# Conversion Audit

> Audit a page or form for friction and return a prioritised change list with expected lift. Use when auditing landing pages, signup forms, or checkout drop-offs.

- Skill: `poorvith-mp/conversion-audit` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add poorvith-mp/conversion-audit`
- Raw SKILL.md: https://api.skillmd.com/api/skills/poorvith-mp/conversion-audit/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: poorvith-mp (https://skillmd.com/u/poorvith-mp)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/poorvith-mp/conversion-audit

---


# Conversion Audit

You're diagnosing friction, not redesigning for taste. Every recommendation should trace back to a specific reason a visitor would hesitate or leave — not "this would look nicer," but "this creates doubt/confusion/effort at exactly the point where the visitor is deciding whether to continue."

## Workflow

1. **Identify the single primary action** the page/funnel exists to drive. If there are multiple competing CTAs, that's itself a finding worth flagging — a page trying to do too much usually converts worse on all of it.
2. **Walk the page/funnel in order**, from the visitor's first impression to the conversion action, checking for friction at each point:
   - **Clarity in the first few seconds** — can a new visitor tell what this is and why it matters to them without scrolling? Vague value propositions ("Empowering the future of X") are a common, high-impact fix opportunity.
   - **Trust signals at the moment of commitment** — near the CTA/payment step specifically, not just somewhere on the page. Social proof, guarantees, security badges matter most right where hesitation peaks.
   - **Friction in the action itself** — every additional form field, click, or decision point is a chance to lose someone. Ask whether each field/step is truly necessary for this specific conversion goal.
   - **Mismatched expectations** — does the traffic source's promise (ad copy, email subject line) match what the landing page actually delivers? A mismatch here creates a bounce that no on-page optimization fixes.
3. **For funnel drop-off data**, focus on the step with the steepest percentage drop, not the step with the most absolute traffic lost — the steepest drop usually indicates the sharpest friction point per visitor who reaches it.
4. **Propose specific, testable changes**, not vague directions. "Move the guarantee badge next to the price, not in the footer" beats "add more trust signals."
5. **Prioritize by expected impact vs. effort** — flag which changes are quick wins (copy tweaks, CTA button text) vs. bigger lifts (form restructuring, pricing page redesign).

## Anti-Patterns & Constraints

- Don't recommend dark patterns (fake urgency countdowns, hidden costs revealed late, forced continuity) even if they'd likely lift short-term conversion — they damage trust and often violate consumer protection regulations depending on jurisdiction, and that's a genuine tradeoff worth naming if the user asks for tactics in this territory.
- Don't give a generic checklist response — ground every recommendation in what's actually on the specific page/funnel being reviewed.
- Don't claim a specific percentage lift for any recommendation — that requires actual A/B testing; frame changes as hypotheses to test, not guaranteed outcomes.

## Output format

```markdown
## Conversion review: <page/funnel>

**Primary action:** [what this page/funnel should drive]

**Friction points found** (ranked by likely impact)
1. [Specific issue] — [why it creates hesitation] — [specific fix to test]
2. ...

**Quick wins:** [low-effort changes]
**Bigger tests:** [higher-effort changes worth a dedicated A/B test]
```

## Verification & Quality Checklist

- [ ] Success metric and its current baseline defined before launch, not after.
- [ ] Target segment named specifically enough to exclude someone.
- [ ] Channel-specific limits respected (character counts, aspect ratios, policy rules).
- [ ] Compliance checked for the channel (CAN-SPAM, GDPR, platform ad policy).