Fast A/B Testing with Facebook Ads

In this article I will explain a fast and simple way to quickly test different ad creative elements and come up with the ultimate ad which you can scale afterwards.

We test different parts of an ad in four steps, using five different versions for each part. This helps us see what grabs people’s attention the most without spending too much money or time.

By the end, we know exactly what mix of ad parts works best and can use that to reach more people effectively.

Minimum Requirements for Testing:
  • 5 versions of the thumbnail headline text
  • 5 versions of the thumbnail image
  • 5 versions of the ad copy text
  • 5 versions of the call-to-action headline
We initiate 5 unique Traffic campaigns, each optimized for link clicks:
  1. Phase 1 – Image Headline – Traffic – [Unique Name]
  2. Phase 2 – Image – Traffic – [Unique Name]
  3. Phase 3 – Copy – Traffic – [Unique Name]
  4. Phase 4 – CTA – Traffic – [Unique Name]
Ad set settings:

Ad placements are exclusively in the FEED to minimize variation, as differences in Unique Outbound CTR% can be significant across various placements. If an ad performs well in the feed, it’s likely to succeed in other placements when we scale the winning ad.

Go to adset settings > Placements > Only Choose Facebook & IG Feed

Setting Facebook Rules

It’s crucial to note that launching the campaign as is would prompt Facebook to allocate most of the budget to the best-performing ad set, preventing each from reaching the desired 10,000 reach.

To circumvent this, we establish Ad Rules directing Facebook to deactivate an ad once it hits 10k reach, forcing the remaining budget onto the other ads. This strategy ensures all ads achieve significant exposure of 10,000 reach.

Choose the ads you want to setup rules for and choose Create new rule
Depending on your budget choose at least 500 reach

The testing sequence prioritizes the thumbnail headline, followed by the thumbnail image, post copy, and finally, the call to action.
Whether you opt for a reach of 10k, 1k, or 500 per ad depends on your budget, but we recommend at least 500 to base decisions on substantial data.

Test Campaign Launch

We kick off with the Phase 1 campaign, which includes one ad set containing 5 ads, all identical except for different Image Headlines.

We setup rules for all 5 ads – (Whether you opt for a reach of 10,000, 1000, or 500 per ad depends on your budget, but we recommend at least 500 to base decisions on substantial data.)

After setting up all campaigns and rules, Facebook executes the campaign, deactivating ads as they hit the reach threshold until all are switched off.

Rules for this particular ads were to reach more than 500 users and then turn off

In the end, we can assess the performance of 5 distinct ads, each with at least a 500 reach, to determine the most successful one based on the unique outbound CTR%.

Rules for this particular ads were to reach more than 500 users and then turn off

In above example Ad number 1 is the winner, which means that the headline that went on this particular ad will move to the second phase.

Replicate the approach

Phases 2, 3, and 4 follow a similar structure, with each phase building on the winning elements of the previous one. By Phase 4, we’ve tested various combinations of images, headlines, ad copies, and CTAs.

  • Phase 1 – Winning Headline
  • Phase 2 – Winning headlines, 5 different images, same copy, same CTA
  • Phase 3 – Winning headline and image combination, 5 different copy, same CTA
  • Phase 4 – Winning headline, image and copy combination, 5 different CTAs

The outcome is the ultimate ad, comprising the best elements from each category.

You can then scale this ad in broader campaigns optimized for Conversions, employing all available placements.

Our method stands out for its speed and cost-effectiveness compared to traditional strategies that use separate ad sets. While some might question its precision, our experience confirms its efficiency in swiftly identifying potent ad combinations.