Investigating Instagram

How does the social media giant's sponsored content affect users?

During the pandemic, when I felt myself lapsing back into old disordered eating habits, I sat down to vent my frustrations through writing. After spending months jabbing the “not interested” button on posts advertising recipes for “nice cream” made with bananas or “cookie dough” made with protein powder and yogurt, I came up with the perfect name for my article: “the Anorexia Algorithm.”

What do you do when, every day, the whole Internet is urging you to be thinner? Telling you the best way to minimize the space you take up? You listen: you shrivel, conditions that are not conducive to writing a shocking expose. But, two years and 100 hours of therapy down the line, I find myself in a place to critically assess the way the tech world infiltrates women’s bodies.

On Sunday, June 26, I decided to undertake an experiment: to capture and analyze all of the sponsored content I received through Instagram on a day-to-day basis. A week later, I found myself with 131 screenshots in my photos: nearly an ad for each of the 180 minutes I spent on Instagram through those seven days.

Findings

Given Instagram’s recent revamp, which supplanted the historic notifications banner with a virtual mall, I was expecting a cascade of the clothing, skincare, and makeup products that I love. But, I was met with something different: an advertising portfolio that centered services over products and catered to skill sets I actively try to develop. 

I decided that the easiest way to understand my data would be to classify the 131 ads I viewed into broad categories that could capture multiple similar products and services. The top contender was surprising: over the week that I studied, I received 18 advertisements for graduate school. UVA Darden, Lasell University, and even Harvard Business School sought to attract potential students through promises of flexible study options. However, though I had indeed spent some time recently researching options through which to pursue a Master of Public Policy, Instagram instead met me with Public Health degrees from Brown and Behavioral Health options from William James. Did my algorithm (rather, the “variety of algorithms, classifiers, and processes” that Instagram likes to make clear that it uses) not understand what I was looking for, or was it actively trying to nudge me toward opportunities I had not previously considered?

My algorithm also pushed me toward opportunities for learning outside of formal education: a variety of seminars, “bootcamps,” and newsletters appeared through the week. Six were geared toward marketing or building a presence on social media. Three focused specifically on coding.

The Instagram algorithm seems to have found its sweet spot: maximize the consumer’s potential by finding the linkage between what they like and what they’ll be excited by. I had only previously interacted with 25 percent of the brands promoted to me. While the other 75 percent of brands highlighted were new to me, they fell under my general umbrella of interests. I got the expected ILIA and Sephora ads, but I also was exposed to new beauty services like Prose.

In some spaces, I felt more vulnerable than in others. Grad school and my product preferences were spaces I had publicly engaged in, whether it be through Google or Instagram likes. But, when I started getting ads for reproductive health advocacy (of which I encountered four) I felt that we were entering new territory. No doubt this was influenced by the June 24 overturn of Roe v. Wade in Dobbs v. Jackson Women’s Health Organization; one ad from NARAL Pro-Choice America called on readers to “Tell SCOTUS: I Dissent!” 

I must admit, as someone who is vocally and passionately pro-choice, I wasn’t mad that the pro-choice message was being pushed. But, I must wonder where the other side of the coin falls. Though the majority of America is pro-choice, the Dobbs case would not have voted in conservative favor if a strong anti-choice evangelical coalition did not exist in the United States. What ads are they getting? Social media has played a role in fostering extremism on both the far-right and the far-left, and from the U.S. January 6th Uprising to Islamic State recruitment. In fact, the National Consortium for the Study of Terrorism and Responses to Terrorism found that social media played a primary role in 56 percent of U.S. extremist radicalization processes between 2011-2016: an increase from 25 percent during the previous half-decade.

Other posts hit a little too close to existing traumas. Six of the ads I encountered, or 4 percent, centered mental health; five (3.5 percent) advertised research studies to individuals with depression, trauma, or anxiety. Though an argument could be made for sharing resources, the research studies in particular made me feel exploited. Instagram users are not blind to the fact that their activity is constantly being ground into statistics; it was painful to know that, not only was my depression clearly identifiable, but that I had been marked as someone who could be used.

I have to wonder how it is that Instagram accesses my innermost thoughts in this way. In a 2021 blog post for Instagram, Adam Mosseri shares that, when deciding how posts will be ranked within users’ feeds, the app generally takes into account metadata connected to the posts in question; metadata connected to the user; a user’s activity, including past likes; and a user’s history of interaction with other users (though ads may be an exception). However, Mosseri leaves out that Instagram collects 79 percent of users’ personal data, as uncovered in a study by cloud storage firm pCloud: more than any other app studied. The pCloud study identified 88 “invasive apps” that are enabled by their privacy policies to collect and share user data with third-parties, including TikTok, Reddit, Amazon, Slack, Snapchat, and Venmo. This means that, theoretically, activity on any of these platforms may inform the content received through peer platforms.

Mosseri maintains that Instagram’s algorithms are designed to “make the most of [users’] time.” Indeed, as I note with respect to pro-choice advocacy, there can be something validating about consuming content that you consider tried and true. However, for every consumer exists a line. Though globalization has shifted social media in recent years from a forum for connecting with close friends to a space for transnational advocacy, my line stands at the boundary of my personal vulnerabilities. Give me the advocacy: but I can’t be in your research trials!