mostakimseoexpatebd
Dołączył: 15 Kwi 2025 Posty: 3
|
Wysłany: Wto Kwi 15, 2025 07:23 Temat postu: Facebook split tests – why and how?! |
|
|
Facebook Red hasn't let Orbiters sleep lately, as it publishes news almost weekly: from split testing to product scans on the website.
Since Facebook doesn't rest, neither do we - that's why in today's post we're introducing you to Facebook's new feature - split testing. You'll find out what we can test, why it's used, and what the benefits are below.
Why Facebook split testing?
In an ideal world, you would set up a Facebook campaign , set up ad targeting, and voila! You would reach your target audience, receive a great response to your ad, and pay a low cost per click. But is that reality? In reality, we often scratch our heads and figure out which targeting would be suitable for reaching our target audience. Now, Facebook has made testing easier for us and introduced split testing. Facebook is well aware that we learn through testing – first we try one thing, and if it doesn’t work, we try something else. But what does this testing look like on Facebook? Read on.
Finding the right targeting for your target audience is like playing poker. You play until you win.
What can we test?
Before you decide to set up Facebook split testing, you should know that the option is not available for all campaign types. You have the option to split test if you choose the following as your campaign objective:
website visits (Traffic),
App installs,
Lead generation and
Conversions.
Within one test, Facebook allows you to test two or three strategies: A/B test or A/B/C test. However, within one mobile database campaign, you can only test one variable. The variables you can test are:
target audience,
advertising delivery (Delivery optimization) and
advertising placements.
Set up a split test in Ads Manager by clicking Create new campaign. Once you select the appropriate campaign type, Facebook will offer you the option to "Create split test".
Target audience
Have you ever thought, “Hey, who hasn’t tried this before?” As budding entrepreneurs like us at Red Orbit, we’ve often tested different target groups within a single campaign. The downside of this approach is that ads compete with each other. In practice, this can mean that individuals have seen a similar ad from two different ad sets, which means less accurate data for us about which target group performed better.
Facebook completely separates the target groups from each other when doing split testing and prevents any competition between ads. Of course, Facebook wouldn't be Facebook if it didn't have its own rules. If you're going to test target groups, you can only do so with your saved audience. So before testing, make sure you have the target group you want to use for the test saved.
Delivery optimization
Delivery optimization has two different tests to choose from. You can test ad delivery or bid amount. Facebook recommends the following tests:
Optimization for conversions (A) vs optimization for link clicks (B).
Optimize for conversions with a conversion window of 1 day (A) vs. optimize for conversions for a conversion window of 7 days (B) vs. optimize for link clicks (C).
Facebook gives you many choices for testing. Choose wisely!
Click To Tweet
Of course, you can choose a different testing method, but the fact is that you can test as many as three different ad sets at once. However, it is important not to test different variables, as this can lead to incorrect data.
Placements
Here, Facebook gives you a lot of options and you can really get creative when testing. You can test if your ads perform better in the news feed or the right column. What can you test when placing your ads?
What now?
Once you have selected your test campaign type and variable, you need to decide on the distribution of money and the testing time. You have the option of distributing the money evenly, or you can decide for yourself how you want to distribute it. Even distribution is recommended, as this will give you the same conditions for testing. The test should last a minimum of 3 and a maximum of 14 days. Do not change any settings during the test; Facebook will notify you when your split test is complete. _________________ fresh mobile database |
|