Partition Sets (Legacy)
Admin access or Domain access to property with feature access to message/scenario partition set management
A partition set allows your organization to divide end-users who visit a property into one or more distinct partitions (groupings) who can receive different messaging scenarios. By utilizing a partition set, you can A/B test different scenarios on property visitors and/or manage a controlled rollout of your messaging experience. In this article, we will cover how partition sets are divided (bucket start/end) and use cases for partition sets.

Partition %

The message scenario that is shown to your end-user is determined by the Sourcepoint messaging library running on the property the end-user visits. Your organization can determine what percentage of end-users are shown each message.
In the example below two messages have been created with different themes. A partition set is a collection of partitions with a total partition size is 100%. Your organization should set each partition to a message and a scenario. The 'Partition %' is the percentage of end-users who are shown a particular message.
Partition settings
Partition example
Setting
Description
Name
The partition of the name assigned by your organization
Partition %
The percentage of end-users shown a specific message
Scenario
The scenario connected to the message

Buckets

The messaging scenario that is surfaced to your end-user is determined by which bucket the end-user has been assigned to when they visit a property. Each end-user that visits a property with the Sourcepoint messaging library running on it will get assigned a random value from 0 to 999. This value is stored in the end-user's cookie and persists until the user deletes the cookie.
By setting the start and end bucket values of each partition, you can control the percentage of end-users who are served a certain message scenario.
The bucket values will determine what percentage of users will experience the selected scenario. For example, 0 - 999 means that 100% of users will experience the scenario whereas 0-499 means 50% of users will experience the scenario.
In the below example. we have evenly split the partition set into two partitions. The first partition (0-499, 50%) will receive the default scenario while the second partition (500-999, 50%) will receive the test2 scenario.

Use Cases

A/B Testing

Partition sets can be used to A/B test messaging experiences where a certain percentage of users see one experience and another percentage of users see a different experience. This requires that you set up as many scenarios as you have partition groups.
In the example below, the partition set has 2 partitions: Group A and Group B. Group A users will be assigned a bucket value between 0 - 499, representing 50% of users. Group B users will be assigned a bucket value between 500 - 999, representing the other 50% of users. Group A users will experience Scenario 1 and Group B users experience Scenario 2. Scenario performance can be measured using Sourcepoint reporting to inform your strategy moving forward.
A/B testing

Controlled Rollouts

Partition sets can also be used to control the percentage of users being exposed to a new messaging experience. For example, if you would only like 10% of your users to experience a messaging scenario, you can create a partition set with a single partition and that partition is set up to only show to users with a bucket between 0 - 99.
Could not load image
Last modified 2mo ago