What is funnel analysis?
Funnel analysis is a method of measuring how many users complete each step in a sequence that leads to a goal, such as sign-up, setup or purchase, and where they drop off. It shows which step loses the most people, so teams know where to focus.
How do you build a funnel analysis?
You build a funnel analysis by choosing a goal, listing the steps users take to reach it as tracked events, and counting how many users complete each step in order within a set time. The output is a step-by-step count with the conversion between each pair of steps.
The decisions that matter:
- Start from the goal. "Paid", "set up and used weekly" or "first report shared". One funnel, one goal.
- Keep the steps few and real. Four to six events that every successful user passes through. Too many steps and every drop looks small.
- Choose ordered or any order. Most product funnels require the steps in sequence. If users can do them in any order, measure completion of each instead.
- Set a conversion window. How long a user has to finish, such as seven days from sign-up. Without one, slow users and abandoned users look the same.
- Segment. Split by plan, channel, device or company size. A funnel that looks fine overall often hides one segment that falls apart.
How do you find the drop-off step?
Look at the conversion from each step to the next, not only at the overall total, and find the step with the biggest relative loss.
Take a project-management tool where a new admin has to create a project, invite the team and get a teammate to act. Say 1,000 people sign up in a week:
| Step | Users | Of previous step |
|---|---|---|
| Signed up | 1,000 | |
| Created a project | 700 | 70% |
| Invited a teammate | 300 | 43% |
| Teammate completed a task | 240 | 80% |
The overall result (240 of 1,000, or 24%) says something is wrong. The step view says where: inviting a teammate loses 57% of the people who got that far. That's the step to study first.
Before acting, check two things. Does the drop hold across segments, or is it one channel bringing in solo users who have no one to invite? And is the step wanted at all, or is the product asking for it too early?
What can't funnel analysis tell you?
A funnel shows where users leave. It doesn't show why. For that, pair it with session recordings of people at the failing step, short interviews and the questions users ask at that moment.
Often the answer is plain: users knew what they wanted (add the team) but not how, or the step needed information they didn't have to hand. Foldspace includes product analytics (funnels, retention and session replay), and its Product Agent lets users at the failing step ask for what they need and does it with them in the product, so you can watch whether the drop at that step shrinks.