By Mehdi HassanWhat is conversion funnel analysis?
Conversion funnel analysis is measuring the share of visitors who move from one step to the next on the way to a conversion, such as landing page, pricing page, signup form and confirmed account. The output is a set of step-to-step rates, and the lowest one is where you look first. It turns "the site doesn't convert" into "40 people a day reach the form and most don't finish it".
It's the quantitative half of the research described in our conversion rate optimization guide. Funnels are good at showing where people leave and bad at explaining why, so they work best alongside recordings, surveys and sales conversations. Treat the funnel as the map and the qualitative research as the investigation.
Which steps belong in a website funnel?
A website funnel should include only the steps a visitor must pass through to convert, usually four to six. More steps make the chart look thorough and make it harder to read, because small optional pages split the traffic into fragments too small to judge.
Business type | Typical funnel steps |
|---|---|
Self-serve SaaS | Landing or homepage → pricing → signup form start → account created → activated |
Sales-led SaaS | Landing page → demo page → form start → form submit → meeting booked |
Waitlist or pre-launch | Landing page → form start → email confirmed |
Services or agency | Homepage or service page → work or pricing → contact page → form submit → call booked |
Content lead magnet | Article → lead magnet page → form submit → download |
Include a form-start step whenever you can. The gap between "reached the form" and "started the form" is a different problem from the gap between "started" and "submitted", and lumping them together hides which one you have.
How do you set up a GA4 funnel exploration?
You set up a GA4 funnel exploration in the Explore section by choosing the funnel template and defining each step as an event or page view. Before that, make sure every step fires a clean event, once per action. Our guide to landing page conversion tracking covers the event setup.
Open Explore and pick Funnel exploration
Start from the template rather than a blank exploration so the visualization and settings are ready.
Define each step
Add steps in order, each as a page view with a page path condition or as a named event such as form_start or sign_up.
Choose open or closed
A closed funnel only counts people who enter at step one. An open funnel lets people join at any step. Use closed for a landing page funnel and open for a site-wide one.
Add a breakdown
Set a breakdown dimension such as device category or session source so each step splits by segment.
Turn on elapsed time and next action
Elapsed time shows how long people take between steps. Next action shows what people who drop off do instead.
Set a long enough date range
Use enough weeks that each step has a meaningful number of users, especially on low-traffic sites.
Google's Analytics Help center has a page on funnel exploration that documents the current settings, which change from time to time. Check it if a menu doesn't match what you see.
What does each funnel drop-off pattern usually mean?
Each funnel drop-off location points to a different likely cause, which is why the pattern matters more than the overall rate. A big drop between landing and the next step is usually a message problem. A big drop inside the form is usually friction. Use the table below as a starting hypothesis, then confirm it with qualitative research.
Where the drop is | Likely cause | Check first |
|---|---|---|
Landing → any next step | Visitors don't understand the offer or it isn't for them | Headline clarity, ad-to-page match, traffic source |
Landing → pricing, but not beyond | Price surprise or unclear plans | Pricing page layout and plan differences |
Reached form → form start | Form looks long, or the ask feels too big | Field count, CTA wording, reassurance copy |
Form start → submit | Friction or errors inside the form | Validation errors, mobile keyboard, required fields |
Submit → booked or activated | Weak next step after the form | Thank-you page, email confirmation, scheduler |
Drop only on mobile | Layout or form broken on phones | Real-device test of the step |
When the biggest drop is on the very first step, also check who is arriving. A broad paid campaign can send thousands of poorly matched visitors, and no page will convert them. Our list of why landing pages don't convert covers the traffic-side causes.
How do you find signup funnel drop-off by segment?
Find signup funnel drop-off by segment by breaking the funnel down by device, source and landing page, then comparing each segment's step rates. The overall funnel is an average of very different visitors, and averages hide the problem: a form that works on desktop and fails on phones looks like a mediocre form overall.
- Device category catches mobile layout and form problems.
- Session source or medium shows whether one campaign sends visitors who never move past the first step.
- Landing page shows which entry pages feed the funnel well and which don't.
- New vs returning shows whether people need several visits before converting, which changes what the first visit should ask for.
Hypothetical example: reading a demo funnel by device
For example, suppose a sales-led SaaS site sees a weak overall rate from demo page to form submit. Broken down by device, desktop visitors move from form start to submit at a healthy rate, but mobile visitors mostly start and abandon.
Watching mobile recordings shows the phone field opens a full text keyboard and the error message for an unrecognised format appears above the fold, out of view. The fix is a correct input type, a forgiving phone format and inline errors. Without the device breakdown, the team would likely have rewritten the headline, which wasn't the problem. This is illustrative, not a real client result.
How do you turn funnel drop-off into fixes?
Turn funnel drop-off into fixes by pairing each leaking step with the qualitative evidence that explains it, then ranking the fixes. The funnel tells you the step; recordings, on-page polls and sales calls tell you the reason. A fix with both kinds of evidence behind it belongs at the top of the list.
For each problem step, watch a set of sessions that reached it and left, add a one-question poll on that page asking what's stopping them, and ask sales which objections come up most. Our conversion research methods guide covers each technique. Then score the resulting ideas with a framework, and decide whether each one needs a test or can be shipped and watched in the funnel.
How often should you review your conversion funnel?
Review your conversion funnel monthly, and after any change to the pages, the form or the traffic mix. A monthly look catches tracking breaks and slow decline before they cost a quarter's pipeline. A review after each change tells you whether the change helped, which matters most on low-traffic sites that can't run formal tests.
Save the exploration so the same steps and breakdowns are used every time, and annotate dates when you changed something. If you're moving platforms, the funnel is one of the first things that breaks: plan event continuity before launch, as our guide to analytics in a website migration explains.
