Events
What Is a GA4 Funnel?
Updated 2026-09-043 min read
Short answer
A funnel is an ordered sequence of steps, showing how many people reach each one. The value is in the drop-off between steps: it points at the exact place people give up.
Counting events tells you how often things happen. A funnel tells you where people stop, which is usually the more actionable fact. If a thousand people start onboarding and four hundred finish it, the interesting question is not the four hundred — it is which step lost the other six.
How a funnel is built
You choose an ordered list of steps, each defined by an event. Analytics then counts the people who reached each step, and the drop between consecutive steps is the loss at that point.
app_open 1,000 users signup_start 620 users (−380) signup_complete 480 users (−140) first_workout 210 users (−270) Overall conversion: 21% Biggest single loss: app_open → signup_start
That last line is the whole point. The largest drop is at the very first step, which means the problem is not the signup form everyone was arguing about — it is that most people never start it.
Reading drop-off honestly
Two numbers describe the same gap, and they answer different questions. Conversion rate is completed divided by started. Drop-off rate is the remainder. Quoting one when you mean the other is a common way to make a step sound better than it is.
| Measure | Formula | For the example above |
|---|---|---|
| Conversion | completed ÷ started | 210 ÷ 1,000 = 21% |
| Drop-off | 1 − conversion | 79% |
| Users lost | started − completed | 790 |
What breaks a funnel
- A step whose event is not actually being sent — it will read as total loss at that point
- Steps in the wrong order, so people appear to skip backwards
- An event that fires more than once per user, inflating a step relative to its neighbours
- A cohort so small that a handful of people swings the percentages
Careful: A step showing near-total drop-off is more often a missing event than a catastrophic user experience. Verify in DebugView that the event fires before redesigning the screen.
Open and closed sequences
Funnels can be defined so that users must pass through every step in order, or so that they may enter partway and skip ahead. The two produce different numbers from identical data, which is worth knowing before comparing a funnel to one somebody else built.
Note: When a funnel result surprises you, check the definition before checking the product. Most surprising funnels are describing something slightly different from what you assumed.
How this looks in Appnaly
Appnaly reads the event data GA4 holds and surfaces the flows around your core features, so a step that collapses is visible without building a funnel by hand each time. Because the underlying counts come from GA4, the same definitional questions apply: an event that is not being sent is missing in Appnaly for exactly the same reason it is missing in Analytics.
For a quick calculation on two numbers you already have, the free funnel conversion calculator gives conversion, drop-off, and the number of users lost, with the formula shown.