By running the request-to-shipped workflow on itself: demand signals from customers, the waitlist, and Heidi's own usage data land in the brain, get connected to context, and the loop closes publicly when things ship.
Dogfooding the prioritization process is the point: the workflow Heidi sells to customers, request in, context attached, decision made, requester told, is the one that runs Falkster's own roadmap.
Your questions are literally input: what people ask on the waitlist form becomes answer pages, and what they ask for becomes signal in the queue. Ask for what is missing; it is read by the person who builds.