Free online toolA/B test calculator

Variation B improved conversion rate by +5.0%, going from 8.00% to 8.40%.

You can be 97.88% confident that this result is a consequence of the changes you made and not a result of random chance.

What would happen if you got more users in the experiment? If B converts better? Play with those sliders and find out!

8.0% of the 50,000 who saw variation A converted. This is also known as the "observed conversion rate".

How far the true conversion rate is likely to be from the observed conversion rate.

There is a 95% chance that the true conversion rate is within ±0.24% of the observed conversion rate.

Increasing the number of users will improve precision.

8.4% of the 50,000 who saw variation B converted. This is also known as the "observed conversion rate".

How far the true conversion rate is likely to be from the observed conversion rate.

There is a 95% chance that the true conversion rate is within ±0.24% of the observed conversion rate.

Increasing the number of users will improve precision.

Variation B has a +0.40% absolute improvement in conversion rate compared to variation A. Also known as the "observed difference".

Variation B has a +5.0% relative improvement in conversion rate compared to variation A.

While the observed difference is +0.40%, the true difference is likely to be within ±0.174% of that value.

The z-score can be used to compare the results of different A/B tests

The further from 0 the value is, the more confident you can be that there is a real difference between the variations.

There is a 2.1% probability that random chance alone is responsible for the observed difference.

Increasing the number of users or the difference in CR will improve the p-value.

A/B testingRun experiments with Tggl and Amplitude

From setting up an A/B test experiment to analyzing results in Amplitude, see how Tggl can help you achieve your goals faster.