ab-test-runnerlisted

# A/B Test Runner Stop guessing. Test everything systematically — one variable at a time. ## When to Activate - User says `/ab-test` or `/ab-test {element}` - User asks "should I test this?" - User asks "which approach works better?" - User wants to optimize a specific content element ## Prerequisites - `~/.content-autopilot/content-history.json` with performance data (from /perf-log) - `~/.content-autopilot/profile.json` ## Data: ab-tests.json Location: `~/.content-autopilot/ab-tests.json` ```json { "version": "1.0", "tests": [ { "id": "test-001", "element": "hook_type", "hypothesis": "Data hooks outperform question hooks on X", "variant_a": { "type": "data_hook", "content_id": "2026-03-20-001" }, "variant_b": { "type": "question_hook", "content_id": "2026-03-22-001" }, "control_variables": ["topic", "posting_time", "format", "length"], "metric": "engagement_rate", "status": "running", "started_at": "2026-03-20", "results": null, "winner": null, "confidence": null } ] } ``` ## Testable Elements | Element | How to Test | Minimum Tests | |---------|------------|---------------| | Hook type | Same topic, different opening | 3 pairs | | Title logic | Same content, different title | 3 pairs | | Posting time | Same content type, different time | 5 tests | | CTA type | Same content, different CTA | 3 pairs | | Thread length | Same topic, different tweet count | 3 pairs | | Content depth