EPISODE · Jun 17, 2026 · 10 MIN
Why Your A-B Test Needs a Segmentation First
from Conversion Rate Optimization with Fexingo: A/B Testing, Landing Pages, and CRO Strategy · host Fexingo
Episode 58 of Conversion Rate Optimization with Fexingo: Lucas and Luna unpack why running an A-B test without pre-slicing your audience by segment can hide winning and losing variants. Using a real-world example from a mid-market SaaS company, they walk through how a test that showed 'no winner' actually uncovered a 23 percent lift for mobile users and a 9 percent drop for desktop users — completely cancelling each other out. They discuss Simpson's Paradox, the importance of logging device type and referrer source before the test starts, and why post-hoc segmentation is a recipe for false conclusions. A practical, data-focused conversation for anyone running experiments. #ABTesting #CRO #ConversionRateOptimization #Segmentation #SimpsonsParadox #MarketingAnalytics #DataDriven #ExperimentDesign #SaaS #UserBehavior #MobileVsDesktop #Statistics #FalseNegative #HypothesisTesting #BusinessPodcast #FexingoBusiness #ConversionRate #DataScience Keep every episode free: buymeacoffee.com/fexingo
Embed this episode
NOW PLAYING
Why Your A-B Test Needs a Segmentation First
No transcript for this episode yet
Similar Episodes
No similar episodes found.
Similar Podcasts
No similar podcasts found.