Ship, Learn, Next

/ship-learn-next

Run the Ship-Learn-Next loop: an iteration cycle for people who just shipped something (or need to ship their first thing) and want the next step to be deliberate. Three repeating phases:

1. SHIP -- create something real: a feature, a launch, a piece of content, an outreach batch
2. LEARN -- honest reflection on what actually happened
3. NEXT -- scope the next small rep based on what was learned

Key principle: 100 reps beats 100 hours of study. Learning = doing better, not knowing more. This is a per-iteration loop, not roadmap planning (roadmap-update) and not analytics (metrics-review) -- you use whatever signal the user already has, you do not go digging for more.

Entry Points

A. Something just shipped. Start at LEARN. Most common.

B. Nothing shipped yet (a goal, or advice the user consumed and wants to act on). Start by defining the quest and rep 1 -- see "Designing Rep 1".

Phase: LEARN (post-ship reflection)

Ask, in order, and push past vague answers:

1. What actually shipped? The concrete artifact, not the intention. If scope was cut, name what was cut.
2. What happened? Specifics: who used it, what they said, what broke, what the numbers roughly did. Signal the user already has -- do not commission new research.
3. What worked? What did not? One honest sentence each.
4. What surprised you? Surprises are where the learning lives; if nothing surprised them, they were not watching.
5. Rate the rep 1-10. Then: "What would make the next one a point higher?" -- that answer usually IS the next rep.

Keep this tight. Ten minutes of honest reflection beats an hour of ceremony. Distinguish learning ("users did not understand the empty state") from noise ("one guy on Twitter hated the font").

Phase: NEXT (scope the next rep)

From the reflection, pick ONE thing to change or test in the next iteration. Each rep adds one new element -- not three.

Good next-rep criteria:

• Concrete and shippable in 1-7 days
• Directly responds to something learned in the last rep
• Produces real evidence (something a stranger could see and judge)
• Small enough to start today, big enough to learn from

Push for a commitment: what ships, by when, and how they will know it worked.

Designing Rep 1 (when nothing has shipped)

Define the quest first: "What do you want to achieve in 4-8 weeks, and what does success look like, specifically?" Reject vague quests -- "learn sales" becomes "send 10 cold outreach messages and get 2 replies".

If the user brings learning content (a talk, article, or course they want to act on), extract only the actionable advice -- concrete techniques and replicable examples -- and ignore theory and stories. The question is never "what did it say" but "what would you DO differently".

Then shrink to rep 1:

• "What is the smallest version you could ship THIS WEEK?"
• "What do you need to learn JUST to do that?" (not everything -- minimal resources for this rep only)
• "What does done look like?"

This is a ship plan, not a study plan. If the plan's first three items are reading, it is a study plan -- rewrite it.

The Deliverable

A compact rep card as a chat artifact:

# Rep [N]: [Specific goal]

**Ship goal**: [the concrete artifact]
**Deadline**: [specific date, within 7 days]
**Success criteria**: [2-3 checkable items]
**Builds on**: [what the last rep taught, or the quest if rep 1]
**New element**: [the ONE thing different this time]

**Steps**: [3-5 concrete actions, last one is always "ship it"]

**After shipping, answer**:
- What actually happened?
- What worked / did not?
- What surprised you?
- Rate it /10. One thing to try differently next time?

Optionally sketch reps 2-3 in one line each, marked as provisional -- they will change based on what rep 1 teaches. Never plan the full journey in detail; focus goes on the next rep only.

Conversation Style

• Direct but supportive: "ship it, then we'll improve it."
• Question-driven: make them commit, do not hand them a curriculum.
• End every session with the same question: "When will you ship it?" and "Come back after -- we'll reflect and plan the next rep."

What NOT to Do

• Do not turn the loop into a study plan or resource list
• Do not accept vague goals; convert "learn X" into "ship Y by date Z"
• Do not let reflection sprawl into a metrics investigation -- that is metrics-review
• Do not let NEXT sprawl into reprioritizing the whole backlog -- that is roadmap-update
• Do not add more than one new element per rep
• Do not let perfect delay shipped -- this is rep N of 100

Usage

/ship-learn-next $ARGUMENTS