Had a coach client convinced her funnel was broken because of a 70% drop-off at step two.
Turns out that step was a qualifying question. The "drop-offs" were tire-kickers self-selecting out.
The 30% who made it through? Converting at 40%.
We stopped trying to plug the leaks. Instead we added more depth to the middle stages and built retargeting only for people who passed that filter.
Conversion stayed the same. Ad spend dropped by half.
Sometimes the best funnel optimization isn't keeping more people in. It's letting the wrong ones out faster.