Home / Use Cases / A/B Test Landing Pages with AI
Use Case

Stop Guessing. Start Testing with Statistical Confidence.

Axerto's built-in A/B testing lets you create AI-generated variants, split traffic with configurable weights, and identify winners with 90%, 95%, or 99% statistical confidence — all without extra tools.

Most A/B testing tools require separate platforms, complex integrations, and statistical expertise. Axerto includes A/B testing natively — create a variant, split traffic, and see results with automatic significance scoring built in.

Step-by-Step Guide

  1. 1
    Create a Variant

    From any landing page, click 'Create Variant' to generate an AI-powered alternative version. The variant inherits your base page and can be edited independently.

  2. 2
    Set Traffic Split

    Configure traffic weight for each variant (e.g., 50/50, 70/30). Visitors are assigned to variants on first visit and tracked with a 30-day sticky cookie.

  3. 3
    Monitor Performance

    View real-time stats including view counts, lead counts, conversion rate, and statistical significance at 90%, 95%, and 99% confidence levels.

  4. 4
    Promote the Winner

    When a variant reaches statistical significance, click 'Promote' to make it your main page. All other variants are deactivated automatically.

Key Benefits

Statistical Significance Scoring

Z-test calculations show 90%, 95%, 99% confidence — know when a result is real, not random.

Weighted Traffic Splits

Configure any traffic ratio (50/50, 70/30, etc.) with sticky cookie tracking for consistent user experience.

AI-Generated Variants

Generate variant pages with different AI prompts — test entirely different approaches, not just button colors.

One-Click Winner Promotion

Promote the winning variant to main page with one click. All other variants deactivated automatically.

Frequently Asked Questions

How does Axerto's A/B testing work technically?
When a visitor hits your landing page URL, Axerto randomly assigns them to a variant based on configured traffic weights using weighted random selection. The assignment is stored in a 30-day cookie (ab_{parent_pk}) to ensure the same visitor always sees the same variant. You can bypass the split with ?v= to directly access a specific variant for testing.
What statistical method does Axerto use for significance?
Axerto uses a two-proportion z-test to calculate statistical significance. Results show significance at three confidence levels: 90%, 95%, and 99%. We recommend waiting for 95% confidence before making decisions, and 99% for high-traffic pages where false positives are costly.
Can I A/B test entirely different page designs?
Yes. Each variant is a full, independent landing page that can be edited separately in the visual editor or regenerated with a different AI prompt. This lets you test fundamentally different offers, layouts, and messaging — not just minor element changes.

Ready to Try Axerto?

7-day free trial — no credit card required. Full feature access from day one.

Start Free Trial