Distribute: Growth Experiments

This skill turns the channels a user has already chosen into a prioritized backlog of growth experiments — small, falsifiable bets that make a working channel work harder — plus a running tracker where results and learnings accumulate. It reads the Go-To-Market Strategy (the chosen channels and their pass thresholds), researches what's working right now in the niche, and walks the user through selecting, designing, and scheduling a tight cycle of experiments for one focus channel at a time.

The goal is not to produce a list of generic growth tips. It is to produce specific, falsifiable bets: each experiment names what changes, what's expected to move, the number that means it worked, and the simple steps to run it this week. "Make better videos" is noise. "Five hook variations leading with the customer's pain, one a day; a winner clears 50% three-second retention" is an experiment. The whole value of this skill is the distance between those two sentences.

This is the loop that runs after the first 30 days. productos/distribute/1-Go-To-Market-Strategy.md picks the channels and gets the user started; this skill is the repeatable engine for month two and beyond — hypothesis, run small, measure against a threshold, decide (double down / iterate / kill), and log. The most common ways it fails are chasing vanity metrics, running too many experiments at once, designing experiments with no threshold, writing them too generic to act on, and never logging results so nothing compounds. The workflow below is built to prevent all five.

Session length: Designed to be completable in 45–60 minutes. All tactic research is Claude's job during the session — finding what's working now in the niche and tearing down how competitors currently distribute. The user reacts and supplies the baseline number, the tools/analytics they have, and their weekly time budget. The skill plans one focus channel's cycle (one to three experiments) per run, not all channels at once.

Inputs

This skill needs channel context (which channels the user is running and what "working" means for each), product context (so experiments are specific), an output target, and the experiments playbook. Channel and product context come one of two ways depending on where the skill runs — the rest of the workflow is identical.

Channel context — Path A (standard): ProductOS in the repo — productos/ at the app repo root, or the current folder in a standalone ProductOS checkout

1. The Go-To-Market Strategy — usually productos/distribute/1-Go-To-Market-Strategy.md. The primary input: the ranked channels, their starter plans, and their pass thresholds. The focus channel defaults to the GTM doc's primary channel. If this file doesn't exist, the user hasn't picked channels yet — offer to run studio-distribute-gtm-strategy first, or establish the channels via the Path B questions.
2. The product summary — usually docs/PRODUCT.md (or the Define docs). Provides the customer, their language, the magic moment, and the north star, so experiment hypotheses are product-specific and metrics ladder to the right number.