll come back to it.

The bigger point is that exits are neutral. They only become meaningful in context. Some pages should have high exits:

• A contact form thank-you page (they submitted it, they’re done)
• A blog post at the end of a long read
• An order confirmation page (mission accomplished)

Exits only become concerning when they’re happening on pages where you don’t want sessions to end — mid-funnel pages, key service pages, a pricing page people are clearly considering.

Before diagnosing a high-exit page as broken, ask what job that page is supposed to do. I’ve already written more about how to give each page a clear job, now pairing that framework with exit data is where it gets really useful. If the page is doing the job you designed it for and the end of that job is an exit, high exits are fine. If it’s not, the exits are a signal worth investigating.

Here’s where it gets inconvenient: there’s no pre-built exit report in GA4. Exits and entrances only live in the Explorations section — they’re not in standard reports, and they’re not in Data Studio either (if you were hoping to build an external dashboard for this, I’m sorry to say that’s still not an option).

What you can do is build a free form exploration, which in GA4’s Explorations is just a fancy term for a table.

Here’s how:

1. Go to Explore in the left navigation in GA4

Click Blank to start fresh.

2. Add your page dimension

In the Variables panel, click the + next to Dimensions and search for your page dimension. You have two options:

• Page path and screen name — this is the cleaner choice for most sites. It strips URL parameters so that similar pages group together rather than each parameter variation getting its own row.
• Page path + query string — use this if your site uses meaningful URL parameters you want to track separately, like product filters.

For most use cases I’d start with Page path and screen name. If you want to go deeper on what these dimension options actually mean and when to use each, I’ve written a full explainer on the difference between page path and page location in GA4.

3. Add your metrics

Click the + next to Metrics and add: Views, Entrances, and Exits.

4. Build the table

Add the dimension and each metric (called “value” in Explorations) to the exploration canvas. You now have a table showing every page on your site with its view count, how many sessions started there (entrances), and how many sessions ended there (exits).

The exits metric on its own tells you something. But exits and entrances together are where things get really interesting.

Once you can see both metrics side by side, you’re actually looking at a three-way read on every page.

High entrances + high exits = a single landing page doing its job. People are arriving directly — from search, from email, from ads — and then doing the thing you want them to do, on this page. That is what you want! Don’t panic about these pages.

High views + high exits + low entrances = worth investigating. This page gets a lot of internal traffic (people navigate to it from elsewhere on your site) but then sessions end there. That’s worth asking questions about: is the page delivering what people came for? Does it have a clear next step? Does it lead anywhere useful?

The timeout tell. This one is the most nuanced, and it’s my favourite to explain. I often describe it to clients this way: a high-exit page can be a place where people start, a place where people leave, or a place where people just time out — maybe they wanted to show it to someone else later, so they left the tab open.

Think about a menu, a detailed product spec, a reference article. These are pages people find, note the information, and then leave the tab sitting open for later. High exits on those pages are almost certainly session timeouts, not actual departures. The tell is when you see both high exits and high entrances on the same page. That could be that people are randomly ending up there and leaving, or they’re returning later on after their session ends. To figure out which option you’re dealing with, look at other engagement factors on that page, such as scroll depth or views of specific pieces of content.

The combination that you want to prioritize evaluating is the opposite: high exits, low entrances, and the page isn’t designed to be a final destination. Start your investigation there.

Once you’ve downloaded your exploration data, a quick way to evaluate the data is by handing it to an AI tool and asking it to find the patterns.

When I do this, I’m not asking the AI to do the exit rate math (AIs are terrible at math!). Instead, what AI is better at is scanning a huge table of page paths and flagging the ones that look anomalous, given context it couldn’t have known otherwise.

Before you export, take a quick look at what’s in your page paths. If your URLs include anything sensitive — internal admin sections, anything that reveals site structure you’d rather keep private, anything that could be personally identifiable — clean or remove those rows first. Once you’ve cleaned up your data, try a prompt like this:

“Here’s my page paths with entrances and exits. [Tell it what you do.] [Tell it about the audiences you serve.] [Tell it about your site and business goals.] This is data from [date range]. What are you seeing?”

Just make sure to give it lots of context. Tossing data into something like Claude and saying “here’s the data from my site” works a lot better when the AI understands what your site actually does, what a normal pattern looks like for you, and whether a spike in exits on your pricing page is alarming or expected.

Once you get that initial analysis back, try some follow-up questions: which pages are getting exits you wouldn’t expect, and which high-exit pages look like they might be session timeouts rather than actual departures?

You can also add source/medium as a dimension to your exploration before downloading if you want channel-level patterns — for example, whether people arriving from email are hitting a particular page and leaving more often than organic visitors are.

Exits and entrances have been buried in GA4 Explorations since the transition from Universal Analytics, but it’s still a good metric to look at, even though it’s annoying to find now. There’s no automated alert when exit rates spike, no pre-built report to check. You have to go build the exploration yourself — and because that takes a step most people haven’t taken, this data often just sits unchecked.

If you’re trying to understand what’s actually happening on your site beyond views and sessions, exits and entrances give you something specific: a page-by-page read of where sessions are ending and where they’re starting, with enough context to tell the difference between a page doing its job and one that needs attention.

The exploration build takes about five minutes. Start with your top 20 pages by views and see what the entrance and exit numbers tell you. I think you’ll find it surfaces things your standard reports might have been glossing over.

Have questions about this? Check out my conversation with Talia Wolf on YouTube, and drop your questions in the comments!