Designs and reads quantitative UX evidence honestly: hypothesis-first A/B tests with a traffic/duration feasibility gate, a decision tree for when A/B is the WRONG tool (why-questions, low traffic, strategic redesigns), statistical vs practical significance, the analytics what-not-why limit with a m
Designs and reads quantitative UX evidence honestly: hypothesis-first A/B tests with a traffic/duration feasibility gate, a decision tree for when A/B is the WRONG tool (why-questions, low traffic, strategic redesigns), statistical vs practical significance, the analytics what-not-why limit with a mandatory follow-up qual spec, and the firewall against reporting metrics from 5-user studies. Use wh