Facing a Hater
By Santo - Buenos Aires
For Coca - Cola Brand Sprite
Highly Commended in category Data-Driven
Insights
Project Description
In the era of self-expression, oversharing, over-exposure, and overflow of opinions, haters comment. Haters hate. Haters hurt.
And teenagers are the most affected group, sometimes suffering from depression or even tragic endings.
Sprite, as one of Gen Z’s closest brands, got involved in the conversation by giving teens a new speech, and new tools to act against the aggression, under the concept: “It’s haters season. Stay fresh.” in campaign called #ILOVEYOUHATER
As part of Sprite’s #ILOVEYOUHATER campaign, we wanted to engage with Sprite’s teenage target audience by being relevant and creating a one-of-a-kind experience. We decided to conduct a social experiment by bringing a real hater face to face with his victims.
And teenagers are the most affected group, sometimes suffering from depression or even tragic endings.
Sprite, as one of Gen Z’s closest brands, got involved in the conversation by giving teens a new speech, and new tools to act against the aggression, under the concept: “It’s haters season. Stay fresh.” in campaign called #ILOVEYOUHATER
As part of Sprite’s #ILOVEYOUHATER campaign, we wanted to engage with Sprite’s teenage target audience by being relevant and creating a one-of-a-kind experience. We decided to conduct a social experiment by bringing a real hater face to face with his victims.
Agency Solution
Haters around the world are constantly posting hurtful comments on social media.
And teenagers are the most affected group, sometimes suffering from depression or even tragic endings.
As one of their closest brands, we got involved in the conversation by giving them a new speech, and new tools to act against the aggression.
Thanks to big data analysis, we found the biggest hater in Argentina and conducted a unique social experiment: We put him in front of 100 of his victims.
During the course of two-months, we developed and trained a custom machine-learning algorithm using 15,00 tweets from key seed users (Argentines notorious for being bullies or being bullied) and a proprietary dictionary of Argentine Spanish words and phrases
Once the training of the model was completed, we had it process more than five million tweets from over 100,000 unique users from Sprite’s Twitter community and other brands accounts with a high affinity to Sprite’s teen target. Classifying each tweet as offensive or non-offensive and differentiating the haters from the non- haters.
To better understand our audience and the current context of bullying in Argentina, we used natural language-processing techniques to determine over-arching themes used by haters to assault their victims. We identified five topics: ideology, body shaming, trolling, gender warring, and identity. Finally, we generated a score based on the frequency of each type of insult, and the topic they pertained to, achieving a final ranking of all the identified haters.
@AguanteElCofler ranked highest among our list of haters so we contacted him, invited him to participate in a lifetime experience. Once our hater was onboard, we reached out to everyone he had criticized during the pervious six months.
We filmed the encounter and distributed the three minute video across the most relevant digital media channels for our target: Youtube and social media platforms.
And teenagers are the most affected group, sometimes suffering from depression or even tragic endings.
As one of their closest brands, we got involved in the conversation by giving them a new speech, and new tools to act against the aggression.
Thanks to big data analysis, we found the biggest hater in Argentina and conducted a unique social experiment: We put him in front of 100 of his victims.
During the course of two-months, we developed and trained a custom machine-learning algorithm using 15,00 tweets from key seed users (Argentines notorious for being bullies or being bullied) and a proprietary dictionary of Argentine Spanish words and phrases
Once the training of the model was completed, we had it process more than five million tweets from over 100,000 unique users from Sprite’s Twitter community and other brands accounts with a high affinity to Sprite’s teen target. Classifying each tweet as offensive or non-offensive and differentiating the haters from the non- haters.
To better understand our audience and the current context of bullying in Argentina, we used natural language-processing techniques to determine over-arching themes used by haters to assault their victims. We identified five topics: ideology, body shaming, trolling, gender warring, and identity. Finally, we generated a score based on the frequency of each type of insult, and the topic they pertained to, achieving a final ranking of all the identified haters.
@AguanteElCofler ranked highest among our list of haters so we contacted him, invited him to participate in a lifetime experience. Once our hater was onboard, we reached out to everyone he had criticized during the pervious six months.
We filmed the encounter and distributed the three minute video across the most relevant digital media channels for our target: Youtube and social media platforms.