Page 7 - Data Transparency White Paper_FINAL_Neat

Addressing Biases in Multicultural & Inclusive Identity Data
7

WHAT MARKETERS CAN DO RIGHT NOW

What can marketers do now, as data quality improves? Ask the right questions of data providers and
strenuously embrace the best practices AIMM puts forward in this document. Data quality is everyone’s
responsibility. Everyone has a role to play: smart, informed buyers and open, transparent providers. Marketers
can lead and encourage data providers to participate in AIMM’s important validation study review by Media
Rating Council and help AIMM establish accuracy benchmarks and monitor improvements over time.

AIMM is not looking for perfection. AIMM is rallying the entire ecosystem for sustained attention about
multicultural data quality to bring about improvements over time. Basic identity data plays a disproportionately
important role in marketing decisions – it is the foundation of all planning, targeting, media-buying decisions,
and considerations for financial impact measurement.

CURRENT PRACTICES IN MULTICULTURAL IDENTITY ASSIGNMENT

HOW IS MULTICULTURAL IDENTITY ASSIGNED TODAY?
There are a wide variety of practices associated with assigning multicultural identity in the industry. First
name, surname, or geography and racial/ethnic concentration are the most common variables used in
identifying multicultural consumers. Philosophies and approaches differ. Some providers begin ethnic
assignments with Census block groups and ZIP+4 physical addresses. Others start with first name and move
to surname if first name fails to distinguish.

Other providers rely on self-reported data for race/ethnicity, as well as language preference and acculturation.
However, providers who make multicultural assignments with naming algorithms or geography distrust self-
reported data, stating that consumers sometimes lie about their race/ethnicity for some advantage or to
avoid unwanted profiling or targeting.

Though not perfect, the U.S. Census is still used for key benchmarks. The Census Bureau also provides
surname tables for naming conventions, as well as racial/ethnic concentrations by block groups that are
widely used. Additional government sources and immigration data are also used in the identity reconciliation
process.

Finally, digital data identity providers infer identity from site visits and online behavior. They capture people
who register on sites or browse in a particular language, which offers clues into language preference and
possibly cultural identity. However, this system is not perfect: many multicultural consumers may not visit or
register on in-language sites, so it is difficult to determine the accuracy and completeness of such methods.