Best Strategy for Split Testing Ad Images in Facebook

by jcss
6 replies
I want to find the best image for a chosen demographic from 5 images. Here is my strategy.

If I put 5 ads with the same title and description which only are different in ad images in a campaign, 1 or 2 of the ads will get the most impression while others will not suffer. To solve this issue, I create 5 different campaigns and put each ads separately in those campaigns. This way, they won't compete with each other and all the ads will have a chance to get impressions and possibly clicks.

Now the question is: Does this method work? Will those 5 ads in separate campaigns compete with each other to grab the impression or are they being displayed evenly?

What is your strategy to split test ads in facebook?
#facebook #images #split #strategy
  • Profile picture of the author ChrisBa
    My strategy is similar, I start with 2-3 different ad texts, 2-3 different ad bodies and 5 images, which gives 20-45 ad variations, then start to weed out the losers and work on the winners
    {{ DiscussionBoard.errors[8632781].message }}
    • Profile picture of the author jcss
      Originally Posted by ChrisBa View Post

      My strategy is similar, I start with 2-3 different ad texts, 2-3 different ad bodies and 5 images, which gives 20-45 ad variations, then start to weed out the losers and work on the winners
      You didn't answer the rest of my questions!
      {{ DiscussionBoard.errors[8633148].message }}
  • Profile picture of the author JustinDupre
    Your method is a sound method. It still hickups from time to time but generally gives a much smoother impression count spread across all your ads.
    Signature
    I offer CPA coaching and investment opportunities for those SERIOUSLY interested in making money directly or indirectly with affiliate marketing. PM me for details.


    Read More about CPA/Affiliate Marketing on my Blog
    {{ DiscussionBoard.errors[8633172].message }}
    • Profile picture of the author jcss
      Originally Posted by JustinDupre View Post

      Your method is a sound method. It still hickups from time to time but generally gives a much smoother impression count spread across all your ads.
      Thank you but would you please comment on my question as well?
      {{ DiscussionBoard.errors[8633249].message }}
      • Profile picture of the author JustinDupre
        Originally Posted by jcss View Post

        Thank you but would you please comment on my question as well?
        That answered your question. It should evenly display ads. Whether or not they compete against each other is completely in the air. If you're targeting 5 ads to a crowd of 100,000 people, there is no sound way to test that unless you have inside knowledge of the algorithm Facebook is using to choose and display ads.
        Signature
        I offer CPA coaching and investment opportunities for those SERIOUSLY interested in making money directly or indirectly with affiliate marketing. PM me for details.


        Read More about CPA/Affiliate Marketing on my Blog
        {{ DiscussionBoard.errors[8633324].message }}
        • Profile picture of the author jcss
          Originally Posted by JustinDupre View Post

          That answered your question. It should evenly display ads. Whether or not they compete against each other is completely in the air. If you're targeting 5 ads to a crowd of 100,000 people, there is no sound way to test that unless you have inside knowledge of the algorithm Facebook is using to choose and display ads.
          Ok, thanks! Do you mind to share your own strategy on how to split test different variations of ads?
          {{ DiscussionBoard.errors[8633460].message }}

Trending Topics