Analyzing gender representation on thousands of ads at business vertical level, we see most brands still rely on gender stereotypes by associating specific gender to specific categories: women are over represented in family and personal care while men are more present in categories representing purchase power, financial independence, entertainment, reinforcing a symbolic message of gender roles while generating a lower sense of identification from consumers with the brand.

Brief

What if, for every important decision making in a campaign development marketers had tools, insights, experiments to make the best decision (business and community wise) driving diverse and inclusive advertising?

Objectives

Drive actionability by developing a data-driven approach to support decision-making towards diverse and inclusive communication.

Approach

Make available to the industry tools to 1. help diagnose brand-level representation biases 2. avoid damaging stereotypes, 3. measure business impact of diverse representation and 4. provide learnings and insights to help the industry advance towards a plural representation of people in ads today.

In order to test the impact of diverse representation, we offered marketers a free tool (Last 5 Ads) for a self assessment on how diverse their communication was. This tool translates unconscious bias into conscious ones, comparing the brand's representation of gender, ethnicity and body type to the population average, giving recommendation of who should be the main character in their next campaign for a balanced - population-like - representation of people.

This was a kick off inspiring them to test new forms of representation, in which we used experimental design to compare performance between diverse and business as usual representation. After running several studies and proving that diversity generates better business results, we translated this information into training, reports, and scientific paper to inspire the industry on this movement.

A brand's journey towards diverse and inclusive advertising may be surrounded by uncertainty and opinions. The best way to manage this transition is using data and scientific experiments that support a continuous cycle of Test and Learn:

HYPOTHESIS> TEST> LEARN> IMPLEMENT

This journey starts with a diagnostic of presence and portrayal of the characters featured in a brand's marketing communication, analyzing gender, ethnicity and body type representation. With that, the brand can consider new forms of representation to better connect with the target audience. This phase can be followed by uncertainties such as: “We have always used protagonists with the same features and we have never had problems – won’t my brand be at risk by transitioning to a more diverse advertising?”

EXPERIMENT DESIGN

With this design, we run two identical campaigns, switching only the protagonists of each ad. By measuring both ads, we can attribute any difference in results to the change in the mix of protagonists, getting insights on how the target audience connects with that communication by seeing themselves represented.

80% of brands testing diverse representation had at least 1.5 incremental Ad Recall lift vs non diverse representation and often outperform traditional representation in metrics such as preference, purchase intent and message recall.

Proving with data that better representation of people in advertising generates better campaigns and brings better business results, we were able to scale this knowledge to +700.000 advertising professionals, generating +4.400 reports and tools downloaded.

We were able to scale the learnings into an intersectional approach and developed campaigns testing gender, ethnicity and body types representation.

1 scientific whitpaper published

+16 campaigns co-created with positive results on diverse cell

9 reports and studies

15 Trainings delivered

+700.000 advertising professionals reached*

+4400 reports and tools downloaded*

+229.300 Campaign clicks*

+1.230.000 Video views*

*Meta internal data