/reports/propensity | Documentation

Report detailed metrics of the propensity modelling for a given site.

This API is experimental and requires the experimental flag, it might change at any time without warning.

The user must be authenticated and have read permissions to the given site.

Request

The request object has the following fields:

Name	Type	Required	Description
`siteId`	String	No	The siteID to get the distribution for.

If no siteId is provided, report for all available sites for which the user has read access will be returned.

The response is an object with the following fields:

Name	Type	Description
`sites`	Array of Object	Array of objects with fields as below, one for each site.

Name	Type	Description
`siteId`	String	The siteID the propensity distribution is for.
`modelInfo`	Model info object	Meta data describing the last training of the propensity model.
`metrics`	Metrics object	Behavioral metrics computed for low/medium/high propensity groups.

The model info object has the following fields

Name	Type	Description
`start`	String	The start time in ISO 8601 for the event data period used to calculate the metrics.
`stop`	String	The start time in ISO 8601 for the event data period used to calculate the metrics.
`usersUsedForTraining`	Number	Number of users included in the last run of the training.
`precision`	Number	Precision between 0 and 1. A high precision score means the model did a good job in correctly predicted who the subscribers were on the test set during the last training.
`precisionLift`	Float	The ratio of additional subscribers identified through our model in comparison to the ones identified based on engagement only.
`lastTrained`	String	The time in ISO 8601 when the model was last trained.

The metrics object has the following fields

Name	Type	Description
`propensityScorePerUser`	Metric of type Range	The average propensity score in the group.
`pageviewsPerUser`	Metric of type Number	The average number of pageviews in the group.
`sessionsPerUser`	Metric of type Number	The average number of sessions in the group.
`activeTimePerUser`	Metric of type Number	The average activeTime in seconds in the group.
`pageviewsPerSessionPerUser`	Metric of type Number	The average number of pageviews per session in the group.
`audienceDistribution`	Metric of type Number	The ratio of users in each audience group.
`pageviewsPerTopUrl`	Metric of type list of Item Value Pairs	The total number of pageviews for the 5 urls with most page pageviews.
`distPerDeviceType`	Metric object	The average ratio of pageviews for each device per user in the group.
`distPerReferrerHostClass`	Metric object	The average ratio of pageviews for each device per user in the group.
`pageviewsPerTopCity`	Metric object	The total number of pageviews for the 5 cities with most page pageviews.

Each metric object of type X has the following fields

Name	Type	Description
`low`	X	The value of the metric for the group of users with low propensity.
`medium`	X	The value of the metric for the group of users with medium propensity.
`high`	X	The value of the metric for the group of users with high propensity.
`anonymous`	X	The value of the metric for the group of users who are anonymous.
`subscribers`	X	The value of the metric for the group of users who are subscribers.

The type Range has the following fields

Name	Type	Description
`min`	Number	The minimal value of the metric.
`max`	Number	The maximal value of the metric.

The type Item Value Pair has the following fields

Name	Type	Description
`item`	String	The item in question.
`value`	Number	The value for the item.

Last updated: December 17, 2021