/ab-setup

Language & Quality Standards

CRITICAL: Respond in the same language the user is using. If Vietnamese, respond in Vietnamese. If Spanish, respond in Spanish.

Standards: Token efficiency, sacrifice grammar for concision, list unresolved questions at end.

Skills: Activate ab-test-setup, marketing-psychology, analytics-attribution skills.

Components: Reference ./.claude/components/interactive-questions.md

---

Interactive Parameter Collection

Step 1: Ask Output Scope

Question: "What level of A/B test planning do you need?" Header: "Scope" MultiSelect: false

Options:

• Quick - Simple test plan
• Standard - Full test design (Recommended)
• Complete - Comprehensive with analysis plan
• Custom - I'll specify requirements

---

Step 2: Ask Test Type

Question: "What type of element are you testing?" Header: "Type" MultiSelect: false

Options:

• CTA/Button - Call-to-action copy or design
• Headline - Headlines, value props
• Layout - Page structure, flow
• Form - Fields, length, steps

---

Step 3: Ask Test Goal

Question: "What's your primary goal?" Header: "Goal" MultiSelect: false

Options:

• Conversion - Increase signups, purchases
• Engagement - CTR, time on page
• Revenue - ARPU, order value
• Retention - Return visits, churn

---

Step 4: Ask Traffic Level

Question: "What's your current traffic level?" Header: "Traffic" MultiSelect: false

Options:

• Low - Under 1K visitors/week
• Medium - 1K-10K visitors/week
• High - 10K-100K visitors/week
• Very High - 100K+ visitors/week

---

Step 5: Confirmation

Display summary:

## A/B Test Configuration

| Parameter | Value |
|-----------|-------|
| Test Element | [selected type] |
| Goal | [selected goal] |
| Traffic Level | [selected traffic] |
| Scope | [Quick/Standard/Complete] |

Question: "Create this A/B test plan?" Header: "Confirm" MultiSelect: false

Options:

• Yes, create plan - Generate test design
• No, change settings - Go back to modify

---

Workflow

1. Hypothesis Formation

   • Create if-then-because statement
   • Identify variable to test
   • Define expected outcome

2. Test Design

   • Define control (A) and variant (B)
   • Set primary and secondary metrics
   • Calculate sample size needed

3. Implementation Plan

   • Technical requirements
   • Tracking setup
   • QA checklist

4. Analysis Framework

   • Define success criteria
   • Plan for edge cases
   • Document next steps

---

Agent Delegation

Task	Agent	Trigger
Test design	conversion-optimizer	Primary task
Psychology review	brainstormer	Behavioral insights
Technical setup	researcher	Implementation guidance
Analytics setup	mcp-manager	Tracking configuration

---

Hypothesis Framework

Format: "If we [change], then [metric] will [direction] because [reason]"

Example: "If we change the CTA from 'Sign Up' to 'Start Free Trial', then signup rate will increase by 15% because it reduces perceived commitment."

---

Sample Size Guidelines

Baseline Rate	10% Lift	20% Lift	30% Lift
2%	19,000	4,800	2,200
5%	7,700	1,900	900
10%	3,900	1,000	450
20%	2,000	500	230

Per variation, 95% significance, 80% power

---

Output Format

Quick Scope

## A/B Test: [Element]

**Hypothesis**: [If-then-because statement]

| Element | Value |
|---------|-------|
| Control (A) | [Current version] |
| Variant (B) | [New version] |
| Metric | [Primary metric] |
| Duration | [Estimated days] |

Standard Scope

## A/B Test Plan: [Element]

**Hypothesis**: [Detailed if-then-because statement]

### Test Design
| Element | Value |
|---------|-------|
| Control (A) | [Description] |
| Variant (B) | [Description] |
| Primary Metric | [Metric name] |
| Secondary Metrics | [List] |
| Traffic Split | 50/50 |
| Sample Size | [Per variation] |
| Duration | [Estimated days] |

### Success Criteria
- Statistical significance: 95%
- Minimum detectable effect: [X%]

### Implementation Checklist
- [ ] Tracking setup
- [ ] QA verification
- [ ] Launch approval

Complete Scope

[Include Standard + Full hypothesis analysis + Detailed sample size calculation + Segmentation plan + Analysis framework + Win/loss playbooks]

---

What to Test (Priority)

High Impact

• Headlines and value propositions
• CTA copy and placement
• Form length and fields
• Pricing presentation

Medium Impact

• Image choices
• Button design
• Page layout
• Copy length

Low Priority

• Minor copy tweaks
• Font changes
• Icon changes

---

Output Location

Save test plan to: ./docs/tests/ab-test-[element]-[YYYY-MM-DD].md