Page 5 - Data Transparency White Paper_FINAL_Neat
P. 5

Addressing Biases in Multicultural & Inclusive Identity Data

              AIMM learned from several of the providers that they don’t hear many concerns about the quality of
              multicultural data from marketers.  People don’t seem to ask questions about how identities are assigned.

              AIMM was not surprised to learn that it is challenging to accurately capture or ascribe segment classifications.
              The reality of America’s racial/cultural milieu is far more complicated today than it once was. Ethnic surnames
              or neighborhood concentrations used to provide strong indications of race/ethnicity on their own, but factors
              such as intermarriages, name changes, and migration patterns have diminished the strong predictive value of
              surnames and geographic location for determining race/ethnic origin for specific segments. As part of this
              research, AIMM has been studying identity classifications for four groups (African-American, Hispanic,
              Asian-American, and LGBTQ+, as well as finer breakouts such as language preference, acculturation and
              country of origin) that are created by third-party data providers specializing in classifying households
              and/or device IP addresses.


              Buyers of multicultural data generally have little insight into the quality of the data they are acquiring. When
              data quality is poor, the entire media-targeting enterprise is undermined. There is no standard measure of the
              relative data quality and no industry guidelines or best practices to improve it.

              AIMM certainly doesn’t understate the complexity of the issue
              or underestimate the extensive experience the providers         When data quality is poor,
              have with identity data. Nor does AIMM expect providers         the entire media-targeting
              to disclose confidential intellectual property. However,
              AIMM believes that the times call for more transparency,        enterprise is undermined.
              validation and benchmarking of underlying data sources and      There is no standard
              methods. Right now, marketers make data purchase decisions      measure of the relative
              without much knowledge of the quality of the data they are
              acquiring. They trust that it is accurate. And the companies    data quality and no
              providing data have had no measure of their relative quality    industry guidelines or best
              until relatively recently, with the advent of data quality
              measurement companies and auditing. But for the most part,      practices to improve it.
              there is still no rising tide, and all boats are adrift.

              Because identity assignment is so important, and so critical to accurate ROI measurement for multicultural
              marketers, AIMM secured the assistance of the Media Rating Council (MRC) to review current practices and help
              generate new best practices. MRC was selected because it operates as the industry’s independent standards
              and auditing body. MRC has designed a confidential review and certification of multicultural data providers. It
              has also created a confidential benchmarking process to help the industry demonstrate improvements in data
              quality. More details about how to participate in the MRC review and benchmarking efforts are included at the
              end of this document.

              To advance the goal of closing these gaps, AIMM sent a letter to all data providers that participate in the RFI,
              signed by leading CMOs and CEOs, encouraging their participation. Three data providers have stepped forward
              to cooperate in this effort. Many more are needed.
   1   2   3   4   5   6   7   8   9   10