What is A/B testing?

A/B testing allows you to test two variations of a single variable, on Facebook and Google, common A/B tests are run to discover the most effective age, gender, placement or lookalike percentage.

To run an accurate A/B test the settings and budget should also be set up identically to ensure that the chosen variable is the only differentiating factor.

If, for example, you wanted to test Females Vs. Males, you would set up the audience in the same exact way except for the gender variable which is the "split" between the two Ad sets. One Ad Set will run to target Male Audiences while the other audience will run to target Females.

Once you've run your experiment for a few days you can identify which test had the best results by looking at  KPIs such as your ROAS or cost per click, click-through rate, CPM, or otherwise.

With Madgicx, setting up an A/B test is really quick and easy. Follow the steps below to run an A/B test within a few clicks.

Step 1:
First, Select the audience from above that you wish to run your A/B test on.

Make sure you have selected your audience, recency, conversion type, location, age, gender, and placement for this audience.

Step 2. 

Click on the duplicate icon which will create a 'clone' of this audience


Step 3.
Select the First Ad-Set and set the differentiating factor for the First Ad-set. In the image example below, we're changing the lookalike percentage. However, you can do the same for any variant such as age or gender. Once you've done so, select

'save changes'.


Step 4.  
Click the 'cloned' Ad-Set and change the same differentiating factor.

Step 5.

Once you're ready, select 'save changes' in the top right corner to continue to launch your Ad Sets.


Please note: You can create as many duplicates as you want just keep duplicating the Ad Sets. You can also run an A/B test to test your best creatives, in this case you can duplicate your audience as we've done here and choose different lookalike percentages for this audience.

And that's how you run an A/B test in a few quick steps! If you have any further questions about A/B testing on Madgicx, be sure to reach out to our support wizards through the chat! 🧙‍♂️

Did this answer your question?