A/B Testing Explained With Real Numbers: Conversion Rate, P-Value, Sample Size

A/B testing explained with real numbers: control vs variant, random assignment, conversion rate, why small samples lie, statistical significance, p-value intuition, sample size, and the common mistakes (peeking and too many variants). We test a blue Buy button against a green one, step by step. All numbers in this video are made up for teaching, not real company data. If this video helped you, please like, share and subscribe to CodeFlow Daily. Key ideas: A = control (what you have today), B = variant (one change) conversion rate = buyers / visitors A: 500 / 10,000 = 5.0%, B: 580 / 10,000 = 5.8% gap = 0.8 points, relative lift = 16% noise about 0.32 points, gap / noise = 2.5, p = 0.012 sample size for a 1 point lift from 5%: about 8,200 per group Chapters: 0:00 Intro: blue button or green button, which wins? 1:10 Control and variant 2:18 Random assignment 3:28 Conversion rate with made-up numbers 4:41 Why small samples lie 5:54 Statistical significance in plain words 7:01 P-value intuition 8:12 Sample size: how many visitors you need 9:23 Common mistakes: peeking and too many variants 10:32 Summary and call to action #abtesting #datascience #statistics #pvalue #experimentation