Creator Office — Release & Launch Plan

Set the rhythm

Engineering is not the constraint — every package fits inside either cadence. What changes is how much room each launch gets for proving and comms, and how often Nate and Sydney are on a call.

Launch timeline

Every package runs the same five-phase cycle. The grey band is build, which runs continuously underneath and is not scheduled by this process. Hover any segment for its dates.

The packages

If a package cannot be described in one sentence a creator understands, it is a list, not a package.

The release SOP

This is not the build process. Build is continuous and flag-gated; the SOP is what turns finished code into a launched package. Each phase has one exit gate — if the gate isn't met, the launch moves, not the standard.

Go / no-go — five narrow vetoes

Phase dates by package

Computed from the launch date and cadence above. Prove is the first thing a compressed cadence eats.

Open decisions

Dates are computed from the launch each one gates. Three block the most expensive packages, and none are engineering problems.

Also outstanding from the meeting

• Staging access to attendees — this is now Prove's entry requirement, not a nice-to-have
• Meeting notes + proposed agenda circulated
• Postmark user-activity recordings access for Nate
• Active user / engaged user / conversion proposals — blocks every package's success measure
• Review Unwrapped (Corey's Gmail-native pitching platform) before P3's Shape phase — it overlaps P2 and P3 combined
• AMA / office hours — attaches to Land from P2 onward, once there's enough live product for an AMA to have substance

Activation, engagement and conversion — by stage

A single definition can't span the roadmap: at P1 you can only measure what P1 ships. So each stage gets its own definition, built from the features that actually exist by then, and each one tightens as more lands. Proposals for Matteo and the team — none of these are agreed yet.

The five rules these follow

• Every definition is a write action, never a page view. Logging in is not activity. If a creator can hit the number without changing anything in their own business, the number is measuring us, not them.
• Definitions are versioned, not revised. A P1 active user and a P4 active user are different things — never compare them on one chart. Stamp each number with the definition it was measured under.
• One denominator, chosen once. Recommend students granted access, not students who logged in. It's a harsher number and the honest one — someone who never logged in is a failure we own.
• Thresholds are low and honest, not aspirational. The team agreed light usage is enough to indicate value for some segments. Set the bar where value actually starts, then watch it — don't set it where we'd like adoption to be.
• No targets until four weeks of data exist at that definition. Week-one retention at 18–20% is a starting signal, not a baseline. A target set before the data matures becomes a justification instead of a test.