Post Analyzer
Does: breaks down why a post worked or didn't and tells you exactly what to change. Solves: posting blind without knowing what's actually driving results. Triggers on: "analyze this post" · "why did this post flop"
How Maxim does this (diagnostic, not vibes)
Wraps the marketing CRO diagnostic discipline applied to social. Scores the post across the levers that actually move engagement — hook strength, scannability, specificity, emotional/identity trigger, CTA clarity — and returns the one or two changes with the highest expected lift, then frames them as a test via experiment-tracker.
Behavioral overlay
- Frameworks (cited per ADR-007): Fogg Behavior Model (B=MAP — was the action easy + motivated + prompted?) · AIDA · Hook Model · Peak-End (how it opened and closed).
- No-fabrication: if performance metrics are supplied, the analysis uses them and cites them; if not, it says so and analyzes structure only (never invents a "why it flopped" from missing data).
- Confidence tag (ADR-010) — 🟡 when metrics are absent.
Output
Diagnosis (what worked / what hurt, by lever) · the 1–2 highest-leverage changes · a testable rewrite + hypothesis.