What an A/B test is deciding
An A/B test compares two versions of an experience on one outcome, such as conversion or average order value. The decision is whether the difference is large and stable enough to ship, not whether a chart looks higher.
The checks analysts are expected to name
You do not need a statistics degree to be credible. You do need these ideas in plain language.
- One primary metric, chosen before you look at the result.
- A sample large enough that a small wobble is not treated as a win.
- Random assignment, so the groups were comparable.
- A readout that states the result, the uncertainty, and the recommendation.
How this shows up in business analytics
The Business Analytics Course treats experimentation as a business skill: define the metric, read the result, and tell a stakeholder what to do on Monday. That is the same muscle as a dashboard insight, with a clearer rule for what counts as a change.

