Page 13 - Data Transparency White Paper_FINAL_Neat

Addressing Biases in Multicultural & Inclusive Identity Data
13

Race/Ethnic Assignments: Accuracy Rates

100%
Truthset Hashed Emails: Percentage 90%
Segment of Accurate Assignment of Race/
Ethnicity vs. Validation Sets 80%
70%
African-American 64% 60%
50%
Asian-American 68% 40%

30%
Hispanic 76%
20%
White (including Hispanic) 90% 10%

AFRICAN ASIAN HISPANIC WHITE
AMERICAN AMERICAN (INCLUDING
Source: AIMM-Truthset Q3 2020 Results. HISPANIC)

We wanted to compare Truthset’s accuracy rates with Nielsen’s while acknowledging that the validation data
sources, match keys (i.e., physical addresses vs. email address), and data vendor segments were very different.
But taken together, the two studies provide insights into the current state of accuracy in multicultural
identity data.

Two Independent Studies Provide Insights into the
Current State of Accuracy in Multicultural Identity Data in 2020

Segment Truthset Nielsen
Accuracy vs. Validation Data Accuracy vs. Validation Data

African-American 64% 67%

Asian-American 68% 71%

Hispanic 76% 73%

White/White Non-Hispanic 90% 89%

Source: AIMM-Truthset Q3 2020 Results, Nielsen third-party data validation studies from 2011–18.
Nielsen datasets used to measure coverage and accuracy rates vary by study.

When we put together Truthset’s Coverage and Accuracy ratings, we get a true sense of the “visible” and
accurate multicultural consumers available for targeting and marketing. We think these are the important
numbers to continue tracking in the future. This is the size of the pool of accurately identified multicultural
consumers in online third-party datasets today.