We increased conversion ~30% and we don't know exactly how

The point isn’t to know, it’s to do:

We spent six weeks on the work, we did our best, and something worked out really, really well...

You could make the argument that we should have tried each thing separately, measured each impact, and then decided where to go next (or known when to stop). You could make the argument that changing so many things at once makes it impossible to know which variables actually mattered. You could have argued we should have been more rigorous in our evaluation so we could learn something fundamental we could apply to a future project.

You could argue all those things. While you were debating those points, spending months teasing out answers, or testing each change in high-traffic succession for statistical significance, we were already basking in the results and moving our product teams to the next project. In six weeks all the work was done, we did our best, and it worked.

The point wasn't to know, it was to do.