A/B test is an experiment where two or more variants of setup or entities are compared to determine which one performs better. In Wise.blue, you will be setting up an A/B test for two or more campaigns and deciding on the audience split which will make sure that a user sees ads only from one of the campaigns.
Before you start, make sure you see the A/B test button in the interface.
If the button is not there, please contact us via the chatbox in the bottom right corner.
Button itself will be available when campaigns for A/B testing have been created and selected.
Note:
- Currently, A/B testing is available only for campaigns. Ad group A/B testing is coming soon.
- A/B testing can be performed only for new campaigns that were not active in the past.
- Campaigns should be scheduled to start in a couple of hours after creation at least.
Here are several ideas what you can test:
1. The same audience and setup, different creative message
2. The same audience and setup, different CTA
3. The same audience and setup, different tweet
4. The same audience and creatives, different bidding type
5. The same audience and creatives, different bidding unit
6. The same creatives and setup, different audience - in this case A/B test will make sure that audience overlap is not happening
Campaign creation
Start with the campaign creation. Campaign, group and ad setup is usual. To help you set up campaigns faster, copy already existing campaigns and just carry out the necessary edits. There are two ways to copy a campaign:
1. Select the campaign, click the Copy button on the top, and select the With ads to the same account option. In the next step, a small window will open and if ads exist in this campaign you will be able to choose if you want to copy them or not.
2. Access the 3 dots menu next to the campaign name and select Copy. In the next step, a small window will open and if ads exist in this campaign you will be able to choose if you want to copy them or not.
Campaigns and groups can be edited via the 3 dots menu next to the entity after which you select Edit. Created ads can't be edited, they should be recreated. Cards for ads can be duplicated and edited. Editing an existing card without duplicating it, will propagate card changes to all existing ads that use this card.
Enabling A/B testing
When campaigns for the A/B test have been created, select campaigns and click the A/B test button. If the button is not available, campaigns' start time is too close, and it should be set for some later time.
In the opened window, give the name to the A/B test, set the time period* for the A/B test to run, and put in audience split ratios - the sum of audience size should be 100%.
*Note on A/B test time period:
- The start date should be 1 minute before the earliest campaigns' start date. It should be automatically set for you by the system.
- The end date should be at least 1-5 days after the latest campaigns' start date:
- For non-app-based campaigns testing time period should be at least 1 day;
- For app-based campaigns testing time period should be at least 5 days.
When all is set, press Confirm.
Filtering out A/B test campaigns
The symbol a of chemical flask under the campaign name means the campaign is participating in an A/B test.
By clicking the symbol, you will filter out campaigns that participate in this specific A/B test.
Other way of filtering out campaigns can be found on the Account level. Click the filter button in the left right corner and filter by specific A/B test(s) or by currently active A/B tests.
Results
Results of A/B test can be accessed the same way as results for other campaigns in Ad Manager by customizing columns for metrics you want to see.