Deciding Data

Dynamic Creative vs Split Tests?

May 24th 2022

Split Testing and Dynamic Creative are tools that both try different content. But, they are designed to accomplish different goals. Here is a quick overview of their differences and how best to use them together.

Spoiler: Split Testing is great for learning how to make the best ad creative. Dynamic Creative is great for scaling your incredible creations.

Ok, let's say you are already convinced that your ad creative is one of the most important components of your marketing. You've read the Nielson studies and been around the block a few times. You want your team to have a process for producing the best creative.

Trying Different Creative

Facebook has two features that take different ad components and try them out:

Dynamic Creative: Takes creative assets, then "...mixes and matches them in new ways to improve your ad performance. It allows you to automatically create personalized creative variations for each person who views your ad, with results that are scalable."

Split Tests: Traffic is randomly split between different creative, then results about how people responded to them is put together so that you can understand what worked.

These two features can seem pretty similar. So it can be confusing to decide when to use each.

Different Aims

These two approaches actually have different aims:

  • Dynamic Creative aims to produce a high performing campaign.
  • Split testing aims to help you understand what creative works best.

These goals are similar, but result in some pretty different setups under the hood. Dynamic Creative is designed to help you scale your campaign, use personalization, and help delay folks getting tired of your ads. Split testing isn't about a single campaign. It is about understanding what creative (and what about that creative) works best. It provides lessons to help you understand your customers and produce a better next generation of creative.

Different Choices

Because of these different aims, these two approaches do some key things differently.

Personalization

Dynamic Creative uses personalization to adjust the ad content. Split testing does the opposite and randomizes which person sees different creative.

Because of this, the interpretation of their results needs to be different. Split Tests tell you about the creative for a whole audience; Dynamic Creative tells you about the creative for the people that it thought would like it. In Dynamic Creative, it's system for matching people with ads (and the data it has about those people) are part of the results.

Time

To get the best immediate performance, Dynamic Creative shifts spending between creative based on performance and sometimes based on time. Split Testing does the opposite. Split Tests randomizes where the spending goes for a particular period of time.

The randomized spending allows Split Tests to compare like-to-like traffic between creative. This way we can see the impact of the creative rather than "confounding variables."

Dynamic Creative associates what creative is shown with a particular time. Because a lot can change from one day to the next (like who is online, competitor offers, world events, etc.) these other changes "confound" the results of the creative test and influence the results.

But, this "confounding" isn't that important if you just want a good campaign, which is what Dynamic Creative focuses on. It isn't aiming for understanding. Instead of trying to remove "confounders" Dynamic Creative is working to use them, adapting the content as things change.

Conclusion

Diagram of how Dynamic Creative and Split Testing can work together

Because these tools accomplish different things, they are actually complimentary. Split Testing can help you understand and make better creative. Think of this as an R&D process. Then you can scale your great creative with a Dynamic Creative campaign, to get the most benefit from it. Think of that as the mass production (and mass customization) process.

Deciding Data helps make ad testing easy and understandable. If you would like a tool to help understand your ads. We're happy to chat.