Note: This feature is only available to users with a Campaigns Pro license. For more information on how to assign a Campaigns Pro license to your account, see Assign Premium Feature Licenses.
What is A/B testing?
A/B testing, also known as split testing or multivariate testing, allows you to experiment with your message content and determine which version is most effective at driving better open rates, reply rates, and overall engagement.
Within any one-time campaign, create up to three different versions of your message and we will distribute them equally to your recipients. Afterwards, you can compare the performance analytics for each variant to inform future messaging strategies.
What are the best practices for A/B testing?
When creating your A/B test campaign, keep in mind the following best practices in order to get the most out of your results:
Test the right pieces
Pick a specific part of your message to experiment with and test each one at a time to get clear data on what strategies are driving success. Your A/B test can include variations in attachment types, the tone of your copy, or calls-to-action but not all of these at once.
Use a large sample size
When running a test, make sure to choose a large enough number of recipients that you can feel confident in your A/B test results. We recommend sending your A/B test campaign to at least 100 contacts. The performance of a message sent to a small group won’t necessarily be indicative of that message’s performance on a larger scale.
Optimize your schedule
Utilize TextUs data to understand your audience segment and the best performing open and reply times of day. Schedule your campaign accordingly to maximize engagement.
Send a campaign with A/B testing
A/B testing is available to Campaigns Pro users on all one-time campaigns. To create an A/B test campaign, complete the following steps:
Go to the Campaigns page and click New Campaign.
Under Choose campaign type, select One-time campaign, then click Next.
On the Select Recipients page, add contacts to your campaign by uploading a spreadsheet, or selecting a group of contacts. After adding your contacts, click Next.
Enter a title and write your first message variation for the campaign. You can use the message options to add customizations such as your signature, emojis, or personalization.
To add a new message variation, click Add Message. Write your second message variation.
If you’d like to add a third message variation, click Add Message on the bottom left corner of the page and compose your third variation.
Once you're done creating your campaign messages, click Next.
On the Review and Send page, review the message previews for your campaign, then complete the following steps:
Use the Date and Time dropdowns to schedule when your campaign message will be sent. If you want your campaign to send immediately, leave the settings at the default date and time.
Use the Assign me to campaign conversations toggle to be automatically assigned to the campaign-related conversations. This setting allows you to be notified of any inbound campaign responses depending on your notification settings.
Click Send Campaign.
Viewing A/B test analytics
After you send your A/B testing campaign, you can track the performance of each message variation to determine which was most effective. To view these metrics, go to the Campaigns > One-Time tab and select your campaign from the list.
On the Overview tab, we display all message variations. When you click the arrow symbol
next to each message, you can view the number of recipients they were sent to and their response rates for quick comparison. We automatically highlight the variation with the highest response rate as the Top Performer.
On the top menu bar next to Overview you can click on the tab for a particular message variation in order for more information about that variation’s performance, such as response times and delivery failures.
In the example below, we’re looking at the detailed performance metrics for Message 2. On the left menu, you can choose to view subsets of your recipients by delivery status (sent, delivered, failed, omitted due to opt-out) or response status (replied, unreplied).
