A conversion funnel is the series of steps a visitor goes through before doing the thing you want, such as buying, booking or sending an enquiry. It is called a funnel because many people start at the top and fewer make it through each step. Looking at it step by step tells you exactly where people give up.
How a funnel is measured
You list the steps in order and count how many people reach each one. For an online shop that might be: viewed a product, added to basket, started checkout, entered payment, completed order. Each step has its own drop-off. Overall conversion rate only tells you how many finished; the funnel tells you which step lost the rest.
In Google Analytics (GA4), the Funnel exploration report lets you define these steps from page views or events and shows completion and abandonment for each. Tools such as PostHog and Mixpanel offer similar funnel charts. For a simple brochure site, the funnel can be as short as: landed on a service page, opened the contact page, submitted the form.
Marketers also use "funnel" for the broader journey: awareness, consideration, decision. That version is a planning model for content and ads rather than something you measure precisely.
Example: finding the leak
A bakery takes cake orders online. Plenty of people reach the order form, but few submit it. Breaking the form into steps in analytics shows most leave at the delivery date field. The calendar only offers dates three weeks out with no explanation. Adding a line saying "custom cakes need 3 weeks; for sooner, call us" and a phone link turns a silent dead end into calls. Without the funnel view, the bakery might have spent money on more traffic instead.
When it is worth setting up
If your site has a multi-step path, such as checkout, booking or a quote builder, a funnel is the quickest way to find what to fix. If the only goal is a single contact form, you mostly need two numbers: visits to key pages and form submissions. Funnels also need enough traffic to mean anything. With twenty visitors a week, one person changes the percentages wildly, so watch session recordings or ask customers instead.
Funnel analysis mistakes
- Too many steps. Tracking every scroll and click produces a chart nobody can act on. Pick the moments that represent real commitment.
- Broken tracking. If an event stops firing after a site update, the funnel shows a fake collapse. Check events after every change.
- Optimising the top only. More visitors into a leaky funnel mostly produces more abandonment.
- Ignoring device splits. A step that works on desktop can be broken on phones; compare the two.
Related terms
- Conversion rate: the end-to-end figure a funnel breaks apart.
- Google Analytics: one place to build and view funnel reports.
- Retargeting: ads aimed at people who dropped out partway.
- Lead generation: the top of the funnel for service businesses.