How to Tap Social Media to Find New Insights

Researchers are finding it more and more difficult to get people to take part in studies. Meanwhile, 4.9 billion people around the world use social media and the average person spends 145 minutes per day on social media. So, the logical question is, ‘Can we tap into social media to better understand people and gain new insights?’

Not surprisingly, by taking advantage of new technology, the answer is, unequivocally, ‘Yes.’ Surprisingly, despite the hype, Generative AI isn’t the best way to do it.

People don’t always say what they mean to researchers. And, in the artificial setting of a study, they may filter their responses to say what they think is the ‘right’ thing or not to offend.

But, in the high-volume world of social media, and other forms of online commentary, people more freely express themselves in context. That creates a unique window into the human psyche, with the potential for a wealth of new insights into people’s thoughts, feelings, and behaviors.

Social media offers researchers the opportunity to harvest valuable new insights about people’s nonconscious responses to brands, topics, experiences, teams and more. By tapping into this rich source of information, researchers can gain a deeper understanding of consumer preferences, brand perceptions, and societal trends.

At this point, you may be thinking that text analysis, a.k.a. sentiment analysis, is the secret to unlocking those insights. It’s true that typed responses can be a valuable source of insights. But how people express themselves when they type is different than when they talk. And, when people leave their comments via video, or audio, the responses are longer and richer.

More importantly, videos and audios offer the potential for researchers to go beyond the words people use to decipher the subtle signals that reveal our nonconscious emotions and gain an understanding people’s nonconscious, ‘System 1’ responses.

This breakthrough holds immense potential for a wide range of applications, including:

• Understanding brands and consumer perceptions: Uncover the emotional resonance of brands, identify areas for improvement, and gain insights into customer satisfaction and deep-seated competitive differentiators.
• Gaining unfiltered feedback on products and services: Discover the hidden gems and pain points that customers experience from unboxing to repeat use, without the artificialities of traditional feedback methods.
• Decoding customer reviews: Analyze the emotional undertones of product reviews, revealing the true motivations behind customer satisfaction or dissatisfaction.
• Unveiling societal trends: Identify emerging topics, understand public sentiment, and track the evolution of social attitudes.
• Learning what’s important to passionate fans of sports teams, bands, etc.
• Saving cost and time relative to traditional fieldwork and as a pre-cursor to traditional feedback

Recently, Day One Strategy, a global strategy and insights agency that serves the healthcare market, did a project where they looked at attitudes about, and receptiveness to, vaccines. Instead of surveying people in a traditional way, Day One tapped into social media where, as you might suspect, vaccine acceptance is a hot topic.

The research looked at YouTube videos over time. By ‘scraping’ the videos to understand nonconscious responses the Day One team found that, although experts feel ‘beaten up’ by loud anti-vaccine voices, the general population’s perceptions of vaccines have improved over time.

The underlying technology deployed by Day One can be used to gather insights not only from YouTube, but from TikTok, product review sites, and other online sources. With it, researchers can understand and quantify the emotional responses of both influencers and the rest of us. Not all sites offer people the opportunity to leave video or audio feedback today, but it can be easily added, including the ability to instantly transcribe, with translation if desired, the words used into the site as if they had been typed.

Using a combination of voice tech and applied behavioral science that are increasingly used in quant, qual, and CX research, voice signals such as pitch rate, tone, etc. are mined to detect and score emotions. Other tools, such as language models, may then be used to identify nouns, for example, brand names, and the adjectives people use to describe a product or service.

The result is a fast, high-level view of the emotional landscape of whatever is being studied.

If it’s available, from profile data or other sources, researchers can use data about the speakers to create and study subgroups by demographic variables. Using those variables, video and audio ‘highlight reels’ can be created by emotion or tagged words or phrases. Those reels enable researchers to reveal the emotion behind responses and truly bring the voice of the customer into the Board Room.

The nonconscious emotional intelligence gained from social media can be combined with traditional research methods including sentiment analysis of typed responses. Generative AI summaries may also play a role in research that taps social media, although summaries of short typed individual responses aren’t usually valuable and summaries across responses are a level of abstraction away from understanding people’s emotions.

The ability to harness nonconscious emotions from social media has the potential to revolutionize market research, providing a deeper, more accurate understanding of people by revealing their emotions and motivations.

As Day One demonstrated, it’s now possible for researchers to tap into social media to quickly gain valuable new insights—to understand and quantify people’s emotional responses in the context of the social media, fan, and review sites where people say what they really think and feel. By combining this newfound insight with traditional research methods, researchers can gain a holistic perspective on the human experience.