Funnel Exploration

Funnel exploration in GA4 visualizes user drop-off across a defined sequence of steps. Unlike standard reports, funnels in GA4 Explore can use any event or page view as a step, apply segment breakdowns, and switch between open and closed completion modes.

GA4 funnels are event-based. Each step is an event occurrence, optionally filtered by event parameters. This makes them more flexible than Universal Analytics goal funnels, which were page-view-based.

Creating a funnel

Go to Explore → Create new exploration → Funnel exploration.
In the Variables panel, add any segments or dimensions you plan to use.
In the Tab Settings panel, click Edit funnel (pencil icon next to Steps).
Click Add step for each step in your funnel.
For each step:
- Choose the condition: event name, page/screen view, or a custom event parameter
- Optionally add AND conditions to filter the step (e.g., event name = purchase AND value > 50)
- Give the step a descriptive name
Set the funnel type: open or closed.
Optionally set elapsed time between steps.
Click Apply to generate the funnel.

Open vs. closed funnels

This is the most important choice in funnel configuration.

Closed funnel: Users must enter the funnel at step 1. If a user visited the cart (step 2) before visiting the product page (step 1), they are not counted. The funnel tracks only the intended linear path.

Open funnel: Users can enter at any step. A user who goes directly to checkout without a product page view is counted at the checkout step. Open funnels show absolute drop-off from each step but do not enforce a linear order.

When to use closed: You want to understand the linear conversion path and enforce order. E-commerce purchase funnel: view_item → add_to_cart → begin_checkout → purchase. You only want to count users who followed this exact sequence.

When to use open: You want to see where users are in the funnel regardless of how they got there. Multi-channel signup flow where some users skip steps because they already completed them.

Step conditions

Each step can be defined by:

Event name: the simplest condition — any occurrence of a named event qualifies.

Page/screen view: users who viewed a specific page (by path or title). Useful for page-based funnels on single-page application sites where route changes fire page_view.

AND conditions: add filters to a step. For example:

Event = purchase AND parameter value > 100 (only count high-value purchases)
Page view AND page path contains /checkout/ (all checkout-related pages)
Event = form_submission AND parameter form_id = signup (specific form)

OR conditions within a step: If multiple page paths count as the same step (e.g., any product detail page), use OR conditions: page path contains /product/ OR page path contains /item/.

Elapsed time between steps

Set a time constraint on how long users have to move between steps. Options range from 5 minutes to 90 days.

Use elapsed time when:

Your funnel has a natural time window (checkout should complete within 1 hour)
You want to exclude users who abandoned for a long period and returned
You are measuring fast-path completion vs. slow-path completion

Leave it at the default (no constraint) when:

Users commonly return across sessions to complete steps (B2B evaluation funnels)
Your funnel is multi-session by design

Segment breakdowns

After building the funnel, apply a segment breakdown to compare the funnel across user groups:

In the Tab Settings panel, find Segment comparisons and add up to 3 segments. The funnel will show parallel bars for each segment, making it easy to compare:

Mobile vs. Desktop conversion rates
Paid traffic vs. Organic conversion rates
New users vs. Returning users conversion rates

This is one of the most powerful uses of funnel exploration — identifying which segments have significantly different drop-off at specific steps.

Analyzing abandoned users

At each funnel step, GA4 shows the number of users who dropped off. Click the drop-off number for a step to create a segment of users who completed that step but did not complete the next one.

Use this abandoned segment to:

Apply it to a path exploration to see what they did instead
Cross-reference with other dimensions (device, country, traffic source) to find patterns
Build a remarketing audience for these users

Funnel exploration examples

E-commerce purchase funnel

Step 1: Event = view_item
Step 2: Event = add_to_cart
Step 3: Event = begin_checkout
Step 4: Event = purchase
Type: Closed
Elapsed time: Within 1 day

This measures the linear purchase path. Closed mode ensures only users who followed the intended sequence are counted.

Step 1: Page view — page path contains /pricing
Step 2: Page view — page path contains /signup
Step 3: Event = form_submission AND form_id = trial_signup
Step 4: Event = onboarding_complete
Type: Open
Elapsed time: Within 30 days

Open because users may view pricing multiple times before signing up, and onboarding may span multiple sessions over days.

Content engagement funnel

Step 1: Event = page_view AND content_type = blog_post
Step 2: Scroll event with percent_scrolled = 90
Step 3: Event = click AND link_text = Subscribe
Step 4: Event = form_submission AND form_id = newsletter
Type: Closed
Elapsed time: Within same session

Exporting funnel data

You cannot export the funnel visualization directly, but you can:

Switch the visualization to a table view for the underlying data
Export the table as CSV

For automated funnel reporting, replicate the funnel logic in BigQuery (see Funnel Analysis in SQL).

Common mistakes

Using open funnel when closed is needed

An open funnel on an e-commerce checkout path will show users entering at the “purchase” step without going through “add_to_cart”. These are likely direct link purchases, API orders, or data quality issues — not users who completed the intended funnel path. Use closed for strict purchase funnels.

Not filtering steps to the relevant event parameters

A step defined as “Event = page_view” in a funnel will match ANY page view. If your funnel step should be the product detail page, add a condition: Event = page_view AND page path contains /product/. Without the filter, every page view qualifies and the funnel loses meaning.

Ignoring elapsed time on long funnels

A B2B demo request funnel where someone viewed pricing 90 days before requesting a demo is a valid conversion. Without setting elapsed time (or setting it to 90 days), this user is counted. With a 1-hour elapsed time, they are not. Match the elapsed time to your actual buying cycle.

Building funnels with too many steps

A 10-step funnel where most steps have 90%+ completion rates hides the meaningful drop-off. Focus on 3-6 steps where meaningful decisions occur. You can always drill down by removing high-completion steps or viewing step-level data separately.

Path Exploration Explore what abandoned funnel users do next.

Funnel Analysis in SQL Replicate and extend GA4 funnels in BigQuery without sampling.

Explore Reports Overview of the full Explore workspace and all exploration types.