Create New Experiment
To create a new Experiment, open the Experiment Tab in the Events section.
Overview
Experiment allows you to test different versions of feature, content, or user experience to determine what performs better. Datablit experiment platform provides:
- Variant Management: Create and manage multiple experiment variants
- Traffic Allocation: Control what percentage of users see each variant
- Real-time Results: Monitor experiment performance in real-time
- Statistical Significance: Built-in statistical analysis for reliable results
Creating an Experiment
Follow these steps to create and configure your first experiment:
Step 1: Navigate to Experiments
- Go to the Experiments tab in the Datablit Console
- Click on + Experiment to create a new experiment
Step 2: Basic Configuration
Provide the following information for your experiment:
| PROPERTY | TYPE | REQUIRED | DESCRIPTION |
|---|---|---|---|
| Experiment name | String | Yes | A string to understand the experiment purpose |
| Description | String | Yes | Description of the experiment |
| Tags | Dropdown | No | Tags to categorize experiments. You can manage tags in the company profile section |
Click Create to save your experiment
Step 3: Advanced Configuration
Experiment will get added on experiment list after its created. Click on the experiment to access the details page to configure advance settings:
| PROPERTY | TYPE | REQUIRED | DESCRIPTION |
|---|---|---|---|
| Status | Dropdown | Yes | By default, the status is set to Active. To disable the experiment, change the status to Disabled. Once disabled, all entities will receive the control variant. |
| Percentage | Number | Yes | Specify the percentage of entities (e.g., users, devices, etc.) for which the experiment should be enabled. For example, if the percentage is set to 10, the experiment will return a test variant for 10% of the entities. The remaining 90% will receive the control variant. |
| Variants | Popup | Yes | Click "Create Variant" to add a new variant. To create a variant, provide a unique name (note: 'control' is reserved and cannot be used) and assign a percentage value to it. |
| Static Allotment | Dropdown | No | To assign a specific test entity ID to a particular variant, enter the ID, choose the desired variant, and click "Assign." This ensures that the specified ID consistently receives the selected variant, regardless of the experiment’s configuration or rollout percentage. |
Understanding Experiment Assignment
Traffic allocation in Datablit experiments works on a two-level system:
Example Configuration
Let's consider an experiment with the following setup:
- Overall Percentage: 20% of total users
- Variants:
v1(60% of experiment traffic)v2(40% of experiment traffic)
Traffic Distribution
- 20% of users participate in the experiment
- 60% of these users (12% of total) receive variant
v1 - 40% of these users (8% of total) receive variant
v2
- 60% of these users (12% of total) receive variant
- 80% of users receive the control variant (no experiment)
Retrieving Experiment Variants
To determine which variant a user should see, you can retrieve the experiment assignment using either the Datablit SDK or direct API calls.
Method 1: Using Datablit SDK
The recommended approach is to use the Datablit SDK for your platform. Check out the Datablit SDK documentation and navigate to your platform's specific implementation.
Method 2: Direct API Call
curl --location 'https://api.datablit.com/api/1.0/experiment/variant?expId=01JVJT9CPXGEJVC7TDSX538Y8F&entityId=1321' \
--header 'apiKey: 01JSWDEZYM0E19ZF3RMHS75PB1'Finding Your Experiment ID
You can find your experiment ID from the Datablit Console:
- Navigate to your experiment details page
- Extract the
expIdfrom the URL
Example URL:
https://console.datablit.com/events/23as2/experiment/01K0KQZ9FW71J8JJVXG8NZ2XN3Experiment ID: 01K0KQZ9FW71J8JJVXG8NZ2XN3
API Response
The API returns the assigned variant for the specified entity:
{
"expId": "01JVJT9CPXGEJVC7TDSX538Y8F",
"entityId": "1321",
"variant": "control"
}