Discovery does not end when the first key is chosen. It stays honest by parking every other key in the open. Hidden backlog is how the lock returns, dressed as "phase two."
Parking is an operating act
Teams often treat deferred work as a kindness to the original vision. The mobile app, the marketplace, the automation layer: all "coming next." If that list is not visible, owned, and tied to a reopen rule, it is not parked. It is a locked door waiting for budget. Discovery fails the moment parked keys can re-enter without answering the three questions again.
What visible parking looks like
A parked-item card
One card per deferred key, small enough to reopen without a mythology.
Core idea: Parking is successful when a new teammate can see what was refused, why, and what evidence would bring it back.
- Name of the deferred object in plain language
- Which question it failed: truth, better option, or not first
- Evidence that would reopen it
- Owner of the reopen decision
- Date it was parked, so it cannot quietly age into destiny
When to reopen
Reopen a parked key only when the live direction has changed in a way the original questions did not cover: a new lock, a first key that failed, or evidence that the better option disappeared. Do not reopen because a competitor shipped a similar surface, because a designer finished a mock, or because the calendar says it is time for a launch.
- The first key produced a repeated, accepted workflow and a new lock appeared
- The better option the team named is no longer available or no longer cheaper
- The truth claim was falsified and a different object now holds value
- Do not reopen to soothe a narrative that the venture still "needs the app"
Discovery as the keep-alive
The cadence matters. After the first engagement, Discovery can run again when a new high-stakes decision appears. That is how it stays a master key instead of a one-time diagnostic. The Keyteller Method and scoped projects exist after that diagnosis, not instead of it. If the parked list is ignored, the studio should name that as a risk, not quietly resume forging every key at once.
Publish the parked list with the first key
Same artifact, same day. Otherwise the locked door returns through side channels.
Review parked items only against new evidence
A review that restates the vision is not a review. Require a changed fact.
Let Discovery prescribe pause again
If the questions fail a second time, park again. The system is working.
What this series was for
Part 1 named the failure mode: a small MVP that still leaves the lock. Part 2 named the three questions that find the door. Part 3 named Discovery as the engagement that runs them. This part names the habit that keeps the master key honest. Together they put MVP Factory inside the Keyteller world: door, lock, first key, Discovery.