Facebook split testing-your passport to better performance

Facebook split testing also known as A/B testing lets you test different versions of your ad so that you can see which one best work for your target audience. For example, you create two versions of an ad with two different designs (image on the left and image on the right)

Suppose you ran the ad campaign with design on the left between two different audiences, one is female age group of 21-30 years old, and the other is female age group of 30-40 years old. Also, you spend $150 for this ad campaign. After the end of the campaign, you got 10,000 impressions. Then you ran another ad campaign with the design on the right. In this case, too, you kept everything the same like the previous campaign. Only you change the design. At the end of your second campaign, suppose you got 8,000 impressions. As you can see the first ad campaign with design on the left is more productive. So it is the winning ad version. In this way, split testing or A/B testing allows selecting the better version of your ad.

Split testing reveals another crucial information. It helps you determine which segment of your audience is more likely buy your products. Let’s go back to our previous example. Suppose, you got more impressions from the age group of 21-30 years old in both ad campaigns than the age group of 31-40 years old.This implies that the age group 21-30 is more responsive and more interested in your product. Consequently, your focus group should be age group of 21-30 years old.

How split testing works

At the heart of Facebook split testing is experimenting with several Facebook audiences. Here how it works:

• Split testing divides your audience into random, non-overlapping groups.
• After dividing the audience into smaller segments the ad set is run with the same design, budget. But each ad set differs with one another in one distinct aspect. This unique aspect is called a variable.Variables available for split testing are

1. Target audience
2. Delivery optimization
3. Placements
4. Creative

• The performance of each ad set is measured according to your campaign objective. An object is what actually you want to achieve. Available split testing objectives are:

1. Traffic
2. App installs
3. Lead Generation
4. Conversions
5. Video views
6. Catalog sales
7. Reach
8. Engagement

• Facebook split testing is based on people. It gathers results on multiple devices.
• After the test is complete, you will get notification and email containing results.

How much will it cost to run a successful Facebook A/B test?

The more ad variations you create, the more impressions and conversions you will receive. This means that you will have to handle more data to analyze to find your desired ad version and target audience. To get a valid A/B test result, you require at least 100 conversions for each ad variation. If your cost per conversion is $2.50 and you want to test four ad variations than your advertisement budget should be $1,000. In order to get an accurate result, you should test at least 2 or 3 ad variations.