the village is dead. long live the algorithm.
I’ve always been fascinated by social media algorithms.
How do they do it?
How do they keep showing me things that make me laugh, make me panic, or have other equally strong emotions? And then when I’m finally done scrolling and look up from my phone, an hour has passed in the blink of an eye, and I’ve done absolutely nothing except maybe “exercised” my thumb a little bit?
But, perhaps more importantly, what does this do to us as humans? Does it connect us with others like us, or does it isolate us from those around us?
The Purpose of the Algorithm
Social media tends to show us what we like, and also takes it one step further. It keeps showing us more of what we are spending time on. It doesn’t care why we paused on something; it just knows we did, and then continues to serve that kind of thing to us.
Why? Gosh, I need to dust off my tin foil hat again for this, I think.
Money…that green paper.
“But I’m not spending any money on the apps!” That’s what I originally thought with my sweet, innocent brain. And then I did some research on why companies are spending so much time and money perfecting their algorithms.
It turns out that we are essentially little data factories, and apps like TikTok, Instagram, and Threads have ways to turn our time into money for them.
The more time we spend on the app, the more certain they become about what we are interested in by analyzing our behavior. More time also equals more strategically placed ads, not only while you’re scrolling now, but also ads they can serve up later based on your activities.
They know everything you look at, how long you look at it, and whether you engaged with it. So, the more you use an app, the more it feeds you content that they know you’ll watch, engage with (rage bait!), or look at closely. They’re getting so good that they are able to predict what you want to see next.
I know it’s fun to find a weird video and think, “wow, I built this algorithm brick by brick to get here” (I’m thinking back to the simple days when I was on the Berries & Cream algorithm). However, now that I look back, I realize that the little dopamine hits I got from hitting “like” or finding a new, completely ridiculous rendition of berries and cream never made me feel better.
In fact, I never started feeling better until I completely deleted social media from my phone for a long while. And that was before the algorithm got so smart!
What happens when we have no support, but we DO have technology?
All this being said, my brain is circling around a question. What happens when you don’t really have a village, you don’t have the government support you need, you don’t have the systems support you need, but you DO have technology? For example, a pregnant or postpartum woman who isn’t getting the support she needs from the community, government, or medical professions (for whatever reason).
What happens when she is doing 3 AM feeds and scrolling deeper into the algorithmic hole? Will it help or will it hurt?
I want to know if the algorithm can actually be the “village” we're all looking for, or if it just mimics support to keep us clicking.
I have an estimated guess based on my own experience with postpartum depression, which happened to be around the time I was a VERY heavy social media user. I also know that I never saw the mental health side of social media, and never knew organizations like Postpartum Support International even existed.
Time to Reverse Engineer that Algorithm
I’m designing an experiment to test this out. My plan is to create “profiles” for postpartum or pregnant women with different behavioral patterns, create Instagram accounts for them, and let them scroll away. Then, I’ll check what they are being shown and evaluate it from a mental health perspective.
Here are my initial thoughts for “profiles” to create:
The Doom Scroller: Clicks only on news with negative sentiment, tragedy hashtags, and "warnings" for parents.
The Hyper Optimizer: Watches "restock" videos, 5 a.m. routines, high-end nursery tours, and sleep training ads.
The Help Seeker: Someone who feels genuinely concerned about her symptoms and is searching for mental health resources and supportive communities.
The Control (Baseline Persona): Logs in, scrolls for 5 minutes, spending approximately the same amount of time on each video without interacting.
I’m going to run this experiment for a while and see what my profiles are being served after a bit. Here’s my initial plan to run the experiment (but, as you know, tech and plans don’t always get along, so I may end up modifying this):
Create new profiles for each of my archetypes
Use Selenium to create code to “scroll” through videos based on each interest (to remove human bias and allow me to run the profiles at the same time)
I know that platforms are really looking for bots, and technically, I'll be creating them. However, I’ve done my research and know there are currently ways around this and potentially existing libraries that can help me with it. I’ll report back with what works for me.
Run this for a bit and see what each person is seeing
Compare and analyze
Let’s call it…. investigative art. Or field research. Basically, I want to know if the algorithm can actually be the 'village' we're all looking for, or if it just mimics support to keep us clicking.
What do you think will happen? Do you have any suggestions for modifications to my proposed profiles or experiment steps?