Why Data Transparency Matters

August 27, 2021
by Santiago Solutions Group
Carlos Santiago for O’Dwyerpr

We all know the indelible mark recent social movements involving diversity, equity and inclusivity have had on the communications landscape. There is a clear need to reassess who we are speaking to, the messages we are using and the resonance these may or may not have. But what if we’re only looking at the issue from the surface? What if the data we are looking at to make these decisions is flawed in and of itself?

Many of us rely on third-party data to inform our programming, recommendations and executions—in fact, it has been the foundation of planning, targeting and buying decisions that help to measure ROI.

And we trust these quantitative measures at face value, rarely second-guessing the methodology or accuracy of such metrics, but putting our full faith and commitment behind these numbers.

But upon further inspection, we realize there’s more to this story, as we are now uncovering a credible means of evaluating multicultural identity data measurement and its providers in an effort to fuel better quality results that are reflective of audiences and their purchasing decisions.

Exposing the Issue of Data Transparency

The first step in the process is understanding what percentage of multicultural consumers’ hashed email addresses (devoid of private identifiable data) appear in third-party data and what percentage of those are accurately flagged as African American, Hispanic or Asian. For this reason, the Alliance for Inclusive and Multicultural Marketing, an arm of the Association for National Advertisers, partnered with Truth{set}, a company dedicated to measuring the accuracy of digital record-level marketing and media, to conduct a study on multicultural consumer identity accuracy and coverage.

So far, 16 third-party data providers, a number that has more than doubled since last summer, have participated in AIMM’s quarterly studies with Truth{set}. They include traditional and disruptor providers: Acxiom, AdZapier, Alliant, Bridge, Epsilon, Fluent, Infutor, LB Digital, Onemata, Speedeon, Stirista, Targetsmart, Throtle, V12, Webbula and 180 by Two Speedeon. Updated results can be found at AIMM’s Data Transparency & Quality initiative page.

The quarterly tracking study uncovered the following key issues:

  • Underrepresentation: In the data provided by Truth{set}, only half of all records were assigned any race/ethnicity (what we call coverage), leaving a wide information gap for multicultural marketers. African American and Asian consumers especially were not represented proportionally in the third-party datasets since AIMM started its tracking. In the last two quarters, AIMM has seen data providers substantially improve both African American and Asian coverage. The coverage gap has narrowed as compared to White consumers.
  • Inaccuracy: Data providers have been very accurate identifying White consumer records, doing so 90 percent of the time since the tracking started. However, a quarter to one-third of the multicultural assignments were incorrect, depending on the target segment. In fact, after the first year of the study, one-quarter of Hispanic and African American records are still classified incorrectly. For example, a Hispanic consumer may be classified as White by most data providers.
  • Visibility: Due to the dual impact of representation (coverage) and accuracy in practical terms, multicultural audiences have been much more invisible in data than White consumers. Compared to about two-thirds of White consumers, only three in ten African Americans and one in four Asian Americans were “visible” to marketers as the initiative started. Through AIMM’s spotlight on multicultural identity data, now half of African Americans are visible and one in three Asians. Unfortunately, the visibility of Hispanic consumers remains flat at half of the segment. This clearly demonstrates marketers’ need for more complete and reliable multicultural identity data sets to enable their campaigns to have the full impact of targeted and relevant messaging.

Setting a New Standard

So where do we go from here? To improve industry-wide standards and set best practices, AIMM created a dashboard of data providers, offering transparent insight into their data’s multicultural consumer identity coverage and accuracy. Since conducting the initial benchmark study, participation in the project has continued to grow as more providers recognize the need for improved practices. So far 15 data providers are on board with AIMM’s goal of improving their data transparently through third-party validation. Data providers not participating with AIMM may not be as focused on systemic improvements of multicultural identity data and continue the legacy practice of self-validating their own data. AIMM expects that more will join as the bar raises industry-wide.

However, this progress will also require the support of brands and their partners. They create the market for third-party data and ultimately act upon the insights within the data.

As a communications professional, you have a role to play as well. Follow these best practices to ensure the data you use is both accurate and fully representative of your entire potential audience set:

  1. Seek transparency in your sources, methods and other data characteristics that will help you understand its suitability.
  2. Validate multicultural assignment accuracy and representation of data before use.

You can also access more information in AIMM’s recent white paper, Addressing Biases in Multicultural & Inclusive Identity Data, which discusses research methods, initial findings, opportunities for growth and best practices for marketers and data providers.

It’s time the entire ecosystem gives attention to multicultural data quality to bring about improvements over time so that we may minimize bias, #SeeALL in advertising and media and deliver stronger ROI. Data quality is everyone’s responsibility—smart, informed buyers and open, transparent providers.

CLICK HERE TO READ THE FULL ARTICLE

Share This