Learning marketing from zero — by running real, pre-registered experiments in public.
This site is a solo project with two parts: a case-study-first curriculum for absolute beginners (currently in Korean), and a lab where every marketing experiment is pre-registered — success criteria locked in a database before the results arrive — then judged honestly, inconclusive verdicts included.
Why pre-registration matters
Marketing content is full of success stories justified after the fact. Here, the hypothesis, success criteria, and measurement date are locked by a database trigger once an experiment starts — amendments require a written reason and are appended, never overwritten. Numbers you cannot quietly edit are the only numbers worth trusting.
Hypothesis and success criteria are written before execution, then locked.
On the pre-committed date, against the pre-committed baseline.
Win, loss, inconclusive, or invalid — every verdict ends in a resource decision, and failures stay published.
Latest issues
All issues- 2026-09-23Shopify Redefined How It Counts a "Session," Starting September 21
- 2026-09-22Google confirmed on September 14 it's testing paying sites for AI answers
- 2026-09-21Meta launches Creator Marketing Hub: content usage rights now expire
- 2026-09-15Adobe acquired Rilo. The product died, but its Slack and Salesforce access didn't.
- 2026-09-14Common Crawl opened 580,000 llms.txt files. Two-thirds were empty templates.