Reworking the onboarding funnel

Before changing anything, I'd split that 40% by where it happens — abandoning the form, never verifying email, or landing in an empty product and leaving.

Those look identical in the aggregate number but have completely different fixes. Form abandonment is usually field count or an unexpected required field. Verification drop-off is deliverability. An empty first session is almost always missing a default path — the user arrives and there's nothing obvious to do.

Do you have the funnel broken out by those three steps?

Then the empty state is doing the work of your onboarding, and it probably isn't built for that.

The highest-leverage change is usually to remove the blank canvas entirely: seed a sample project the user can poke at and delete, so the first action is editing something rather than creating something from nothing. That converts a hard question — what should I build? — into an easy one.

If you want to test it cheaply, ship it to half of new signups and measure time-to-first-edit rather than time-to-first-project. It moves earlier and gives you a signal within days instead of weeks.