design-first-user-looplisted

# Design First User Loop Turn a channel idea into a bounded experiment that can create learning without exceeding delivery capacity or making unapproved promises. ## Inputs Require: - a stated segment; - capacity, budget, and channel constraints; - external communication policy. Evidence, a formal metric contract, and a decision record are useful when they already exist, but none of the three is a precondition to draft this experiment — this batch is often how a founder gets the first evidence at all. State whichever of them exist by ID; for whichever do not, say so honestly instead of blocking or inventing one. ## Workflow 1. State the segment, trigger, and evidence behind the experiment — real evidence IDs if they exist, or the working hypothesis if this is the first experiment with this segment. Either is fine; an unstated segment is not. 2. Define the promised value and the first value moment. 3. Start in `direct_seeding`: choose one primary channel based on access and fit, not popularity. 4. Cap the learning batch at the first five users. State who counts toward the five, the support capacity per user, and the condition that stops further invitations. 5. Create a value-moment observation plan; do not turn the first five into a referral or incentive experiment. 6. Propose 1–2 candidate success metrics and one counter-metric, with a short reason each one would actually indicate the promised value was delivered — reference an existing `metrics.jsonl` reco