What is 'A/B testing' in data analytics?

A/B Testing process:  • Define hypothesis (Version B will improve conversion by 10%) • Randomly split users (50% see A, 50% see B) • Measure outcome metric (click-through rate, conversion, revenue) • Apply statistical significance test (t-test, chi-square) to determine if difference is real or chance • Make data-driven decision.  In this testing, statistical power, sample size, and running time are critical.  Banking A/B tests can be done to find: • Which loan offer email subject line generates more applications?  • Does a simplified KYC flow improve account opening completion?  • Does a chatbot reduce call centre volume?