GuidesAPI Reference
Submit Documentation FeedbackJoin Developer CommunityLog In

Get Variation

Returns a variation where the visitor will be bucketed, without triggering an impression.

Version

SDK v3.0, v3.1

Description

Takes the same arguments and returns the same values as Activate, but without sending an impression network request. The behavior of the two methods is identical otherwise.

Use Get Variation if Activate has been called and the current variation assignment is needed for a given experiment and user. This method bypasses redundant network requests to Optimizely.

See Implement impressions for guidance on when to use each method.

Parameters

This table lists the required and optional parameters for the Python SDK.

Parameter

Type

Description

experiment_key
required

string

The key of the experiment.

user_id
required

string

The ID of the user.

attributes
optional

map

A map of custom key-value string pairs specifying attributes for the user that are used for audience targeting and results segmentation. Non-string values are only supported in the 3.0 SDK and above.

Returns

Variation key representing the variation in which the user will be bucketed. None if user is not in experiment or if experiment is not running.

Example

attributes = {
  'device': 'iPhone',
  'lifetime': 24738388,
  'is_logged_in': True,
}

variation_key = optimizely_client.get_variation('my_experiment_key', 'user_123', attributes)

Notes

Activate versus Get Variation

Use Activate or Is Feature Enabled when the visitor actually sees the experiment. Use Get Variation when you need to know which bucket a visitor is in without showing the visitor the experiment.
For example, use Get Variation in the following circumstances:

  1. To pre-bucket users (for example, in order to render a page correctly) before actually exposing them to the experiment. This way, you avoid triggering redundant impressions that might skew experiment results.
    For example, suppose you want your web server to show a visitor variation_1 but don't want the visitor to count until they open a flag that isn't visible when the variation loads, like a modal. In this case, use Get Variation in the backend to specify that your web server should respond with variation_1, and use Activate or Get Feature Enabled in the front end when the visitor sees the experiment.

  2. To align your third-party analytics and your Optimizely results.
    For example, suppose you use client-side third-party analytics. Use Get Variation to retrieve the variation, and even show it to the visitor, but only call Activate when the analytics call goes out. This way, your Optimizely results and your 3rd-party analytics stay aligned.

  3. To avoid redundant network requests. For example, if you called Activate or is Feature Enabled already, you can use Get Variation if you need the current variation assignment for a given experiment and user.

🚧

Important

Conversion events can only be attributed to experiments with previously tracked impressions. Impressions are tracked by Activate, not by Get Variation. As a general rule, Optimizely impressions are required for experiment results.

Source files

The language/platform source files containing the implementation for Python is optimizely.py.


Did this page help you?