feedback-loop

name	feedback-loop
description	Creates iterative improvement cycles that feed deployment metrics and product feedback back into research for the next iteration. Use to close the loop from Zero-to-Revenue workflow and enable continuous product improvement.

Feedback Loop

Context

Use this skill to create iterative improvement cycles that take deployment metrics and product feedback and feed them back into research for the next iteration. This closes the loop from the Zero-to-Revenue workflow, enabling continuous product improvement.

Instructions

Phase 1: Collect Metrics

1. Fetch deployment metrics (response times, error rates, uptime)
2. Gather analytics (traffic, conversions, bounce rates)
3. Collect user feedback (support tickets, surveys, reviews)
4. Review content performance (engagement, shares, clicks)

Phase 2: Analyze Gaps

1. Identify underperforming features
2. Map metrics to JTBD statements
3. Find disenchantment moments (when users stop using)
4. Surface new jobs-to-be-done

Phase 3: Research New Iteration

1. Use jtbd-research to explore new market segments
2. Apply socratic-method to probe root causes
3. Use chain-of-thought to reason through solutions
4. Challenge with devils-advocate

Phase 4: Update Strategy

1. Update content pillars based on engagement data
2. Refine SEO keywords based on traffic
3. Adjust pricing based on conversion data
4. Plan next iteration with superpowers

Phase 5: Execute

1. Launch updated features
2. Deploy with deployment-manager
3. Monitor with audit-analyzer
4. Schedule next feedback loop

Constraints

• Never skip Phase 1 (collect metrics) — data drives decisions
• Allow 7-day minimum between loops for meaningful data
• Include at least 3 data sources per cycle

Examples

Example 1: Content Performance

Input: Blog post traffic dropped 40% Process: JTBD → "What job were users trying to do?" → New SEO keywords → Updated content Output: New content strategy for next quarter

Example 2: Churn Analysis

Input: Users canceling at Day 7 Process: Socratic probe → "What stopped working?" → Competitive gap → New feature Output: Feature roadmap for next sprint