Count how many people reach each step in order, and the biggest drop tells you where to spend next quarter.
Counting how many people walk in, pick something up, reach the till, and pay. The biggest gap between two of those numbers is where next quarter goes.
An overall conversion rate is unactionable; the step with the biggest absolute loss is a work item.
A funnel measures ordered progress: viewed, added to cart, started checkout, paid. Its value is that it localises a problem — a 4% overall conversion rate tells you nothing actionable, while '68% of people who start checkout abandon at shipping cost' tells you exactly what to fix. The analysis is harder than it looks because three decisions change the answer completely: whether steps must occur in order, what time window a user has to complete the journey, and whether you count users or sessions. State all three, or two analysts will produce two different funnels from the same data and neither will know why.
A funnel counts users reaching each ordered step, so you can find where the largest drop is. Three choices determine the numbers: strict ordering versus any-order, the conversion window a user gets, and user-level versus session-level counting. The insight is never the overall rate — it is which step loses most people, and which segment loses them.
Deep dive into Funnel Analysis as a Product Analyst | Free Consultation for lucky subscribers — Elaine G, 3:48