Enter your base conversion rates and we will show you how many visitors do you need for your test.
The reason why you need to calculate the sample size before starting a test is that it helps maintain the validity and reliability of your results.
Think about it this way, if your sample size is too small, your outcomes might be skewed by random chance or unexpected variables. On the flip side, if your sample size is too large, your study may be time-consuming and costly without necessarily improving the accuracy of the results.
By calculating your ideal sample size at the start, you can strike the perfect balance. You ensure enough participants to represent the broader population you're studying, while making the best use of your time and resources. Moreover, if the results are to be used for decision-making processes, a suitable sample size enhances the credibility and reduces the risk of incorrect decisions being made e.g. the risk of type II errors.
In short, knowing the right sample size benefits you in conducting a precise, credible, and cost-effective test or study.
The Base Conversion Rate is the conversion rate of your defined goal, before any changes or tests are made.
It is used as a benchmark when you start experimenting with changes to see if they cause an increase or decrease in that rate, it helps you understand how many observations you need to detect a significant change or effect.
The Minimum Detectable Effect (MDE) is a statistical term referring to the smallest effect that a particular test or experiment can reliably detect. It's a crucial concept in areas like A/B testing, research design, and statistical power analysis.
The MDE depends on factors such as your sample size, statistical power, and the statistical significance level you've set for your test. For example, if your sample size is larger, or if you're willing to accept a higher level of uncertainty in your results, then your MDE will be smaller.
Essentially, MDE gives you a sense of the sensitivity of your experiment. It helps you determine whether your test is appropriately designed to detect the level of effect that you're interested in. If the expected effect is smaller than the MDE, it suggests that the experiment might not be powerful enough to reliably detect this effect and therefore, adjustments might need to be made.