Relative lift compares the treatment outcome with its control baseline. It's different from an absolute percentage-point difference. If the baseline is small or zero, relative lift can be unstable or undefined. Keep the original counts, denominator and timing available so the apparent improvement is interpretable.
A positive point estimate alone doesn't establish a reliable winner. You also need a suitable design, enough information, conversion maturity and a decision rule defined before reading results. The browser shows a deterministic fictional lift to practice reading a report. It supplies no significance test. Use its readout to learn the arithmetic and limits, then use a real experiment's supported statistical evidence for a live decision.