What Is A/B Split Testing
A lot of marketers recommend you ab split test everything.
You should split test your landing pages, you ads and everything possible :).
But most marketers don't do split testing.
So what is split testing or A/B split testing?
It is simply comparing two things, in your case it could be ads or landing pages and after some time picking a winner based on conversions usually.
Sometimes you'll also want to split test to get the best CTR.
For example, if you have two versions of a landing page, over time, one will convert better than the other.
But how do you know when to stop?
If you get one conversion, should you stop?
Two conversions? Three?
There are many statistical significance calculators that will help you determine when to stop your a/b split testing and call a winner.
Let's do an example.
We have two landing pages LP1 and LP2.
Let's call LP1 the control (the original one) and LP2 the variation (the one I am testing).
I use this calculator a/b split testing calculator.
After sending some traffic to both LPs I plug my results into the calculator and I get these results at t1:
You can see that LP2 is doing better than LP1, but it has not reached the 95% I need to call it a winner.
So I send more traffic to both LPs.
I plug again the results in the calculator and I get these results at t2:
As you can notice, things have changed now.
LP1 is doing better than LP2.
But still LP1 has not reached 95%, so I need to keep on sending traffic to see what happens.
Here are the results after some more time, at t3.
Finally I have a winner. LP1 is my winner.
You see what a disaster it would've been if I had called LP2 a winner at t1?
This is why you need to be patient and let the calculator decide the winner.
Here is a good calculator if you want to check the best CTR for your ads
I hope this helps clarify how to split test your ads and your landing pages.