When should you end a split test?
It’s a question I’m asked often and the answer isn’t as simple as “end a test when you reach 100 clicks”.
In short however, you should end your test whenever it is statistically significant.
This means that there is at least a 95% chance that the results that you get are NOT attributed to pure chance.
Anything lower than that and it’s possible that there was some flaw in your test and the results you got were basically random.
So how do you know if your results are statistically significant or not and you should end your split test?
You can use this free resource:
This is Neil Patel’s statistical significance calculator that will tell you when the results you get from your test are valid.
You simply plug in the number of clicks you got to your site and the number of conversions, and Neil will tell you if the results are statistically significant.
Now I often get asked if there’s a certain time period you should test your offers for – a week or two and so on.
But truth is, everyone’s traffic is completely different, so you can’t use this as the sole factor in determining when to end your split test.
For example one person could generate 10 clicks per week, and another person could have generated 1,000 in that same time period and gotten enough data to plug into the calculator.
So now that we know that the time period is off the table in determining when to end your split test, we can move on to the next point – number of clicks.
So using the previous example, one person generated 10 clicks and another 1,000 – so the person with 1,000 clicks has enough data to end their split test, right?
Although you SHOULD aim for at least 100 clicks in each split test variation, the number of clicks alone does not determine whether you should end the split test or not.
Aim for 100 clicks, and then plug in your numbers into Neil’s statistical significance calculator.
In the example I show in the video, you will see that regardless of the number of clicks, if the approximate number of conversions stays the same, the results can still stay either significant or insignificant.
So focus on getting as many clicks as possible, but then rely on the calculator to tell you if your results as significant and you should end your split test.
👉 Get your free affiliate marketing guide to help you get started: https://www.ivanmana.com
👉 Like my Facebook page!: https://www.facebook.com/onlinemarketingessentials
👉 Also be sure to check out my courses at https://www.amarketingessentials.com/all-courses