The name comes from the shape: many people enter at the top, fewer survive each step, a small number arrive at the bottom. A typical version runs awareness, interest, consideration, purchase — but the useful funnel is not the textbook one, it is the specific set of steps your actual customers take. Saw a post, visited the site, read pricing, started a trial, invited a colleague, paid. Six steps you can count.
Its value to a founder is that it localises problems. "Sales are slow" is not fixable; "78% of trials never complete setup" is. Mapping the funnel forces you to instrument each transition, and the biggest percentage drop is almost always where your next week's work belongs. It also stops you over-investing at the top — pouring more people into a funnel that leaks at step four just makes a more expensive leak.
A concrete example: your consultancy funnel is 1,200 site visitors → 90 guide downloads → 24 discovery calls booked → 14 calls attended → 4 proposals sent → 2 clients signed. The worst drop is 24 booked to 14 attended: 42% of the people who actively asked to speak with you never show up. No amount of extra traffic fixes that. A reminder email and a calendar hold might, and it is a two-hour job that could plausibly add a client a month.