Digital Marketing

GA4 Funnel Exploration: How to Find Conversion Leaks Step by Step

Talha Aslan 18 min read 3 views

Short answer

What is GA4 funnel exploration and how does it find conversion leaks?

GA4 funnel exploration is a report that shows the steps visitors take toward a goal and how many people continue or drop off at each step. You define steps with events, find where the loss is largest, and then use breakdowns and segments to narrow down the reason.

Keep reading for the full details and examples

So this guide shows you how to build a funnel from scratch and read it correctly. We checked each setting name against Google's official funnel exploration help page. Menu labels can change over time, so if your interface shows a similar name, that is expected.

Our goal is to move from "how many people came" to "where did we lose them." The overall conversion rate tells you a problem exists. The funnel, in turn, tells you where it sits.

What do you need before you open GA4 funnel exploration?

A funnel is only as accurate as the events behind it. So your first job is to check your measurement. For every step you plan to add, first confirm that GA4 actually collects that event.

Prepare these items first:

  • Event inventory: Write down the names of events such as page view, view item, add to cart, begin checkout and purchase.
  • Setup verification: Trigger the events with a real test flow. For missing conversions, see GA4 conversions missing: causes and fixes.
  • General setup check: The guide Is your GA4 setup correct? 50 checks lists the typical mistakes.
  • Goal clarity: Make sure the end of the funnel is a clear business outcome.

Also, funnel analysis gets noisy when visitor counts are very low. In that case, extend the date range or, alternatively, cut the number of steps. Measurement matters, but so does the page itself. If you want clean funnel steps, plan analytics needs from the start of any business website design project.

Which funnel should you build first?

However, no site has just one funnel. Start by choosing the main flow that brings in revenue. Also, trying to build all of them at once produces a report nobody can read.

The table below shows starter funnels for common business models. Then adapt the event names to your own setup.

Business modelExample funnel stepsWhere leaks often appear
EcommerceView item, add to cart, begin checkout, purchaseBetween cart and checkout
Service or lead formPage view, open form, start form, submitBetween form start and submit
Membership or SaaSHome, pricing page, start signup, finish signupAfter the pricing page
Content or newsletterArticle view, signup box, subscribeAt the moment the box appears

So treat this table as a starting point, not a rule. The real leak sits in a different place on every site, because visitor behavior changes with industry, offer and traffic source. For the general logic, see what is a conversion funnel and how to build one.

How do you open GA4 funnel exploration step by step?

According to Google's help page, you create a funnel exploration from the Explore area by choosing the Funnel exploration template. In short, the steps below summarize the logic.

  1. Open the explorations area in your GA4 property and start a new exploration.
  2. Choose the funnel exploration template.
  3. Give the funnel a clear name, for example "Product page to purchase."
  4. In the Steps section, define the first step and its condition.
  5. Add the next step and continue in order, checking each condition.
  6. Apply your settings and review the table.

However, menu paths and setting names can change between releases. So follow the logic above and pick the closest match in your interface.

How do you define the steps of a GA4 funnel exploration?

Each step has a name and conditions based on events or dimension values. Per the official page, you can add up to ten steps, and you cannot use metrics as conditions. Therefore, the step structure is your most important design decision.

Follow these principles when you define steps:

  • Few but meaningful steps: Five to seven steps fit most flows. More steps make the chart hard to read.
  • Clear conditions: Each step should capture one event, or one event combined with a page value.
  • Consistent naming: Use plain names your whole team understands.
  • Logical order: Steps should follow the real path visitors take.

For example, on a service site, you define "contact page view" by combining the page view event with a page path condition. Then you capture "form submit" with the form submission event.

How do you combine event and page conditions in one step?

At first, a step can rely on the event name alone. However, an event name is often not enough, because the same event fires on many pages. The page view event, for example, runs on every page. So you add a dimension condition such as page path to the event condition.

According to the official help page, you base conditions on events or dimension values, and when you set several conditions, all of them must be met. That rule gives you two benefits:

  • Narrowing: "Page view" plus "page path contains pricing" captures only the pricing page.
  • Avoiding double counting: You separate two flows that share an event by adding a dimension condition.

Also, watch out for exact matching. If your page paths end with a slash or carry parameters, make sure your condition covers those variants. Otherwise the step looks smaller than it is and creates a false leak.

Then, after you save the step, check that the numbers make sense. The first step should land close to the page view count you expect. This quick test catches most condition mistakes early.

Should steps follow each other directly or indirectly?

First, for each new step, you choose how it links to the previous one. Google's help page offers two options: "is indirectly followed by" and "is directly followed by." With indirect linking, other actions can happen in between. With direct linking, the step must come immediately after.

So in most sales funnels, indirect linking is more realistic. A visitor may add a product to the cart, browse other items and still buy. However, direct linking would count that person as lost.

In contrast, direct linking only makes sense when the order is strict. For example, you might want tight tracking inside a short checkout sequence where any detour means a break.

There is also a time limit. Also, with the "Within" setting, you can require that a step happens within a certain time. For long buying cycles, keep that window wide. Otherwise you push slow but genuine buyers out of the funnel.

Should you make the funnel open or closed?

First, the funnel exploration has a "Make open funnel" toggle. According to the help page, in an open funnel users can enter at any step. In a closed funnel they must enter at the first step. This one choice changes what your report means.

QuestionClosed funnelOpen funnel
Who counts?Only people who enter at step onePeople who enter at any step
Best useMeasuring the whole path end to endSeeing the true volume of one step
RiskLeaves out people who land deep in the siteCan inflate step rates at the start

For example, if ads send visitors straight to a product page, a closed funnel that starts at the home page misses them. Therefore, in such flows, an open funnel gives a truer picture. Even so, it is useful to build both and compare the difference.

In short, do not read percentages until you know who the funnel counts. So write the type you chose in the report notes. Otherwise someone who opens the funnel three months later cannot explain why the numbers differ.

Which breakdown should you choose in GA4 funnel exploration?

In practice, a breakdown splits each step into separate data series by a dimension you pick. According to the help page, users are attributed to the first breakdown value that applies to them. That detail matters, because a user with several values counts in only one.

These breakdown dimensions help most when you narrow down a leak:

  • Device category: It separates desktop, mobile and tablet behavior. Google's dimensions and metrics reference defines this dimension.
  • Session default channel group: It shows which channel's visitors drop where. For the logic behind channels, see GA4 default channel group explained.
  • Landing page: It reveals which entry page loses more people.
  • Browser and operating system: They catch technical incompatibilities.

Also, the "rows per dimension" setting controls how many data series show in the table. Instead of opening many series, first focus on the top three to five values. Also avoid reading several dimensions at once. Go one at a time, such as device first, then channel, then landing page.

How do segment comparisons help you locate the leak?

Segment comparisons place different user groups side by side inside the same funnel. Google's help page says you can apply up to four segments. In short, it is the fastest way to see which group carries the loss.

Three comparisons are enough to start:

  • Mobile and desktop: Is the leak heavy on one device? For more on this, read mobile conversion gap: why mobile converts less.
  • New and returning users: Do new visitors need trust, or do returning visitors hit a blocker?
  • Paid and organic traffic: Do ad visitors drop at a different point?

You can also create a segment by right-clicking in the data table. For instance, each comparison helps you form a hypothesis. For example, if only mobile loses people at checkout, you look for the cause in the form layout or the payment method display.

On the other hand, too many segments carry a risk. As the number of segments grows, each group gets fewer users and the percentages swing. So compare the two or three most meaningful groups first.

How do elapsed time and next action explain the leak?

Once you find where the leak sits, two tools help you understand why. The first is the "Show elapsed time" toggle. It shows the average time between funnel steps.

If the time between two steps is much longer than you expect, users may hesitate or hit a blocker. Meanwhile, if it is very short, the flow may be simple or automatic. On its own it proves nothing, but it raises a question.

The second tool is the "Next action" card. For the dimension you choose, it lists the top five actions users take after each step. So you can see where dropped users went. You might notice, for example, that many people return to the product list after the cart.

However, pay attention to one caveat. Google's help page says the "(no next action)" entry does not always mean a drop-off for page or screen dimensions. Do not read that row as hard proof that people gave up.

Instead, watch session recordings of users who stopped. Recordings show where a visitor stalled or whether they switched tabs. Microsoft Clarity heatmaps and session recordings offer a free option for this.

How does a conversion rate calculator turn funnel counts into insight?

First, before you read the report, do the basic math. A step rate is the share of people who complete a step compared with the previous step. The overall rate is purchases divided by the users in the first step. To speed up the math, you can use a conversion rate calculator.

Example scenario (the numbers are invented, so replace them with your own): on an ecommerce site, ten thousand users view an item. Then two thousand add it to the cart. Of those, one thousand begin checkout, eight hundred enter their details and four hundred buy.

Let's calculate the step rates:

  • Item view to cart: twenty percent.
  • Cart to begin checkout: fifty percent.
  • Begin checkout to details entered: eighty percent.
  • Details entered to purchase: fifty percent.

First, the overall conversion rate is four percent. The biggest absolute loss sits between item view and cart, but that is a normal loss on most sites. So the suspicious point is the fifty percent loss between details and purchase, because those people were very close to buying. You can enter your own step counts and calculate your conversion rate in seconds to see the same pattern.

Is a leak a real problem or a natural loss?

However, not every loss is a problem. Most people who do not add a product to the cart were only browsing. The real question is whether the loss is bigger than you would expect. Three comparisons help you answer.

  • Your own history: How did the same step look in earlier months? A sudden break suggests a technical issue.
  • Segment gaps: If one group does much worse than another, the problem may live in that group.
  • Industry expectations: There is no official benchmark here. So rely on your own data instead of hunting for a number.
SymptomLikely causeWhat to check
Sudden drop on one deviceLayout or button issueDevice breakdown, test on a real device
Drop on one step across all channelsTechnical error or slownessError logs, page speed
Drop only on paid trafficMismatch between promise and pageAd copy and landing page
Very long time between stepsHesitation or a blockerElapsed time, next action card

For a broader diagnosis of falling conversions, our guide conversion rate dropped: where to start is a good companion.

What mistakes make you misread funnel data?

A funnel is powerful, but a wrong setup also gives misleading results. These mistakes show up most often.

  1. Missing or duplicate events: If a step's event never fires, or fires twice, your rates break.
  2. Wrong step order: Defining a sequence visitors never follow makes every user look lost.
  3. Forgetting deep landings in a closed funnel: A funnel that starts at the home page cannot count visitors who land on a product page from an ad.
  4. Consent and measurement limits: Visitors who decline cookies may be only partly measured. Do not mistake that for a leak.
  5. Domain changes: If checkout finishes on a different domain, sessions can break. Do not read that as a drop-off.
  6. Unit mismatch: A funnel counts users while a standard report counts events, so the numbers may not match.

The official page says the funnel reports only the first completed sequence per user. So knowing this helps you explain number gaps. To learn how to measure smaller steps, read what are micro-conversions in ecommerce and GA4 tracking.

How do you audit a funnel exploration setup?

So before you share a funnel, run a short quality check. A badly built funnel sends your team after problems that do not exist. Use this list:

  • Does the first step's count match the traffic you expect?
  • Is each step's count smaller than or equal to the previous step? If not, a condition may be wrong.
  • Does one action fire the same event twice? Run a test flow and count the events.
  • Are the date range, filters and segments right?
  • Does the form or checkout finish on another domain, and does that break the session?
  • Are step names written in your team's language?

If you cannot answer yes to these six questions, fix the setup first. Accurate measurement comes before accurate interpretation. What is GA4 and what are its benefits reminds you of the measurement model behind it.

How do you set up funnels for service and lead form sites?

Also, funnels are useful beyond online stores. On service, course and local business sites, the end goal is often a form submission, a phone tap or a message tap. Design the funnel around how visitors build trust on these sites.

Example scenario: a visitor sees a service page, moves to references and the contact page, then starts and sends the form. You define four steps: service page view, contact page view, form start and form submit.

Keep these points in mind:

  • If you track a form start event, treat the loss between start and submit as its own step.
  • Record phone and message taps as a second goal, because some visitors choose them instead of the form.
  • If you collect form error events, include them. They show which field blocks people.

For a deeper look at how form design affects conversion, read our guide to web form design.

How does an investigation with GA4 funnel exploration unfold?

Now let's put all the pieces together. Example scenario (the data is imaginary): you notice that the overall conversion rate of an ecommerce site slipped in recent months. You open a funnel exploration and build a five-step flow.

At first glance, the loss after begin checkout looks higher than before. So you ask these questions in order:

  1. Is the loss on all devices or just one? Open the device category breakdown.
  2. Is it on all channels? Switch to the channel group breakdown.
  3. When did it start? Check the trended view and narrow the date.
  4. What changed on the site that day? Review release notes, campaigns and payment settings.

For example, you see that the loss began only on mobile and after a certain date. Then it makes sense to look for the cause in the mobile checkout layout or in one payment method. At that point, you walk through the flow yourself on a real phone.

In practice, such an investigation can take a few hours. Still, it costs far less than trying fixes at random for weeks.

What should you test after you find the leak?

The funnel tells you where you lose people, not why. So your next step is to form a hypothesis. First, look at the lost step's page through the visitor's eyes. Then narrow the cause with session recordings and feedback.

Typical areas to fix include:

  • Form and checkout friction: Field count, error messages, payment options. For cart losses, see how to reduce cart abandonment.
  • Missing trust: Return terms, contact details, reviews.
  • Unclear offer: The visibility of price and delivery details.
  • Technical issues: Slow loading, a broken button, a mobile layout bug.

After you form a hypothesis, a controlled test is the safest way to prove a change works. Our guide to getting started with AB testing covers that process in detail. However, always confirm the leak is real before you test.

Also, rank your hypotheses by impact, ease and evidence. Apply small, fast fixes first, and gather more evidence before a big redesign. Also write the date of each change in a log, which helps you interpret the funnel later.

How do you save the funnel and monitor it regularly?

Instead of building the funnel once and forgetting it, create a steady review rhythm. According to the help page, you can save the exploration as a funnel report for faster access. Then open the same funnel each month and ask the same question: "Where did we lose the most people this month?"

A simple routine looks like this:

  • At the start of each month, open the same funnel and copy the step rates into a table.
  • Compare with the previous month and note clear changes.
  • After each campaign or site change, check the funnel again.
  • Every three months, confirm that the step conditions still capture the right events.

To tell campaign traffic apart by source, tag your links consistently. A UTM builder makes that easier. That way, the channel information in your breakdown stays clean.

How do you explain funnel results to your team?

Because data does not produce decisions unless people understand it, tell a story. When you share the report, lead with the story rather than the chart. First state the goal, then where the biggest loss sits, then the likely cause, and finally the test you suggest.

Try this order:

  1. Introduce the funnel and the period: which flow, which dates?
  2. Show the most suspicious step and explain why it looks suspicious.
  3. Add the segment gap: who carries the problem?
  4. State your hypothesis and the test you propose.
  5. Name the success metric and the monitoring plan.

Instead, do not present the numbers as hard proof. "The data shows this, and we suggest testing that" is both honest and persuasive. This way, non-technical team members also own the decision.

How does funnel exploration differ from other GA4 reports?

Put simply, standard reports give totals, while a funnel shows sequential behavior. The table below shows which tool fits which question.

QuestionBetter toolWhy
How many conversions this month?Standard reportsThe total is enough
At which step do visitors drop?Funnel explorationIt shows ordered steps
Which channel brings visitors?Channel reportsSource breakdown is ready
What are the first ten things to measure?First 10 GA4 metrics guideIt gives a starting list

Automatically collected events often give you a cheap start. Our article on GA4 automatically collected events for CRO explains how to use that data. So before you set up custom events, try building a first funnel with the events you already have.

When do you need web design support to fix a funnel leak?

First, finding the leak is a data job. Fixing it is often a design and development job. If the losing step is a form, a checkout page or a mobile layout, the fix touches the structure of the page. You can handle small copy and layout changes yourself, but structural problems take time.

Outside support makes sense in these cases:

  • The same step keeps losing people and you cannot see why.
  • Your page template makes changes hard.
  • Mobile and desktop experiences differ a lot.
  • Your tracking is incomplete and events are unreliable.

At this point, Talha Aslan and team turn funnel data into page decisions. However, we do not promise results or conversion lifts. Our approach goes from data to hypothesis and from hypothesis to a controlled change. If you want to work on this together, take a look at our web design services page.

What does a short action plan for GA4 funnel exploration look like?

Let's gather everything into one plan. If you are building your first funnel, this order is enough:

  1. Pick one main flow that brings in revenue.
  2. Write five to seven steps of that flow with event names.
  3. Verify with a real test that the events fire.
  4. Open the funnel exploration and define the steps and link types.
  5. Compare open and closed funnels and decide which one fits your question.
  6. Add the breakdown and segments with device category and channel groups.
  7. Narrow the cause of the leak with elapsed time and the next action card.
  8. Write your hypothesis, plan the test and monitor the funnel regularly.

The first time, this plan can take an hour or two. Then, in later months, each review drops to a few minutes. What matters is using the funnel as a habit, not once. This way you can show the leak instead of guessing at it.

Core guide: What Is Digital Marketing? Its types, examples and how it is done, on one page with an interactive film that travels through the brain.

Frequently Asked Questions

How many steps can a GA4 funnel exploration have?
According to Google's help page, a funnel can have up to ten steps. Each step needs a name and conditions based on events or dimension values, and you cannot use metrics as conditions. In practice, five to seven steps give most sites a readable funnel without hiding the real leak behind noise.
What is the difference between an open and a closed funnel?
In a closed funnel, users must enter at the first step to count. In an open funnel, users can enter at any step. For example, someone who lands directly on the cart counts in an open funnel but not in a closed one. Pick based on whether you study the full path or one step.
How do segment comparisons help in a funnel?
Segment comparisons put different user groups side by side in the same funnel. Google's help page says you can apply up to four segments. When you compare mobile with desktop, or new with returning users, you see quickly which group carries the leak, so you can aim your fix at the right audience.
Does "no next action" always mean a user dropped off?
Not always. Google's help page notes that for page or screen dimensions, the "(no next action)" entry does not necessarily mean a drop-off. The user may have left a tab open, the session may have timed out, or an event may not have fired. Confirm with session recordings and other signals before you decide.
Why do funnel numbers differ from my conversion counts?
A funnel counts users who complete the steps in order and reports only the first completed sequence per user. Standard conversion reports count events. Because the unit differs, the numbers can differ too. Before you compare them, check whether you are looking at users or events, and make sure the date range and filters match.
Is it enough to build the funnel once and save it?
Saving gets you started, but it is not enough alone. Your site, campaigns and tracking setup change over time, so step conditions can go stale. Open the funnel on a regular schedule, for example monthly, and confirm that each step still captures the right events. That protects you from false leak alarms.
  • ga4 funnel exploration
  • conversion funnel
  • ga4 explorations
  • conversion leak
  • ga4 events
  • segment comparison
  • cro
  • conversion optimization
Share:
Talha Aslan

Google Partner digital marketing expert. Hands-on with SEO, Google Ads, web design and e-commerce projects since 2012; every post here comes from that experience.

Next project

Get a quote for your project.

Your brief goes straight to Talha Aslan and team: strategy led by Talha, delivery by an experienced team. We reply the same day and send the scope, timeline and price as a written quote within 1 business day.