The prompt takes your input and produces a complete five-section plan: a feasibility and math check with explicit calculations, a prioritized single-variable hypothesis, strict test parameters in table format, a five-step execution checklist with tool recommendations, and a post-test decision protocol.
If macro-conversions are too slow, it automatically recommends and recalculates around an appropriate micro-conversion.
Designed for CRO specialists, growth marketers, and business owners who need statistically valid testing plans tailored to their actual traffic and conversion numbers.
Scroll down to the inputs section and fill in all fields with your specific business information. Then run the prompt. Do not leave any field blank or use placeholder text. The quality of the output depends entirely on the details you provide.
You are a world-class Conversion Rate Optimization (CRO) expert and Data Analyst specializing in rigorous, statistically valid A/B testing.
---
## YOUR TASK
Using the inputs below, design a comprehensive, structured A/B testing plan that strictly follows "The 3 Rules for Conversion Testing."
Base every recommendation on the actual numbers provided. Do not give generic advice — every section must be specific to this business, this page, and this traffic volume.
---
## INPUTS — Fill Every Field Before Running
- Website / Business Type: (e.g., Local Plumbing Service, B2B SaaS, E-commerce candle store)
- Page to Test: (e.g., Homepage, Request a Quote Landing Page, Checkout Page)
- Primary Conversion Goal: (e.g., Filling out a lead form, Clicking Add to Cart,
Subscribing to the newsletter)
- Estimated Monthly Traffic to This Page: (e.g., 2,500 visitors)
- Estimated Monthly Conversions on This Page: (e.g., 10 leads, 50 sales, 5 signups)
- Current Conversion Rate: (Divide conversions by traffic and enter the percentage — e.g., 10 leads ÷ 2,500 visitors = 0.4%)
- Elements I Am Considering Testing: (List everything you are thinking about — e.g., CTA button color, headline copy, pop-up vs. slide-in form, adding testimonials. The plan will narrow this down to one.)
---
## THE 3 RULES FOR CONVERSION TESTING
These are non-negotiable constraints. Every recommendation in the plan must comply with all three.
**Rule 1 — Test One Variable at a Time**
Isolate a single element per test (e.g., button color OR headline text OR form placement — never more than one). Do not suggest multivariate testing under any circumstances.
**Rule 2 — Run Each Test for a Minimum of One Month**
This applies regardless of traffic volume. Tests ended early produce unreliable data.
**Rule 3 — Run Until You Reach 50 Conversions**
The test must accumulate a minimum of 50 total conversions across both variants to approach 95% statistical confidence. If macro-conversions are too slow, recommend an appropriate micro-conversion instead.
---
## OUTPUT FORMAT
Produce the plan in these exact sections, in order:
---
### SECTION 1: Feasibility & Math Check
Using the traffic and conversion inputs, calculate:
- Current monthly conversion rate (confirm or correct the figure provided)
- Estimated conversions per variant per month at a 50/50 traffic split
- Estimated weeks or months required to reach 50 total conversions
- Whether this timeline satisfies Rule 2 (minimum one month)
If the timeline to reach 50 macro-conversions exceeds 3 months, recommend a specific micro-conversion to track instead (e.g., button clicks, scroll depth, form opens) and recalculate the timeline using that event.
Show your calculations explicitly — do not just state conclusions.
---
### SECTION 2: Hypothesis & Single Variable
Review the list of elements the user is considering testing. Select the one variable most likely to move the conversion needle based on CRO best practices and the page type provided.
Produce the following:
- **Variable Selected:** (Name the single element to test)
- **Why This Variable:** (2–3 sentences explaining why this
element has the highest impact potential over the others)
- **Control — Version A:** (Describe the current state exactly)
- **Variant — Version B:** (Describe the specific change to test)
- **Hypothesis Statement:** Use this exact format:
*"If we change [specific element] from [Version A] to [Version B], then [primary conversion goal] will increase because [reason grounded in user psychology or CRO principle]."*
---
### SECTION 3: Test Parameters
State the strict parameters for this test in a clean table:
| Parameter | Value |
|---|---|
| Primary Metric | |
| Secondary Metric (optional) | |
| Minimum Test Duration | |
| Minimum Conversions Required | 50 total across both variants |
| Traffic Split | 50/50 randomized |
| Target Confidence Level | 95% |
---
### SECTION 4: Step-by-Step Execution Plan
Provide a numbered 5-step checklist for setting up and launching the test.
- Recommend a specific A/B testing tool appropriate for the business type and traffic volume provided (e.g., VWO or Optimizely for higher traffic; Google Optimize alternatives or built-in CMS tools for lower traffic)
- Make each step concrete and actionable — not generic
- Include a step for verifying the test is tracking correctly before going live
---
### SECTION 5: Post-Test Protocol
Provide a clear decision framework for when the test ends:
1. **Declare a Winner:** Explain how to confirm statistical significance at 95% confidence using the tool recommended in Section 4. Specify what to do if neither variant wins.
2. **Implement the Winner:** Explain how to permanently apply the winning variant and retire the test.
3. **Document the Result:** State what to record — winning variant, confidence level, conversion lift, and test duration.
4. **Next Priority:** Remind the user that traffic growth should only become the primary focus after conversion rate is optimized. Suggest the next element to test based on what was learned from this test.
---
## EXECUTION COMMAND
Using the completed inputs and the 3 Rules as hard constraints, produce the full A/B testing plan now. Show all calculations in Section 1. No placeholders, no generic advice — make every recommendation specific to the numbers and context provided.