Online Learning Module 9: Learn how to use split testing (commonly known as A/B testing) to identify the audience that resonates most strongly with your ad campaigns.

As an advertiser, have you ever wondered how to analyze the best audience in your ad campaigns? For example, is there a segment within your audience that responds to your ads better than others? Split testing, also known as A/B testing, is a testing tool provided by the Tuke Ads Manager platform. It helps you test different versions of your ads to find the best-performing one, thereby improving future ad campaigns.

The split testing tool divides your ad audience into two groups, ensuring that each group of users only sees one version during the test period. Each user group will then see a different test version, helping you identify the best combination strategy. For example, if you want to conduct an A/B test on different audiences, you can define the targeting characteristics of Audience A and Audience B when creating ad groups, while keeping all other ad settings the same. This automatically creates two ad groups with identical ads and settings, except for the targeted audience. The system will then automatically identify the winning group, which helps you find a more valuable audience. Compared to creating two ad groups separately, using the split testing tool is a better choice. This ensures you get more scientific results, and each user will not see both versions of the ad.

On the Tuke Ads Manager platform, you can easily create a split test with one click. When creating an ad campaign, simply toggle the icon to create a split test. Choose the type of split test you want to perform, including audience split testing and creative split testing. For each split test group you create, you can only select one variable. Once the ad starts running, you cannot change the test variable, but you can modify other options in the ad group. One factor to consider when building a split test in the Tuke Ads Manager is the ad group-level budget. In the campaign budget, you can find the Estimated Testing Power tool. This is a visualization tool for the confidence level of test results. We recommend a confidence level above 80%. Below 80% means the data volume may not be sufficient to ensure the statistical significance of the test results. After the ad has run for 24 hours, go to the ad group and click "View Data" to see real-time data for the test. The page provides easy-to-read visual charts, allowing you to easily view breakdowns of key data. The system determines the winning test group based on the cost per result for each group.

Split testing may produce one of the following results: a winning ad group is found, or no winning ad group is found. If no winning test group is found, click "Create New Split Test" and try adjusting the budget, test duration, and test variables. If a winning test group is found through testing, click "Run Winning Ad" to save the ad settings and publish it as a new ad group. After learning how to set up split tests and view results, please review our best practices: 1. Set an appropriate test budget; 2. Allocate sufficient test duration; 3. Set a broad audience group; 4. Set test groups with significant variable differences; 5. Ensure the confidence level is above 80%.

For advertisers, running split tests in the Tuke Ads Manager platform is a highly valuable tool that helps scientifically determine the best-performing ads, audiences, or optimization strategies.