A/B test
An A/B test randomly splits traffic between two versions of something to see which performs better against a pre-set metric.
A good A/B test has one clear hypothesis, one primary metric, a sample size worked out in advance, and a fixed end date. Stopping early the moment one version 'wins' is the most common mistake, and it produces false winners. Low-traffic UAE sites should test bold changes (new offer, new page structure) rather than button colours, because small effects need huge samples to detect. Ad platforms offer their own A/B tools for creative and audiences.
Hypothesis: adding an Arabic headline raises lead CVR for Arabic-browser visitors. Metric: form submits. Run 4 weeks or 1,000 visitors per variant, whichever comes later.


