Funnel analysis follows users through the ordered steps towards a goal and measures how many continue at each step.
Each step's conversion rate shows where users drop out.
For example, a shop's funnel might run from product page to basket to checkout to payment, losing users at each step.
The biggest drop is usually the best place to look for improvements.
Funnels should count each user once per step, in order, within a set time window, or the rates mislead.
Funnel analysis turns a journey towards a goal into a series of measurable steps.
It exists because overall conversion hides where users actually give up.
Knowing the leaky step focuses work where it can make the biggest difference.
A funnel is defined by an ordered list of events, such as visit, sign-up and first purchase.
Step conversion is the share of users at one step who reach the next.
Overall conversion is the share of users at the first step who reach the last.
Overall conversion equals all the step conversions multiplied together.
A strict funnel counts only users who did the steps in order, while a loose one counts anyone who did each step.
A time window, such as 7 days, stops a purchase months later counting as part of the same journey.
Splitting a funnel by device, channel or country often shows that one group drops out far more.
A drop at a step can mean the step is hard, or that the earlier step attracted the wrong users.
Improving an early step can lower later rates, because the extra users arriving are less committed.
Funnels show where users leave, not why, so interviews or session recordings fill that gap.
In practice, funnels guide where to run A/B tests, which then show whether a fix works.
The usual mistake is counting events rather than users, which lets one user's repeats inflate a step.
One question on this card. Every option explains itself once you pick, right or wrong.
12 concepts in this module, then a graded test with every answer explained.