Third-Party Onboarding

• Using this guide, you will create the Snowflake data share tables (Taxonomy, Audiences, Trigger, and Tracking).
• Provide the Snowflake Share region to Experian. We will then share our account locator with you so that we can pull your audience data.
• Upload data to the tables that will be shared with Experian.
• Experian’s tracking table process will identify when new data has been published. Once updated, it will pull your data to create the onboarding files.
• Experian will resolve the consumer identifiers within your audience table to an Experian household or LUID (Living Unit ID).
• Experian will create the required onboarding files and evaluate the data to ensure it is valid and meets the initial file ingest requirements.
• Experian will process the files through our onboarding tools and distribute your segments to the specified end platform.

Segment Code (segmentCode)

• Each segment must have a unique segment code. The system will error if it detects two segments sharing the same segment code.

• Segment codes should be consistent. Once a segment code is assigned to an audience, please do not change it. Maintain the same segment code for the assigned audience for all ongoing updates. 

• If you retire/remove a segment from your taxonomy, try not to reuse that code.

• If a segment code appears in the metadata, it must also appear in the Data file. 

• Segment codes should be 45 characters or fewer to accommodate downstream platform distribution.

• Segment codes should not include a human-readable word (e.g. segment123, sports_soccer35, etc. will not be accepted). Segment codes should be an anonymized string. 

Segment Description (segmentDescription)

• Describe your segment. Ensure the description is appropriate and easy for the end user to understand.

Rates (cpm and percentOfMedia)

• Include your base rates for CPM (cost per thousand in dollars) and Percent of Media. If you require different rates for different end platforms, these can be customized by platform before distribution. Work with your PM to accomplish this.

Folder Path (folderPath)

• The folder path represents a hierarchy of how your data will be displayed within the end platform using parent/child categorization. This allows users to drill down to the segment name (last folder populated). Examples are below.
          Demographics|Age Range|20-25
          Auto and Trucks|Jeep|Grand Cherokee
          Purchase Data|Sports|Equipment|Baseball

• Each segment within your taxonomy must have a unique folder path. The system will error if the same folder path exists for multiple segments.

• Do not include your brand name in the folder path. Your segments will appear under your brand name at the end platform; however, it is not part of the folder path.

• Do not include spaces before/after the delimiter (pipe, comma)

• If submitting the metadata in the pipe-delimited file, ensure your folder path field is in double quotes (e.g., “Demographics|Age Range|20-25”).

• It is recommended that you maintain a consistent folder path. This ensures that downstream users can easily save, reuse, and find information.

• We recommend limiting your folders to no more than 7 to improve discoverability and viewability.

Please refer to the Column/Keywords in this section for required fields and specific requirements for field length, type, and character restrictions.

• For delimited files:
          Header = Use Column/Keywords Name (this section)
          Row = Each row represents an individual segment 
• Column/Keywords can be in any order.
• The contents of each column must match the column keyword. 
• If your field includes an embedded delimiter, the entire field must be enclosed with double quotes.
• Do not include spaces before/after the delimiter. 
• Do not include a Byte Order Mark (BOM).
• All segments should be listed in one metadata file. Multiple metadata files cannot be processed. 

Specific file naming is required for the system to process correctly. The first four name attributes of the Metadata, Data, and Trigger file must match.

Name format

AM_[clientfilename]_[version]_[YYYYMMDD]_Metadata 

Name example

AM_ABCClient_1_20241024_Metdata.psv, .csv, or .json 

• Columns = segment metadata
• Rows = each segment represents a row 

For visual purposes only. The below shows how the data is organized in Excel before converting to a delimited file (do not send data in Excel)

• segmentDescription – the A_BBFYO description has an embedded comma, so the entire description is in double quotes
• folderPath is separated by pipe – note that there are no spaces before/after the pipe separator (applies to all files)

• FolderPath – since the folder path includes an embedded pipe to separate the folders, the entire folder path is in double quotes.

Experian supports receiving the following offline and digital identifiers. 

Offline: Known Identifiers, including Personally Identifiable Information (PII) 

• Name
• Postal Address
• Email: plaintext or hashed with MD5, SHA-1 or SHA-256 

Digital: Device Identifiers

•  Mobile device ID (MAID): plaintext or hashed with MD5, SHA-1 or SHA-256 
•  Internet Protocol address (IP): plaintext or hashed with MD5, SHA-1 or SHA-256 
•  Email: hashed with MD5, SHA-1 or SHA-256 

Client ID. Please add a unique ID so we can easily exchange information if we find an error or formatting issue.

•  It must be unique to each row.
•  It cannot be duplicated across files.
•  It can be sequential.

Segment Codes

• Each unique segment code in your metadata file will have its own column in the data file.

Please refer to the Column/Keywords in this section and this section for required fields and specific requirements for field length, type, and character restrictions.

Identifier columns. Each column represents an identifier type. 

• The header includes Column/Keywords for each identifier in the file and a client ID. For the consumer identifier Column/Keywords, refer to this section.
• Column/Keywords can be in any order.
• The contents of each column must match the column keyword. 
• A file can only contain one Column/Keywords; it cannot contain two columns with the same column keyword.
• Do not include Column/Keywords for identifiers not present in the file. 
• If a consumer does not have one of the identifiers for the column keyword, leave an empty string (,, or ||).
• The Column/Keywords table outlines the field-specific requirements (e.g., character length, special character use, etc.). Data input that does not comply with the field requirements will result in an error. 
• Do not include a Byte Order Mark (BOM).

Segment code columns. The segment code columns will follow the identifier columns. 

• Each column represents a unique segment code that is in the metadata file.

• There should be only one column for each unique segment code.

• You have two options for assigning segments within the data files:
          1.  The header within the data file is the same across all data files. If the headers do not match across all data files, the process will fail.
          2. The headers will vary across the data files (identifiers and/or segments). This allows you to create data files that represent only certain segments and avoids ‘wide file’ issues. NOTE: If you create your data files using this format, an additional column in the Trigger file must be included (mergeData).

•  You will add a 1 (one) or 0 (zero) in each column that defines if that consumer belongs in a segment:
          1 – in the segment
          0 – not in the segment
• Each segment code must have consumers assigned to it. We cannot process segment codes that have no consumers assigned to it.

Each row represents a consumer with offline or digital identifiers and a client ID. 

Formatting offline data

• Only one address is allowed per row. If you have multiple addresses for the same person, each address must be in a separate row.
• If you have multiple emails for a consumer, you can include them in the same row; however, each email must be in its own column. Up to four plain-text and four hashed emails of each type can be included in a single row. Each email must appear in a separate column and correspond to the column keyword.
• If you are unsure if an email belongs with other consumer identifiers, place it in a separate row.
• If there is an embedded delimiter in the field, the field must include double quotes.
          123 Main St., Apt B within a comma delimited file should appear as: “123 Main St., Apt B”
• If there are special characters in the field, it is recommended that you use double quotes.

Formatting digital data

• Plaintext and hashed digital IDs can appear in the same file. 
• If sending both plaintext and digital IDs, ensure that the identifier appears in the correct column based on the column keyword.
• If you have multiple digital identifiers for a consumer, each row can contain one plaintext, one MD-5, one SHA-1, and one SHA-256 for MAID and IP.
• If you have multiple digital HEMs for a consumer, each row can contain 1 MD-5, 1 SHA-1, and 1 SHA-256.
• If you are unsure if a digital identifier belongs to the same consumer, place it in a separate row.

Hashing data: Please hash identifiers using lowercase and trim before hashing.

IPv6 (raw and hashed): if sending IPv6, please ensure that the IP address is normalized.

Steps for normalization:

1. Remove leading zeros
2. Remove the last four IPv6 groups
3. Two trailing colons are added

Original ipv6 address: 
2603:8001:00f0:0da0:7cd1:6410:25fa:7bf8 

Normalized/Truncated ipv6 address: (The format we need for IPv6) 

2603:8001:f0:da0:: 

The number of data files you will need to create can vary and is driven by the number of identifiers and segment codes.

Specific file naming is required for the system to process correctly. The first four name attributes of the Metadata, Data, and Trigger file must match.

Name format

AM_[clientfilename]_[version]_[YYYYMMDD]_Data_[#]

Name example

AM_ABCClient_1_20241024_Data_1.psv or .csv

Name example if the partner sends in 3 files in a comma delimited file type.

• AM_ABCClient_1_20241024_Data_1.csv
• AM_ABCClient_1_20241024_Data_2.csv
• AM_ABCClient_1_20241024_Data_3.csv

Data Provider ID (dataProviderId)

Provided by Experian.

Taxonomy Type (taxonomyType)

Syndicated for syndicated taxonomies. Syndicated taxonomies are updated monthly and sit on a public shelf.

Data Usage Type (dataUsageType) is ‘3rd’ for Third Party.

Segment Data Files (segmentDataFiles)

This is the total number of Data files accompanying your metadata and trigger file. Do not include your Trigger and Metadata file in this count.

Segment Metadata Exists (segmentMetadataExists) is Y, Yes, T, or True. 

Merge Data (mergeData) is Y if the header on your data files varies across Data Files. This will occur if a data file contains only identifiers and/or audience information for a specific number of segments. Example below:

     Data File 1 contains segment1, segment2, segment3

     Data File 2 contains segment4, segment5, segment6

If your data file will maintain the same exact header across all data files, do not include this column.

Please refer to the Metadata Column/Keywords in this section for required fields and specific requirements for field length, type, and character restrictions.

• For delimited files:
      Header = Use Column/Keywords (this section)
      Row = Only one row is required, and it will include the data mentioned in the section above.
• Column/Keywords can be in any order.
• The contents of each column must match the column keyword.
• Do not include spaces before/after the delimiter.
• Do not include a Byte Order Mark (BOM).
• If your data file headers will not be the same (e.g., segment data is only represented in one file vs. all), remember to include the mergeData column in your Trigger file and include a Y value.

Specific file naming is required for the system to process correctly. The first four name attributes of the Metadata, Data, and Trigger file must match.

Name format

AM_[clientfilename]_[version]_[YYYYMMDD]_Trigger

Name example

AM_ABCClient_1_20241024_Trigger.psv, .csv, or JSON

Please refer to the Column/Keywords in this section for required fields and specific requirements for field length, type, and character restrictions.

• Table structure
          Header = Use Column/Keywords Name (See this section.)
          Row = Each row represents an individual segment

• Column/Keywords can be in any order.

• The contents of each column must match the column keyword.

• All segments should be listed in one taxonomy table. 

EXPERIAN_TAXONOMY_SHARED

Please refer to the Metadata Column/Keywords in this section. and this section for required fields and specific requirements for field length, type, and character restrictions.

Identifier columns: Each column represents an identifier type. 

• The header includes Column/Keywords for each identifier in the file. For the consumer identifier Column/Keywords, refer to this section.

• Column/Keywords can be in any order.

• The contents of each column must match the column keyword.

• The table can only contain one Column/Keywords; it cannot contain two columns with the same column keyword.

• Do not include Column/Keywords for identifiers not present in the file.

• If a consumer does not have one of the identifiers for the column keyword, leave the cell empty.

• The Column/Keywords table outlines the field-specific requirements (e.g., character length, special character use, etc.). Data input that does not comply with the field requirements will result in an error.

Segment code columns. The segment code columns will follow the identifier columns. 

• Each column represents a unique segment code that is in the taxonomy table.

• There should be only one column for each unique segment code.

• You will add a 1 (one) or 0 (zero) in each column that defines if that consumer belongs in a segment:
          1 – in the segment
          0 – not in the segment

• Each segment code must have consumers assigned to it. We cannot process segment codes that have no consumers assigned to it.

Each row represents a consumer with offline or digital identifiers and a client ID. 

Formatting offline data:

• Only one address is allowed per row. If you have multiple addresses for the same person, each address must be in a separate row.

• If you have multiple emails for a consumer, you can include them in the same row; however, each email must be in its own column. Up to four plain text emails and four hashed emails of each type can be included in one row. Each email must appear in a separate column and correspond to the column keyword.

Formatting digital data:

• If sending both plaintext and digital IDs, ensure that the identifier appears in the correct column based on the column keyword.
• If you have multiple digital identifiers for a consumer, each row can contain one plaintext, one MD-5, one SHA-1, and one SHA-256 for MAID and IP.  
• If you have multiple digital HEMs for a consumer, each row can contain 1 MD-5, 1 SHA-1, and 1 SHA-256. 
• If you are unsure if a digital identifier belongs to the same consumer, place it in a separate row. 

Hashing data: Please hash identifiers using lowercase and trim before hashing.

IPv6 (raw and hashed): if sending IPv6, please ensure that the IP address is normalized.

Steps for normalization:

1. Remove leading zeros
2. Remove the last four IPv6 groups
3. Two trailing colons are added

Original ipv6 address: 
2603:8001:00f0:0da0:7cd1:6410:25fa:7bf8 

Normalized/Truncated ipv6 address: (The format we need for IPv6) 

2603:8001:f0:da0:: 

EXPERIAN_AUDIENCES_SHARED

Please refer to the Metadata Column/Keywords in this section for required fields and specific requirements for field length, type, and character restrictions.

• Header = Use Column/Keywords (this section) 

• Row = Only one row is required, and it will include the data mentioned in the section above.

• Column/Keywords can be in any order. 

• The contents of each column must match the column keyword.

EXPERIAN.TRIGGER.SHARED

Field type: TIMESTAMP_LTZ

Example: 2025-04-18T08:38:08.959-07.00

The date 2025-04-18T08:38:08.959-07:00 is in ISO 8601 format, which is a standardized way to represent date and time. Here's a breakdown of its components:

• 2025-04-18: The date in YYYY-MM-DD format (Year-Month-Day).

• T: The separator between the date and time.

• 08:38:08.959: The time in HH:mm:ss.SSS format (Hours:Minutes:Seconds.Milliseconds).

• -07:00: The time zone offset from UTC (UTC-7 hours).

EXPERIAN_TRACKING_SHARED