Model: ExperimentResults

1 min

Model Parameters

ExperimentResults
- confidence_thresholdnumber<double>
  Example
  0.9
  Description
  The significance level at which you would like to declare winning and losing variations. A lower number minimizes the time needed to declare a winning or losing variation, but increases the risk that your results aren't true winners and losers.
  Format
  double
- end_timestring<date-time>
  Description
  End of the time interval (exclusive) used to calculated the results. The time is formatted in ISO 8601.
  Format
  date-time
- experiment_idinteger<int64>
  Example
  3000
  Description
  The unique identifier for the Experiment
  Format
  int64
- is_staleboolean
  Description
  When true, indicates the results were delivered from a cache that is out of date.
- metricsExperimentMetricResults[]
  Description
  The breakdown of Experiment results by metric. ExperimentMetricResults object ordering in the array is consistent with the order that metrics are attached to the Experiment in the Optimizely UI and REST API (i.e. index 0 is the primary metric, indices 1-4 are secondary metrics, indices 5+ are monitoring metrics). See here for an explantion of the impact of metric ordering on results calculations.
  - aggregatorstring<unique | count | sum | exit | bounce | ratio>
    Description
    The aggregation function for the numerator of the metric. 'unique' measures the number of unique visitors/sessions that include the specified Event. 'count' measures the total number of occurrences of Event for the scope (visitor/session). 'sum' is the sum of the 'field' value. 'exit' measures the ratio of sessions with last activation occurring on the target page to the sessions that activated the target page at least once during the session. 'bounce' measures the ratio of sessions that with first and last activation occurring on the target page to the sessions with first activation on the target page. For both 'exit' and 'bounce', the eventId must be the ID of a Page.
    Allowed
    uniquecountsumexitbounceratio
  - conclusionExperimentMetricConclusion
    - loserstring
      Description
      ID of the variation that has reached statistical significance and has performed worse than all other variations.
    - winnerstring
      Description
      ID of the variation that has reached statistical significance and has performed better than all other variations.
  - event_idinteger<int64>
    Description
    The ID for the Event to select data from. Omitted for global metrics that are not relative to a specific Event, i.e. "overall revenue"
    Format
    int64
  - event_propertiesMetricEventProperties
    Description
    The conditions used to filter the metric on specific event property values.
    - filterMetricFilter
      Description
      The set of event property conditions a conversion must match to be counted toward the metric.
      - combine_operatorstring<and | or>
        Description
        Determines whether all conditions (and) or any condition (or) must be met.
        Allowed
        andor
      - conditionsMetricFilterCondition[]
        Description
        defines the event property values and operator used to check whether the conversion meets the condition.
        namestring
        Description
        The name of the event property to match to.
        operatorstring<equal_to | not_equal_to | greater_than | greater_than_or_equal_to | less_than | less_than_or_equal_to | contains | does_not_contain>
        Description
        The type of match to be performed. equal_to and not_equal_to are available for all property types. greater_than, greater_than_or_equal_to, less_than, and less_than_or_equal_to are only available for properties with type number. contains and does_not_contain are only available for properties with type string.
        Allowed
        equal_tonot_equal_togreater_thangreater_than_or_equal_toless_thanless_than_or_equal_tocontainsdoes_not_contain
        typestring<boolean | number | string>
        Description
        The data type of the event property.
        Allowed
        booleannumberstring
        valuestring
        Description
        The value the conversion's property should match, according to the operator.
  - fieldstring<revenue | value>
    Description
    The field to aggregate for the numerator of the metric. Required when 'aggregator' = 'sum', otherwise omitted
    Allowed
    revenuevalue
  - metricsCompoundSubMetric[]
    Description
    A list of all metrics that will be used in the calculation of a ratio metric.
    - aggregatorstring<unique | count | sum>
      Description
      The aggregation function for the numerator of the metric. 'unique' measures the number of unique visitors/sessions that include the specified Event. 'count' measures the total number of occurrences of Event for the scope (visitor/session). 'sum' is the sum of the 'field' value.
      Default
      unique
      Allowed
      uniquecountsum
    - event_idinteger<int64>
      Description
      The ID for the Event to select data from.
      Format
      int64
    - event_typestring<custom | click | pageview>
      Description
      The type of this Event.
      Allowed
      customclickpageview
    - fieldstring<revenue | value>
      Description
      The field to aggregate for the numerator of the metric. Required when 'aggregator' = 'sum', otherwise omitted.
      Allowed
      revenuevalue
    - scopestring<session | visitor | event>
      Description
      Specifies how Events should be grouped. Can also be thought of as the denominator of the metric. 'session' divides by the number of sessions. "Influenced sessions", or sessions that do not contain a decision Event but carry a decision from a previous session are not included in counts for numerator or denominator. 'visitor' divides by the number of visitors. 'event' divides by the total occurrences (impressions) of the specified Event.
      Default
      visitor
      Allowed
      sessionvisitorevent
  - namestring
    Description
    The name of the Metric
  - resultsmap<string, VariantResults>
    Description
    A map of results for each variation in the Experiment keyed by variation ID. For Personalization Campaigns, the special variant 'baseline' represents visitors that have been held back from any change in experience for the Experiment
  - scopestring<session | visitor | event>
    Description
    Specifies how Events should be grouped together. Can also be thought of as the denonimator of the metric. 'session' divides by the number of sessions. "Influenced sessions", or sessions that do not contain a decision Event but carry a decision from a previous session are not included in counts for numerator or denominator. 'visitor' divides by the number of visitors. 'event' divides by the total occurrences (impressions) of the specified Event
    Allowed
    sessionvisitorevent
  - time_windowstring
    Description
    The maximum amount of time a numerator event can differ from a denominator event in order to be included in a ratio metric. This can be an integer, or an integer followed by a letter denoting the unit of time (d = days, h = hours, m = minutes). If no unit of time is supplied, hours is assumed. If no value is supplied, a time window of 48 hours is assumed.
  - winning_directionstring<increasing | decreasing>
    Description
    The winning direction of this metric
    Allowed
    increasingdecreasing
- reachExperimentVariationReach
  - baseline_countinteger<int64>
    Description
    Baseline count
    Format
    int64
  - baseline_reachnumber<double>
    Description
    Baseline reach
    Format
    double
  - total_countinteger<int64>
    Description
    Total number of visitors exposed to the Experiment
    Format
    int64
  - treatment_countinteger<int64>
    Description
    Treatment count
    Format
    int64
  - treatment_reachnumber<double>
    Description
    Treatment reach
    Format
    double
  - variationsmap<string, VariationReach>
    Description
    A map of reach for each variation keyed by variation ID
- start_timestring<date-time>
  Description
  Start of the time interval (inclusive) used to calculate the results. The time is formatted in ISO 8601.
  Format
  date-time
- stats_configStatsConfig
  Description
  Stats Engine configuration settings
  - confidence_levelnumber
  - difference_typestring<absolute | relative>
    Description
    The type of test to compare the variant to baseline
    Allowed
    absoluterelative
  - epoch_enabledboolean
    Description
    Indicates if epoch-based statistics were used