Our calculator uses the Chi-Squared test (χ² test), which is the most appropriate statistical test for comparing conversion rates between two groups in A/B testing. This test compares the observed frequencies (your actual conversion numbers) with the expected frequencies (what we would expect if there was no difference between versions) to determine if any observed difference is statistically significant.

A result is statistically significant if it's unlikely to have occurred by random chance alone. In A/B testing, when we say a result is statistically significant at the 95% confidence level, it means there's less than a 5% probability that the observed difference between versions is due to random variation rather than a real effect of the change you made.

For most A/B tests, 95% confidence (p-value < 0.05) is the standard threshold. This means you're willing to accept a 5% chance of a false positive (concluding there's a difference when there isn't one). Use 99% confidence when the cost of implementing a false positive is very high (e.g., major website changes that would be expensive to reverse). Note that higher confidence levels require larger sample sizes to achieve significance.

The duration depends on your traffic volume and the expected effect size. As a general rule:

• Run tests for at least 1-2 full business cycles (usually a week to account for weekly patterns)
• Ensure you have enough visitors to reach statistical significance (our calculator can help estimate this)
• Don't stop tests too early - early results often show false positives due to random fluctuations

The required sample size depends on your baseline conversion rate and the minimum effect size you want to detect. As rough guidelines:

• For small changes (5-10% lift): 10,000+ visitors per variation
• For moderate changes (10-20% lift): 5,000-10,000 visitors per variation
• For large changes (20%+ lift): 2,000-5,000 visitors per variation

Our blog has a detailed article on sample size calculation with specific formulas you can use.

If your test doesn't reach statistical significance, it typically means one of three things:

1. The change you tested had no meaningful impact on conversions
2. The effect was too small to detect with your sample size (you might need more traffic)
3. There was too much variability in your data (consider segmenting your audience)

In this case, you might want to continue the test with more traffic, test a different variation, or accept that the change didn't produce a measurable improvement.

This calculator is designed specifically for A/B tests (comparing two versions). For multivariate tests (testing multiple variations simultaneously), different statistical methods are required. However, you can use our calculator to compare any two variations from your multivariate test by isolating their data.

Statistical significance tells you whether an observed difference is likely real (not due to chance), while practical significance considers whether the difference is large enough to matter for your business.

For example, you might find a statistically significant 0.5% increase in conversions, but implementing the change might not be worth the effort for such a small improvement. Always consider both statistical and practical significance when making decisions.

The p-value is calculated using the Chi-Squared test statistic, which compares your observed conversion counts with what we would expect if there was no difference between versions. The exact calculation involves:

1. Creating a contingency table of your conversion data
2. Calculating expected values for each cell
3. Computing the Chi-Squared statistic
4. Determining the p-value from the Chi-Squared distribution

Our calculator handles all these complex calculations automatically so you don't need to worry about the math.

While our calculator uses proper statistical methods, it's primarily designed for marketing, UX, and business A/B testing. For medical research or other scientific applications where stricter standards apply, we recommend consulting with a statistician or using specialized statistical software that can account for additional factors specific to your field.