A/B test analysis

Act as a data scientist. Here are my A/B test results: [variant sizes, conversions/metrics]. Tell me which variant won, whether the difference is likely meaningful (and what to be careful about with significance and sample size), the estimated effect, and whether to ship, iterate, or keep running.