I have heard of A/B Split Testing since forever it seems and it has always frightened me to bits in how complex it looked and all that interminable reading. So here, I am going to try to make it as simple as possible so that even I can do it.
I am going to assume a number things here:
Say you have a page that has a square ad running on it. You wonder if having a rectangle ad in the same place, on the same page would make any difference.
In A/B Testing, ideally you want to:
Then from the results (gathered over a couple of weeks at least), you will know which gives you a better return.
I am going to use one of my current ads as illustration for this exercise.
Name of my ad unit/custom channel: JEM.SQ.T3.CTOPLEFT.IMG
(I have named using the same naming format for Custom Channels)
This ad also has an identical Channel name.
Create another rectangle ad that is to be rotated 50% of the time, on the same page, in the same location, for a specific period of time and to analyse the comparative performance of each.
Since I already have one ad in place, I only need to create another ad (with only one different feature - here being: square vs rectangle [new])
1) Create New Ad:
Log into AdSense Account > create a Medium Rectangle ad (300x250) and everything else between the 2 ads are identical except for the size.
==> new ad/custom channel: JEM.MRECT.T3.CTOPLEFT.IMG
2) Get Code of Ad 1 (sq) and Ad 2 (med redct) > paste each into Notepad (you can tell which code is which by the name you have given it).
To the left here, you can see I have copied and pasted both ad's code into my Notepad.
You can see I can tell the difference by the Name (as well as the size in this instance)
I have "(purple) boxed" each code for you to see - as these are the portions of code I want from each, viz.
from the opening to closing comment tag
3) Create one block of code.
This will be what you insert into your page (where you want the 2 rotating ads)
Copy this A/B Test block of code and paste into a new Notepad:
Then copy the "boxed" code from Ad1 (sq) and replace into where the yellow highlighted code is in the A/B Text block.
Coy the "boxed" code from Ad2 (med rect) and replace into where the green highlighted code is in the A/B Text block.
The final combined (AB Test) code will look something like:
Copy this entire block of code and paste into your page where you had the original Square Ad running.
4) That's it. Monitor the performance of both channels
over a period of 2 weeks or so and check against each other.
You should also check against the baseline performance (if you had an original ad running, what was it's performance before you started the A/B Split Testing?
Comparative Performance Criteria you might want to track are:
Article written on 26 Jan 2012