实用计算器
A/B 测试显著性计算器
输入对照组和实验组的访客数与转化数,本工具在浏览器内完成双比例 z 检验,输出两组转化率、相对提升幅度、双尾 p 值,并按 95% 置信度给出是否显著的明确结论。当结果尚不显著时,还会按当前观察到的提升幅度估算每组所需的最小样本量(80% 统计功效),帮你判断该继续跑测试还是提前结束。适合落地页优化、广告素材测试和 CRO 从业者使用。
主要用途
This calculator decides whether the gap between two variants in an A/B test is a real effect or just noise. Enter the visitors and conversions for your control and variant and it returns each conversion rate, the relative lift, and a statistical significance result — typically a p-value and confidence level — telling you how likely the difference is to hold beyond your sample. It exists because eyeballing a few percentage points of lift routinely fools marketers into shipping changes that were random fluctuations. Use it before declaring a winner on an ad creative, landing page, or email test, and to judge whether a test has collected enough data to trust at all.
所在链路
It sits at the analytics stage, separating real winners from noise before you act on test results.
核心功能
- Conversion rate for control and variant from raw counts
- Relative lift between the two variants
- Statistical significance with p-value and confidence level
- Clear winner / no-significant-difference verdict
- Quick read on whether sample size is sufficient
适合谁用
- Marketers validating ad, landing-page, or email tests
- CRO teams confirming a lift is real before rolling out
- Anyone avoiding decisions based on random fluctuation
新手提示
- Set your sample size and test duration in advance; stopping early when a result looks good inflates false positives.
- Significance answers whether an effect exists, not whether it is large enough to matter commercially.
- Run tests over full weeks so weekday and weekend behaviour are both represented.