Use scoring data

Last update: 2025-02-12
  • Topics:
  • Models
    View more on this topic
  • Created for:
  • User
    Admin

As part of scoring a model, scoring data is persisted within a dataset in Experience Platform. When you have enabled muti-touch attribution during model creation, additional event score data are persisted within a dataset in Experience Platform.

Each of these datasets conforms to a schema. This article documents these schemas.

Aggregate scoring data schema

The schema for scoring data is named like AMM AI Schema - <name of model> <id>. For example: AMM AI Schema - Model for Online Conversion 10120.

The dataset, persisting the scoring data for a model, is named like AMM AI Aggregrate Scores - <id>, for example AMM AI Aggregrate Scores - 10120.

The schema includes a field group with an object containing details about the scores. The object consists of the following fields.

Field Name Type Definition
campaignGroup String Name of the campaign group.
campaignName String Name of the campaign.
contribution Double Contribution attributed to this conversion for the given touchpoint.
conversionEndDate Date End date of the conversion window.
conversionName String Name of the conversion that was created during the conversion definition setup step.
conversionStartDate Date Start date of the conversion window.
geo String The geographic location where the conversion has occurred.
mediaChannel String Name of the channel that was used during the touchpoint setup step.
mediaSubChannel String Name of the subchannel.
revenue Double Revenue attributed to this conversion for the given touchpoint.
scoreCreatedTime DateTime Timestamp when this score record is created.
touchpointEndDate Date End date of the touchpoint window.
touchpointName String Name of the touchpoint that was created during the touchpoint definition setup step. Currently the touchpoint is defined on the media channel.
touchpointStartDate Date Start date of the touchpoint window.

Event scoring data schema

The schema for scoring data is named like Attribution AI Scores - <name of model> <id> - Schema. For example: Attribution AI Scores - Model for Online Conversion 10120 - Schema.

The dataset, persisting the scoring data for a model, is named like Attribution AI Scores - <name of model> <id>, for example Attribution AI Scores - Model for Online Conversion 10120 .

The schema includes a field group containing an object containing details about the cores. The object is named like attibution_AI_scores__<name of model> id.

The field group contains the following fields.

Field Name Type Description
conversion Object Conversion metadata columns.
    passThrough Object
        eventType String
        channel_typeAtSource String
     dataSource String Globally unique identification of a data source.
Example: Adobe Analytics
     eventSource String The source when the actual event happened.
Example: Adobe.com
     eventType String The primary event type for this time-series record.
Example: Order
     geo String The geographic location where the conversion was delivered placeContext.geo.countryCode.
Example: US
     path String
     priceTotal Double Revenue obtained through the conversion
Example: 99.9
     product String The XDM identifier of the product itself.
Example: RX 1080 ti
     productType String The display name for the product as presented to the user for this product view.
Example: Gpus
     quantity Integer Quantity purchased during the conversion.
Example: 1
     receivedTimeStamp DateTime Received timestamp of the conversion.
Example: 2020-06-09T00:01:51.000Z
     skuId String Stock keeping unit (SKU), the unique identifier for a product defined by the vendor.
Example: MJ-03-XS-Black
     timestamp DateTime Timestamp of the conversion.
Example: 2020-06-09T00:01:51.000Z
     totalDaysToConversion Integer
     totalTouchpointCount Integer
customerProfile Object Identity details of the user used to build the model.
     identity Object
          id String
          namespace String Contains the details of the user used to build the model such as id and namespace.
touchpointsDetail Object[] The list of touchpoint details leading to the conversion, ordered by touchpoint occurrence or timestamp.
     scores Object Touchpoint contribution to this conversion as score.
          algorithmicInfluenced Double Influenced score is the fraction of the conversion that each marketing touchpoint is responsible for.
          algorithmicSourced Double Incremental score is the amount of marginal impact directly caused by a marketing touchpoint.
          decayUnits Double Rule-based attribution score where touchpoints closer to the conversion receive more credit than touchpoints that are farther away in time from the conversion.
          firstTouch Double Rule-based attribution score that assigns all credits to the initial touchpoint on a conversion path.
          lastTouch Double Rule-based attribution score that assigns all credit to the touchpoint closest to the conversion.
          linear Double Rule-based attribution score that assigns equal credit to each touchpoint on a conversion path.
          uShape Double Rule-based attribution score that assigns 40% of the credit to the first touchpoint and 40% of the credit to the last touchpoint. The other touchpoints splitting the remaining 20% equally.
     touchPoint Object Touchpoint metadata.
          passThrough Object
               eventType String
          campaignGroup String
          campaignName String
          campaignTag String
          eventId String
          geo String
          mediaAction String
          mediaChannel String
          receivedTimeStamp DateTime
          timestamp DateTime
     isFirstInThePosition Integer
     lag Integer
     position String
     touchpointCountToConversion Integer
     touchpointName String Name of the touchpoint that was configured during setup.
Example: PAID_SEARCH_CLICK
conversionName String Name of the conversion that was configured during setup.
Example: Order, Lead, Visit
scoreCreatedTime DateTime
segmentation String Conversion segment such as geo segmentation, which the model is built against. When segments are absent, segmentation is same as conversionName.
Example: ORDER_US

See Schemas for more information.

On this page