Pricing → A/B tests manages experiments and recorded results. Creating or starting a test does not currently split normal booking-widget visitors or apply different checkout prices automatically.
Quick answer
Pricing → A/B tests manages experiments and recorded results. Creating or starting a test does not currently split normal booking-widget visitors or apply different checkout prices automatically.
Open Pricing → A/B tests to create an experiment record, name its two strategies, choose the split, and manage draft, running, paused, or completed status.
The platform has strategy-assignment and conversion endpoints, but the normal booking widget does not currently call that experiment flow. Starting a test in the dashboard is therefore not evidence of a live customer pricing experiment. Do not expect views, bookings, or revenue to appear automatically from ordinary archive traffic.
Before using this for a real pricing decision, have the intended integration and data collection checked. The normal checkout also does not apply the separate dynamic-pricing preview engine.
Where events have been recorded, the detail view compares views, bookings, conversion, and revenue. Statistical or suggested-winner panels describe those recorded samples; they cannot prove that traffic was collected correctly or that other conditions were equal.
Keep enough data for a fair comparison, choose the outcome you care about before starting, and consider seasonality and sample size. A manually completed experiment or a highest-conversion badge is not proof of statistical significance.
Open the test's recommendation settings to choose a 90%, 95%, or 99% confidence level, the minimum total sample size across both strategies, and an optional maximum duration in days. Leave the duration blank for no deadline. Select Save to apply your changes; editing a control alone does not save it.
The analysis and scheduled checks use the same saved confidence and sample requirements. Enabling recommendations can flag a qualifying result for your review. You must explicitly approve a suggested winner to end the test; approval checks the current evidence again. Reaching the duration limit without a qualifying result prompts a review and does not invent a winner or automatically end the test.
Completing a test records its result. It does not rewrite your lesson prices. Use the live lesson editor and check the customer quote to apply a chosen price, respecting the terms of existing bookings.
See dynamic pricing [blocked] and seasonal planning [blocked] for the related tools and their limits.
Still need help with this article?
Send the article URL and the step that needs more explanation.