Twitter x EyeSee: Does adjacent content affect brand reputation?

After joining forces with Twitter to explore how adjacency to controversial or divisive social media content impacts brand performance, we bring you the key insights and outtakes from the session. Twitter’s research analysts Kelsey Capobianco and Isabel Suede shared the study findings, while EyeSee’s CTO Vuk Pašković discussed the new tech that’s changing the future of remote market research. Make sure you give the full webinar a listen!

Why is brand safety a top priority for Twitter?

Even before social distancing protocols and lockdowns, Twitter was the place to be – users are open to sharing their unfiltered and raw opinions, while also being receptive to brand content on their feeds. And while this is why Twitter has reached its popularity both for users and brands, it also raises the question – is it safe for brands to advertise on Twitter? 

“At Twitter, we are determined to make the service a safe place for everyone. To that end, we are committed to policies that lead, products that protect, and partnerships that drive industry-wide change” said Isabel Suede, pointing out the importance of ensuring both brand safety and fulfilling their purpose of serving the public conversation.

This is what nudged the Twitter team to further their understanding of how brands should think about the sensitive content in social media feeds and if it affects their brand – with the help of remote behavioral research.

Simulating feeds for perfect experimental conditions

Using a newly developed tool, EyeSee recreated over 150 controlled Twitter timelines, which allowed strategic positioning of both the sensitive content and tested ads.

“We partnered with EyeSee to better understand how adjacency to controversial or divisive content may impact brand performance. EyeSee’s ability to offer neurometrics within a closed timeline environment was key to the success of our study,” remarked Kelsey Capobianco.

The study was divided and conducted into 2 phases. Tweets from five different categories were internally sourced and categorized as Political, Sensitive News, Misinformation, Offensive, or NSFW. They were then rated based on the level of offensiveness and the negative emotions they evoke. The top 5 most controversial Tweets were then used in phase 2, during which the respondents were exposed to simulated Twitter feeds and tasked to scroll through, while their facial expressions and eye gaze were measured. After the behavioral part of the study was completed, the respondents filled out a survey of various brand metrics.

So, does adjacent content affect brand reputation?

Across the five tested categories of Tweets, brand favorability and consideration were not significantly impacted by the presence of controversial content surrounding its ads! However, driving these lower-funnel brand metrics often relies on more than one impression to make a significant impact, meaning that each impression contributes to a stronger brand lift. With this in mind, it is highly unlikely that multiple ad impressions appear next to or close to sensitive content in busy social media feeds – it’s a 1 in 2.5 billion chance to have 6 impressions adjacent to a controversial Tweet. For brands, this means that Twitter is a safe place for advertising.

In a previous study, Twitter explored whether its users wanted to hear from brands and marketers during the COVID-19 pandemic. Study findings showed that during challenging times, people look for a steady and familiar voice – and that’s exactly where an opportunity for brands to be that voice lies. These insights stayed consistent during the height and throughout the health crisis – making it clear that brands can feel confident embracing the conversation occurring on social media timelines. People on Twitter want to hear brands stand up and support key events regardless of whether they personally impact them.

How did this study inspire ground-breaking new tech?

Tech truly has the power to change research as we know it! New elements and approaches are constantly being invented and added into market research toolkits across industries – from AI, social listening to data science and behavioral methods! 

“Technology is one of the integral parts of EyeSee’s DNA”, EyeSee’s CTO Vuk Paskovic added – “there is an underlying desire to get the most and the best out of tech!” Our long-lasting partnership with Twitter has proven to be a true catalyst for growth that pushes us to develop new solutions – and this project was no different.

In this study, in particular, the interaction of content and context was essential. This meant that over 150 Twitter feeds had to be entirely simulated! Enabling full control of the stimuli content, experimental conditions, while also being time effective was crucial for the project – and this is what led to the development of a completely new solution that would enable exactly this. This innovation cut down the study setup time by 80%, and the same tool can be applied to all other social media research as well – on Facebook, Instagram, YouTube, or any other website.

Typically, there are a few ways brands can test behavior on social media – from desktop or mobile studies to blended approaches, and from existing apps to CMS platforms. But regardless of the path, the key element is recreating that digital environment and making the context look as realistic as possible to achieve natural and smooth content exploration. And 2020 made it clear – remote in-context research is here to stay for good.

Providing lifelike experiences for the respondents, while having a fully experimental design that researchers can control is by far the biggest advantage of using a behavioral contextual approach.

Want a deeper dive into the study findings? Watch the full webinar here!