As the number of data-driven concerns — AI recognition,geolocation data in machine learning models and digital ads settings — bump into real-world concerns, marketers are faced with a growing issue. On the one hand, they are tasked with using an increasing amount of data-driven solutions to help personalize digital experiences for consumers, on the other, as they increase their reliance on these solutions, they also increase the risk of sending out potentially offensive messages.

Increasing the diversity of creative and data management teams is one step in countering potential data bias, but it's only the first. As brands encounter increasingly diverse consumer audiences with varied interests, they should look to the origins of cultural marketing — an extension of conscious consumerism in which consumers seek products and services that support cherished values and needs — for some clues. 

Diversity or Going Through the Motions? 

An early use of cultural marketing can be seen in the advertising aimed at African-American consumers during the 1950s and 1960s. Johnson Publications, for example, launched Ebony and Jet magazines to pioneer coverage of African-American issues and interests. Founder John Johnson encouraged advertisers to feature black models when showcasing products and services — which was all too uncommon at the time.

While many good pioneering campaigns and ads came forth, some early efforts performed poorly. Some brands tried to jump on the trend by perfunctorily including vernacular and stereotypical imagery in an effort to seem relevant. It would take years for some brands to refashion questionable brand imagery, such as Quaker Oats revising the image of Aunt Jemina syrup in 1989 from the antebellum era mammy archetype to a working black woman. But brands got away with issues like these in part because they remained limited to readers and their immediate circle — a contrast to the social media era.

Social media has cut the response time dramatically, leaving brand managers scrambling to respond when consumers quickly discover bad news and share it even faster. For example, critics and consumers denounced a 2017 digital ad for Dove Deep Moisture body wash that implied dark skin was undesirable. Dove apologized on its Facebook page, stating "We missed the mark in thoughtfully representing women of color."

Advertising towards the African-American middle class has evolved into more nuanced messages over the years. The population grew, as did the number of interest groups within it. It also raised the number of opportunities to display true-to-life experiences that could naturally connect to customers.

Preferences of African-American consumers, as well as other consumer segments, have become more precisely measurable through the adoption of smartphones and mobile retail. The result is data associated with retail behaviors. The availability of this data has sparked debates among marketers and industry analysts on the consumer response to controversial marketing messages, such as the sales and financial impact to Nike from its ad featuring Colin Kaepernick. 

What Marketers Can Do

To better plan messages, brand managers must develop a platform of inclusion that blends diversity research from agencies, advertising media assessments and data usage. This approach uses researched insights rather than choices that reapply stereotypes. Nielsen, Advertising Age and eMarketer have conducted studies noting how shifts in buying power and consumer trends among ethnic consumers have revealed new segments and opportunities. 

Learning Opportunities

Social history can also improve how campaign imagery and messages are framed as a hypothesis — a starting point used in advanced analytics and machine learning models. A hypothesis is essentially a question you want to answer statistically speaking. A question framed with cultural marketing in mind may answer how well a given message might be received by consumers with shared interests. 

For example, a Type I (false negative) error — claiming a condition exists when it does not — can imply a campaign message that overlooks a social norm can be captured as a statistical outcome. By enhancing the quality of a hypothesis, teams can help ensure programmatic campaigns and associated media are vetted for multiple audiences.

This approach also raises the question of predictive model assumptions. Many of the current predictive ad technologies are designed to act on a particular word usage or action on an app, but neither words nor actions completely encapsulate social norms. Thus, programmatic marketing can issue media correctly while still producing an undesirable message.

Facebook, for example, revised its digital ad filter options after the US Department of Housing and Urban Development claimed it encouraged housing discrimination by skipping over prospective home buyers and renters by race. Facebook ad campaign settings were blamed, yet this is just another instance of the complexity that arises when programmatic tech bumps against historic social concerns.  

The Risks of Careless Branding

Brands can ill afford to lose consumer attention. Nielsen noted that the spending power among African-American consumers reaches $1.2 trillion annually across a number of consumer goods, so consumer brands are reevaluating how they can fight for this market. A terrific example is Proctor & Gamble’s purchase of a Black-owned health and beauty startup Walker & Co., which makes shave kits, lotions and grooming products aimed at people of color.

Good branding is meant to stir positive feelings for customers. But consumers turn against brands that issue offensive messages carelessly. Brands that show genuine appreciation for diversity will always succeed, gaining good will among consumers that can last a long time.

fa-solid fa-hand-paper Learn how you can join our contributor community.