Create an Experiment

post

https://api.optimizely.com/v2/experiments

Create an Experiment in a Project

Recent Requests

Time	Status	User Agent
Retrieving recent requests…

Loading…

❗️
Warning
The Create an Experiments endpoint is not available in Optimizely Feature Experimentation.

Query Params

action

string

enum

Action to change the state of the experiment. publish saves and stages your experiment. If you have not started your experiment or the experiment is paused, your changes will not be visible to visitors when you publish. start makes your experiment live to all visitors who are in your targeted audience and changes the status to running. pause stops the experiment and changes the status to paused. No new visitors will see a paused experiment until you restart it. See Differences among publish, start, and pause for details.

Allowed:

Body Params

A string in JSON format that includes all the fields to create an Experiment

audience_conditions

string

Defaults to everyone

The audiences that should see this experiment. To target everyone, use the string "everyone" or omit this field. Multiple audiences can be combined with "and" or "or" using the same structure as audience conditions

campaign_id

int64

For Personalization experiences, this ID corresponds to the parent Campaign. For standalone experiments this campaign_id does not correspond to a campaign object.

changes

array of objects

Custom CSS or JavaScript that will run before all variations in the Experiment (for Experiments in Web Projects only)

changes

description

string

The description or hypothesis for an Experiment

environments

object

String identifier for the Experiment's status in each Environment based on the environment key.

feature_id

int64

The ID of a Feature to attach to the Experiment. This turns an Experiment into a Feature Test.

feature_key

string

The key for the Feature attached to the Experiment. Applies to Feature Tests only. Valid keys contain alphanumeric characters, hyphens, and underscores, and are limited to 64 characters.

feature_name

string

The feature flag name to display in the Optimizely app. Whitespaces and other non-alphanumeric characters allowed. Defaults to feature key if left empty.

holdback

integer

0 to 10000

Percent of traffic to exclude from the experiment, measured in basis points. 100 basis points = 1% traffic. For example, a value of 9900 would mean that 1% of visitors will be eligible for the experiment. This is only applicable for Web.

key

string

length ≤ 64

Unique string identifier for this Experiment within the Project. Only applicable for Full Stack and Mobile projects.

metrics

array of objects

Defaults to []

An ordered list of metrics to track for the Experiment. Required for Web, Full Stack, and Mobile Experimentation. Not applicable for Web Personalization Experiences.

metrics

multivariate_traffic_policy

string

enum

For Experiments of type multivariate, this specifies how the weights and statuses of combinations will be decided. In full_factorial mode, | combination weights are read-only, and are generated by multiplying together weights of section variations.

Allowed:

name

string

Name of the Experiment. Required for Web Experimentation. Optional for Web Personalization experiences and Full Stack experiments. Not applicable for Mobile Experiments.

page_ids

array of int64s

A list of Page IDs used in the Experiment. url_targeting or page_ids, but not both.

page_ids

project_id

int64

required

The Project the Experiment is in

results_token

string

temporary token based on experiment id, used to access data platform services from other parts of the product

schedule

object

traffic_allocation

integer

0 to 10000

Percent of traffic allocated for the experiment, measured in basis points. 100 basis points = 1% traffic. For example, a value of 5500 would mean that 55% of visitors will be eligible for the experiment. This is only applicable for Full Stack.

type

string

enum

Defaults to a/b

Indicates whether this is an a/b, multivariate, feature, or multiarmed_bandit test or an experience within a personalization campaign. Note that the default for this field is a/b. If another test type is desired, populate this field with the appropriate string (from one of the possible values).

Allowed:

url_targeting

object

variations

array of objects

A list of variations that each define an experience to show in the context of the Experiment for the purpose of comparison against each other

variations

whitelist

array of objects

A list containing the user IDs and variations of users who have been whitelisted

whitelist

Responses

Warning

201Return the created Experiment

400Invalid request body content

defaultUnexpected error