Split testing in SMS campaigns helps you identify which message variant performs best with your audience. You can test on a sample and automatically send the winning variant, split your audience evenly across all variants, or set a control variant to benchmark against. Each method provides performance insights to help you optimise future SMS campaigns.
Once your SMS split test variants are ready, you can choose your audience, set the send time, and configure how the test runs.
Before you start
Things you need to know:
You must already have set up your split test campaign and added your content to each SMS variant.
Split tests must only include one link in each variant, as the winning variant is determined by the number of link clicks.
Configure and send a split test campaign
After you create and save your split test variants, you continue to the Contacts step.
Set up who receives your campaign, when it sends, and how your split test runs.
Select the contacts to send this campaign to
Choose the lists, segments, or marketing preferences you want to send your campaign to.
At this stage, you can add or remove recipients to tailor your audience.
Select send time and speed
Set SMS expiry:
For undelivered SMS, select Expire after three days to use the standard three-day expiry time.
To expire earlier, select Expire on, then choose a date and time.
The expiry must be between 1 hour and 72 hours (three days).
Set up split test
Once you’ve chosen your audience and send time, configure how your split test runs. There are three configuration options:
TEST THEN SEND WINNER
Send each campaign variant to a sample of your audience, measure results, and then automatically send the winning version to the remaining contacts in your audience.Select the percentage of audience to test - use the slider to choose what portion of your audience receives the test versions.
Set test duration - choose how long to measure responses before picking a winner, in hours.
Test summary displays variant count and test group sizes as you adjust settings.
SPLIT AUDIENCE EQUALLY
Divide your audience evenly among all variants.Sends each variant to a portion of your entire campaign audience, dividing contacts evenly across all versions. There’s no automated winner send, every contact receives one variant, and you review the campaign results afterwards to identify which performed best.
Example: With 4 variants and 8,000 contacts, each variant is sent to 2,000 recipients.
Test summary shows how your audience is split.
SET CONTROL VARIANT
Use a control variant to establish a baseline for comparison. This is usually your standard version. When you run the test, part of your audience receives the control, while the rest receives one of the test variants.
Unlike other split test types, this method doesn’t send an automatic winner. You’ll need to review performance manually to decide which variant worked best.
Select control variant - choose which variant is the control.
Control audience percentage - set the proportion of your audience to receive the control variant.
Test summary updates to show variant and control audience allocation.
Send your campaign
Select CONTINUE.
Review your campaign details. To send, select SEND CAMPAIGN IMMEDIATELY or CONFIRM SCHEDULED SEND.
The option you see here depends on whether you chose to send an immediate or scheduled campaign in Step 3.
Undelivered SMS
If a network can’t deliver an SMS at the first attempt, they make multiple further attempts to deliver up to the message expiry time.
The custom expiry is useful if your SMS is time-sensitive, for example, a 24-hour sale or travel updates, so you don’t want the network to continue trying to re-send the SMS for the full three days.
SMS split test campaign report
To learn more about split test reports for SMS, see SMS/MMS report.