A/B Testing
A/B testing compares two versions of something — a page, a button, a flow — by showing each to a portion of real users and measuring which performs better against a target metric. It's a product and marketing technique more than a QA one, but QA teams often build and verify the infrastructure that runs it.
From a testing angle, A/B testing raises specific QA concerns: does the variant assignment actually split traffic correctly, does each variant render and function correctly, and does the analytics tracking attribute results to the right variant. A broken A/B test doesn't crash — it just quietly produces wrong conclusions, which is arguably worse.
Feature flag systems usually power A/B tests under the hood, so testing an A/B rollout often means testing the flag logic itself: that users land in the correct bucket, stay there consistently across sessions, and see a fully functional experience in every variant, not just the default.
