Lumen's team assumed their problem was traffic quality. It wasn't. Visitors were arriving fine and leaving before they understood what the product actually did the real leak was structural, not top-of-funnel.

The rebuild started by treating the homepage as a single decision tree instead of a list of sections. Every scroll had to answer one question before introducing the next: what is this, who is it for, why trust it, what does it cost, what do I do now. Anything that didn't serve that sequence — including sections the old site had leaned on for months — got cut, regardless of how nice it looked in isolation.

In Framer specifically, that meant building the pricing section as its own clean component early, testing it with real plan data immediately, rather than styling a static mockup and hoping the real numbers would fit later. The CTA moved up three sections. The credit-card ask moved back, past activation, not before it.

None of this involved new ad spend or new audiences. It was the same traffic hitting a site that finally got out of its own way.