Experiments
| Experiment | Target Page | Status | Actions |
|---|---|---|---|
| No experiments launched yet. Click "Create Experiment" to start. | |||
💻 External Website JS Embed Snippet
Paste this snippet into the <head>of any external website (e.g., Happy Diwali landing page) to enable 50/50 split testing & conversion tracking:
<script src="https://kreate-locations-b5bjgef6eyakfahc.canadacentral-01.azurewebsites.net/api/snippet?id=test1"></script>
How to tag your external HTML elements:
- Variant A Element:
<div data-ab-variant="A">Original Headline</div> - Variant B Element:
<div data-ab-variant="B">Diwali Offer Headline</div> - Track Button Click:
<button data-ab-convert="true" data-ab-value="149">Buy Now</button>
Choose outcome
What should this test improve?
Configure details
Internal hosted page or external website?
Create variants
Link your control and challenger variants.
Launch & monitor
Review and launch your experiment.
What would you like to improve?
Select an experiment from the dropdown above to view saved variants and preview options.
Live Variant Previews & Source HTML
Variant A HTML preview will appear here when an experiment is selected.
Variant B HTML preview will appear here when an experiment is selected.
Lead Submitted
Determines the winning variant.
CTA Click
Explains user intent before conversion.
Bounce Rate
Prevents a winner that hurts page quality.
🎯 Conversion Tracking Implementation
Active for Experiment: All Experiments
Add data-ab-convert="true" to any button or link. The embed script automatically detects clicks and logs the conversion.
<button
data-ab-convert="true"
data-ab-value="1">
Submit Lead
</button>Call window.kreateAB.track() inside your React/Vue/vanilla event handlers upon form submission or checkout:
if (window.kreateAB) {
window.kreateAB.track('lead_submitted', {
value: 1 // conversion value
});
}Allocation
Keep users sticky with a cookie so the same visitor always sees the same version.
Audience rules
A vs B comparison
| Variant | Visitors | Conversions | Conversion Rate | Avg Order Value | Result |
|---|---|---|---|---|---|
| Variant A | ... | ... | ... | ... | Control |
| Variant B | ... | ... | ... | ... | Challenger |
Recent Events (Raw Data)
| User ID | Variant | Landing Page | Time on Page | Status | Revenue |
|---|---|---|---|---|---|
| Loading data... | |||||
Baseline Conversion Rate
Sample Size
AB Test Assumption
Chi Squared Test
Two Sample T Test
Paired T Test
Mann Whitney U Test
Wilcoxon Signed Rank Test
Why Variant B is likely winning
Variant B uses a more concrete promise and a clearer action-oriented CTA.
- “5 minutes” creates a measurable speed promise.
- “Create My Website” matches visitor intent better than “Get Started”.
- SEO-ready copy adds a business outcome beyond design.
Auto-stop recommendation
Winner detected
Variant B has enough evidence to roll out. Suggested action: deploy B to 100% traffic and archive A.
Suggested next experiments
Shorter Form Test
Reduce fields to increase lead submissions.
Trust Badge Test
Add testimonials and “500+ sites generated”.
Industry Landing Test
Show plumber, dentist, consultant examples above fold.
Lead Generation Pack
- Headline test
- CTA test
- Short form test
Trial Signup Pack
- Free trial CTA
- Pricing section
- Demo placement
Local Business Pack
- Location headline
- Reviews block
- Call button
Tracking setup
Safety rules
Experiment
Name the test and state the hypothesis.
Signals
Choose a control and at least one candidate.
Data & validation
Define an auditable, point-in-time window.
Review
Confirm the design before running.
Experiment definition
🧪 Create New Prompt A/B Test
Define two prompt templates with variables like {{question}} or {{topic}} to evaluate LLM output quality.