BlueConic's Profile sampling feature assigns a random number between 1 and 100 to each profile, enabling marketers to create randomized test groups for campaign experimentation and measurement. BlueConic automatically creates the Sample ID profile property for all profiles. This facilitates A/B testing, control group creation, and performance analysis across various channels and touchpoints.
These values are evenly distributed across all profiles in the BlueConic tenant. This applies to profiles that were imported as well as profiles created from a web visit.
Consider using Profile Sampling if you want to test a campaign across multiple channels and touchpoints. Profile sampling also supports Uplift modelling to increase the Return on Ad Spend (ROAS), and helps train machine learning models faster by using a representative sample of profiles.
Examples of uses include
You can use the Sample ID property to create random groups of profiles. Ways to use random sampling to optimize your marketing include:
Use test groups to perform experiments across multiple messages or dialogues.
Perform comparative multichannel marketing campaigns.
Measure the effects of using an orchestrated Lifecycle against an existing campaign.
Create a 'hold-out group' of profiles and exclude them from particular campaigns.
Run an experiment for a holiday sales campaign, for example, where 10% of your customers see a dialogue and receive an email, while a control group sees only a dialogue.
Target 80% of customers with a personalized email vs. 20% with a non-personalized email and compare opens or clicks.
Measure web push notification success, using BlueConic insights.
In AI Workbench, you can train AI models against a random sample of profiles instead of against your whole profile database.
Before you begin
The Sample ID profile property is automatically generated and populated for all profiles. No prior setup is required to use this feature.
Determine whether you need to measure the effectiveness of a campaign across multiple channels or compare different campaign strategies.
Collaborate with marketing and analytics teams to define objectives and KPIs.
Ensure you have existing segments or criteria to define your target audience.
Sample ID property
You can use the Sample ID property to create random groups of profiles.
The assigned numbers are evenly distributed across all of your profiles, including visitor profiles and those imported via connections. The BlueConic Global Listener assigns Sample IDs (randomly) between 1 and 100 to each profile in your database. This number is written once to the profile and is not updated or changed.
Create Segments for sample groups
To use profile sampling, create segments based on the Sample ID property. Each Sample ID value (1 to 100) represents 1% of your total profiles.
1. Select a Segment
Select an existing segment to serve as an audience for testing. In this example, we start with a 'Holiday Campaign' segment of 52.8k profiles, including customers who have consented to receive personalized emails.
To create a test audience of 50% of these profiles, click the Save button's dropdown menu and select Save as.
Rename the new segment 'Holiday Campaign - Sample A', and click OK.
To refine the segment definition click AND to choose an additional profile property for the segment definition.
2. Search for the sample ID profile property
Search for the 'sample ID' profile property, which BlueConic creates and fills automatically with a random number between 1 and 100.
For the profile property condition, select values from 1 to 50. Click Save.
The test segment contains a random set of 50% of the Holiday Campaign audience, or 26.4k profiles.
3. Create a control group
To create a control group, repeat the steps above but use the remaining Sample ID values (e.g., "Between 51 and 100"). This creates a segment of profiles that can be excluded from a campaign for comparison
Create a Control Group segment using sample ID values from 51-100.
With profile sampling, you can create test segments such as these for targeting, testing campaigns, and trying out new offers.
Use sample IDs in Lifecycles
Create a segment from a Sample Group to measure campaign performance.
For example, if you want to compare 10% of your total profile population against the remaining 90%, you can create two segments based on the Sample ID range:
One segment with Sample ID values from 0 to 90 (representing 90% of the population)
One segment with Sample ID values from 91 to 100 (representing 10%)
Example
You’re running a cross-channel Lifecycle campaign aimed at converting unknown visitors into known ones. To measure the impact of the Lifecycle compared to doing nothing:
Include xx% of profiles in the Lifecycle campaign.
Exclude profiles in the Sample Group from the Lifecycle.
Export sample groups to other systems
You can export segments based on Sample ID values to external systems. For example:
Exclude a holdout group from campaign delivery.
Deliver variant experiences to test effectiveness.
Example
Export all profiles with Sample ID 1–90 for campaign targeting. Exclude Sample ID 91–100 as a 10% holdout group.
Tip: Use BlueConic insights to compare engagement or conversion between the targeted group and the excluded holdout.
Dialogue control groups vs. sample IDs
A control group in a Dialogue is typically used only to measure the performance of a single Dialogue, or when performing an A/B test for a given dialogue.
Dialogue Control Groups are different for each Dialogue set up, while the Sample ID is set for each individual profile and keeps the same value.
A control group can still be used if a dialogue is served to a specific sample group to determine the success of the individual dialogue or for A/B testing for a dialogue.
Next steps
Use BlueConic insights to visualize uplift between test and control groups.
Add Sample ID-based segmentation to Lifecycles or external exports.
FAQ
What happens to the Sample ID during profile merging?
The first stored Sample ID is retained and is not updated after profiles are merged. Each profile stores only one Sample ID value.