Facebook Scandal May Impact Healthcare, Marketing Research
For a lot of verticals, Facebook users are Ground Zero for research, predictive modeling and targeted marketing. Scientists and marketers are worried that the social network’s scandal may have fallout that impedes their access to user data that they then use for healthcare marketing and other efforts.
So reads “Fallout From Facebook Data Scandal May Hit Health Research,” published on March 22 by the New Scientist. Annabel Latham writes that the alleged privacy violations related to Facebook releasing data to Cambridge Analytica that was then used in targeting likely Donald Trump voters with tailored marketing messages — now the subject of congressional and FTC investigations — could mean Facebook users’ data will be harder for researchers and marketers to come by in the near future.
“Facebook is already extremely protective of its data,” she says. “The worry is that it could become doubly difficult for researchers to legitimately access this information in light of what has happened.
“Clearly, it’s not just researchers who use profile data to better understand people’s [behavioral] patterns,” Latham continues. “Marketing [organizations] have been profiling consumers for decades — if they know their customers, they will understand the triggers that prompt a purchase of their product, enabling them to adjust marketing messages to improve sales. It has become easier with digital marketing — people are constantly tracked online, their activities are [analyzed] using data analytics tools and personal recommendations are made.”
Should Marketers Worry About Losing Predictive Modeling Data?
Latham cites the predictive modeling that’s famously used by Amazon and Netflix, to the point that it’s common knowledge among consumers, not just Amazon and Netflix customers.
In order to create predictive models, marketers need to gather layers of data. Latham used a non-marketing example of the practice, in which researchers in the UK layered Twitter data to predict riots faster than the police.
Professors from Cardiff University write in their abstract:
“We present an end-to-end integrated event detection framework that comprises five main components: data collection, pre-processing, classification, online clustering and summarization. The integration between classification and clustering enables events to be detected, as well as related smaller-scale ‘disruptive events,’ smaller incidents that threaten social safety and security or could disrupt social order. We present an evaluation of the effectiveness of detecting events using a variety of features derived from Twitter posts, namely temporal, spatial and textual content. We evaluate our framework on a large-scale, real-world dataset from Twitter.”
Emotions Drive Marketing Conversions and Facebook Has That Data
“Information from online [behavior] can be used to predict people’s mood, emotions and personality,” writes Latham. “My own research into Intelligent Tutoring Systems uses learner interactions with software to profile personality type so it can automatically adapt tutoring to someone’s preferred style. Machine learning techniques can combine theories from psychology with new patterns found — such as Facebook ‘likes’ — to profile users.”
Personalization Depends on Finding Target Audience Data
In January 2017, Target Marketing reported on Trump’s digital campaign in “How Trump Won.” Facebook users contributed the most to the presidential campaign and Brad Parscale, the Trump campaign's digital director and co-founder of San Antonio-based Web design, online marketing and branding firm Giles-Parscale, credited Facebook and Twitter with the election victory.
In order to win, Trump’s campaign personalized messages to Facebook users, he told Wired. Target Marketing summarizes:
Each day, the Trump campaign ran “40,000 to 50,000 variants of its ads, testing how they performed in different formats, with subtitles and without, and static versus video, among other small differences. On the day of the third presidential debate in October, the team ran 175,000 variations,” Wired reports.
What do you think, marketers?
Please respond in the comments section below.