Tessa Everly
Reddit has officially launched its split testing feature, making it accessible to all marketers on the platform.This announcement follows a successful testing phase with select advertising partners over the past several months..
This new feature allows marketers to simultaneously test two different ad variations to identify which performs better. The top-performing ad will then serve as the basis for a more extensive advertising campaign.
Reddit explains, “In a controlled experiment, you divide your audience at the user level, execute two ad flights with just one variable changed, and determine the winning ad with a 65% confidence level. There’s no overlap, no contamination, and no guesswork involved.”
The duration of Reddit’s split tests can range from two to six weeks, ensuring that advertisers gather a substantial response sample to accurately identify the winning ad variant.
Advertisers can access Reddit’s split testing as a self-service option through the Reddit Ads Manager. This means they can activate the feature without needing direct assistance from a Reddit ads representative.
Furthermore, Reddit provides a library of pre-designed templates for split testing, simplifying the process of setting up experiments.
“Select what you want to test, and the system manages the rest — no specialized knowledge is necessary to conduct a well-structured experiment,” Reddit stated.
This approach opens new avenues for advertisers to explore different strategies on Reddit, helping them connect more effectively with the platform’s unique audience.
In initial testing phases, Reddit reported that four out of five split tests successfully identified a winning variant based on return on ad spend (ROAS).
David Bridges
Cameron Steel
