Offline Segment Transfer

Prev Next

This document is for demand partners who wish to provide all over-the-air (OTA) segments to their customers. It describes the process and specifications used for downloading OTA audience segments from Triton Digital.

Overview

  • Partners who wish to use our audience segments must have a contract with Triton Digital in order to access the data.

    • You must work with a Triton Digital Partner Integration Specialist to get credentials.

  • The data is made available in AWS under this S3 bucket:

    • s3://triton-ota-segments

  • Most segments are "public", they can be made available to any buyer.
    Some segments are only provided for specific buyers, they are referred as "private" in this document.

  • Files are updated once per day at most. One should keep track of the most recent version of each downloaded file by using the date indicated in the filename. Newer versions of a file completely replace previous versions.

Taxonomy

The taxonomy files provide a list and description of all the audience segments available. If a segment is no longer mentioned in the taxonomy, it should be removed (it can still be used in existing campaign). 

Taxonomy File Locations

  • Public segments are under:

    • s3://triton-ota-segments/PUB/{COUNTRY_CODE}/taxonomy/

  • Private segments are under:

    • s3://triton-ota-segments/{DSP_CODE}/{SEAT_ID}/taxonomy/

  • COUNTRY_CODE is the country code for public segments

  • DSP_CODE is the code assigned to each DSP, only that DSP has access to its folder (taxonomy of private segments)

  • SEAT_ID indicates the user account ID under that DSP which is allowed to target private segments in this folder

Example:
The URL for the taxonomy of public segments in the US is:
s3://triton-ota-segments/PUB/US/taxonomy/

The URL for the taxonomy of private segments for buyer with seat 12345 on DSP ABC is:
s3://triton-ota-segments/ABC/12345/taxonomy/

Taxonomy File Format

Filename

  • The naming convention for a taxonomy file is as follows:

    • {YYYYMMDD}.taxonomy.{VOLUME_NUMBER}.tsv.gz

  • YYYYMMDD is the date when the file was uploaded

  • VOLUME_NUMBER is a number allowing to split a large file into many smaller ones

    • You need to fetch and concatenate all the volumes to load all the segments

Example:
20220803.taxonomy.000.tsv.gz

Content

The taxonomy file is a tab-separated formatted file which includes the following columns in the following order:

Segment ID

  • Mandatory

  • Data type: string

Segment Name

  • Mandatory

Price

  • Mandatory

  • In USD

  • Data type: double. Decimals are optional; 2 should be interpreted as 2.00

  • Can be 0 (free to use)

Company

  • Optional

Segment Category

  • Optional

Status

  • Mandatory

  • Active or Inactive

Empty fields appear as two consecutive tabs.

Lines starting with a hash sign (#) can be ignored.

Example:

Segment Data

Each segment is provided in a separate file (a segment file only has data for one audience segment).

Segment File Locations

  • Public segments are under

    • s3://triton-ota-segments/PUB/{COUNTRY_CODE}/segments/

  • Private segments are under

    • s3://triton-ota-segments/{DSP_CODE}/{SEAT_ID}/segments/

  • COUNTRY_CODE, DSP_CODE and SEAT_ID have the same meaning as for taxonomy file locations

Example:
The URL for the data of public segments in the US is:
s3://triton-ota-segments/PUB/US/segments/

The URL for the data of private segments for buyer with seat 12345 on DSP ABC is:
s3://triton-ota-segments/ABC/12345/segments/

Segment File Format

Filename

  • The naming convention for a segment file is as follows:

    • {YYYYMMDD}.{SEGMENT_ID}.{VOLUME_NUMBER}.tsv.gz

  • YYYYMMDD and VOLUME_NUMBER have the same meaning as for taxonomy file names

  • SEGMENT_ID indicates the ID of segment this file contains

Example:
20220803.3918.000.tsv.gz

Content

The segment file is a tab-separated formatted file which includes the following columns in the following order:

Content ID

  • Mandatory

  • Data type: string

  • Should be matched with the ota.content.id parameter of the Bid Request

Demographic group

  • Optional

  • Data type: string, format: <GENDER><AGE_FROM>(+|-<AGE_TO>), where

    • GENDER is 'M' for men or 'W' for women

    • AGE_FROM and AGE_TO are 2 integers indicating an age range. AGE_TO is optional when AGE_FROM ends with a ‘+’ sign

AQH

  • Mandatory

  • Data type: integer

Empty fields appear as two consecutive tabs.

Lines starting with a hash sign (#) can be ignored.

Example:

Data Processing Details

  • For each segment, there can be multiple entries for the same content ID: each entry is for a different, non-overlapping demographic group.

  • Since demographic groups are not overlapping, it is possible to add the AQH values from multiple groups to get the AQH of a larger group.
    For instance:

    • the AQH for "W25-34" can be added to the AQH for "W35-44" to get the AQH of the "W25-44" group;

    • the AQH for "W25-34" can be added to the AQH for "M25-34" to get the AQH of the "25-34" group (men and women).

Feedback

Send any comments or corrections to docs@tritondigital.com. (This is a NO REPLY email address used only for documentation feedback. For customer support, go to the Triton Digital Support Center.)