multimedia storytelling with data and imagery

Cataloging 20,000 Social Media Posts

Approximately 40% of Instagram and Facebook posts are sponsored or suggested content rather than the accounts we actually follow.

Words and Visuals by Will Koeppen

I'll admit that I have a deep-seated aversion to most advertising. When someone says they watch the Superbowl for the ads, I think, "Gross." Banksy says it best, comparing public-space advertising that you can't avoid to "a rock someone just threw at your head" (Wall and Piece, 2006).

Although social media platforms are owned by private companies, whether they are considered public spaces seems to be a matter of confusion. Before he purchased Twitter (now rebranded "X"), Elon Musk referred to it as "the defacto public town square". And despite obvious legal issues, US state legislators — on both the left and right — have introduced legislation attempting to control how companies display speech made on their platforms, in effect trying to regulate them as public forums.

Can we avoid the rocks being thrown at our head? Sure, but getting off the apps is easier for some than others. Social media is useful. It's useful for staying in touch with our extended family, for posting life updates that old friends can find, for checking up on our exes and unrequited loves, and for seeing beautiful art and media that we may never get to experience in person. It's also useful professionally. I personally know a number of once-ardent detractors of social media who created Instagram profiles as soon as they started their own small business. As a creator of data visuals and photography, I'm in the same boat as most modern artists where a presence on social media is widely regarded as a necessity.

A Too-Brief History of Social Media Ads and Me
If you're just sitting there and consuming stuff, I mean, it's not necessarily bad, but it generally isn't associated with all the positive benefits you get from being actively engaged or building relationships. ~Mark Zuckerberg, in an interview with Joe Rogan (August 2022)

I started using Facebook in 2007. Back then, text posts were formatted to be a status update and came directly after your name. Apparently, I liked to use that to its full effect and write in the third person, so on the current platform my old, artsy posts barely make any sense at all. "[Will Koeppen] is winding down on dancing, winding up on science", "is in the land of cold and casseroles", or "is coming to the conclusion that it's a sham", etc. The network of people that you could see was limited to your school or university, there were no parents, and there were no ads.

In July 2011 I began using Tumblr for my photography (aw, Tumblr). As a platform, it attracted a wonderfully quirky, clever, and darkly humorous community. I arrived late to Tumblr, but its users had a culture of complimenting and encouraging others that made me feel welcome right away. After a few years, I had gathered a pretty solid 10K followers, but when Yahoo purchased Tumblr in 2013 they began messing with the dashboard and introduced ads with abandon. Many people, including me, cursed the changes and left their blogs behind. Six years later, Yahoo (by then purchased by Verizon) declared that they couldn't monetize Tumblr's user base and infamously sold their $1.1 billion acquisition for just $3 million.

Relative area comparison of the value of Tumblr's purchase price in 2013 (large box) to its sale price in 2019 (small box). The justification given for the massive loss was that neither Yahoo nor Verizon could figure out how to "leverage Tumblr's data" (i.e., generate revenue through ads).

How many ads we can handle in our social media feeds before we put down our phone in frustration is a personal question. In 2014 I started an Instagram account, and since then I have watched the ratio of posts to ads slowly drop from 10:1 to 3:1. Perhaps learning from Yahoo, most social platform companies today seem to know that too much advertising too early can drive away users, which is why all of them start out blissfully ad-free before turning up the dial, trying to boil frogs. Meta's Threads is the latest such example. It's currently ad-free, but not for long. This is their honeymoon period when early adopters, Twitter refugees, and tech journalists alike get to gush over the platform's user experience, basking in both its newness and its refreshing absence of sponsored posts.

Previous... Work?

It's not easy to find reliable data describing how much of our social media feeds are user-created content vs. everything else. If you try to Google the answer for yourself, you'll find the internet is frothy with anecdotes from reddit users and/or flooded with irrelevant articles from content farms describing how to advertise.

The one source I did find was from the Australian cell phone comparison website WhistleOut. In November 2022 they said they gathered 8,750 social media posts from 175 different accounts, and they paid for a survey of social media users to estimate average ad percentages on various platforms. It's unclear to me if those surveyed actually counted the ads, or if users were asked to estimate, but I think it's likely the latter. Regardless, WhistleOut reported that Facebook and Instagram feeds had an average of 20% ads with peak saturations of more than 40%. They claimed that TikTok had only 2.4% ads, but that is almost certainly an undercount as advertising on TikTok can be relatively subtle and unlabelled, such as when companies pay to boost the posts of normal users that mention their brand.

Ads aren't the only content served to us from outside our direct network. Instagram introduced both reels and suggested posts to try to compete with the experience provided by rival TikTok. In July 2022, Mark Zuckerberg told The Verge that suggested posts made up 15% of the Facebook feed and that number would increase to 30% by the end of 2023. The Verge wrote that suggested posts on Instagram were already greater than 15%, though they didn't report exact numbers. Currently, users are allowed to snooze Instagram's suggested posts for 30 days at a time, but we can't disable them completely. Facebook also shows a spectrum of self-referential ads for its marketplace as well as for people and groups it suggests you follow. And it produces posts informing you of actions that your network is taking on Facebook, which could be anything from one friend wishing another, "Happy Birthday" to an acquaintance introducing themselves to their local Buy Nothing group.

For another reference, we could also consider broadcast television — the kind that needs a cable, satellite dish, or antennae and is still popular in many American households, despite taking some hits. The constraints imposed by scheduled, broadcast television led to the enduring format of the classic, three-act sitcom, and each episode of Friends, The Simpsons, Schitt's Creek, and The Good Place are 22 minutes long (give or take a minute) to be able to accommodate ~8 minutes of advertising in a half-hour broadcast. A 2014 Nielsen's Advertising and Audiences Report lamented the 16 minutes/hour (27%) of ad content in broadcast television as "an enormous bombardment of advertising exposure." Ad saturation statistics along with any opinions about them were dropped from subsequent, public-facing reports.

With the growth of suggested content (which can be boosted through payments), social media platforms Instagram and Facebook serve less "expected" content than broadcast television. The graph was compiled with data from my personal social media feeds, which had slightly different numbers than the WhistleOut study.

Of course, streaming TV unburdens its producers in both directions: they can charge for ad-free experiences or introduce as many ads as the audience can stand without adhering to time slots. “Netflix Standard with Ads” currently serves 4-5 minutes of commercials per hour (7.5%). Those numbers seem downright paltry, but it's a new service, and I'd be surprised if they don't dial up the ads in coming years.

Invaders in the Feed

I wanted to find out, for myself, how much of my Facebook and Instagram feeds are generated from outside of my intentional network. That is, how many posts am I shown that are not from my friends, groups I've joined, or companies that I follow? To answer that, I spent 100 days from May 4 to August 11, 2023 cataloging 100 social media posts on both Facebook and Instagram every day. This was my only social media activity during the time period. For some heavy users, 100 posts may not seem like very much, but I found that the daily patterns emerged very quickly and the dataset of 20,000 posts is visualized below.

20,000 posts from my personal Instagram and Facebook feeds were observed between May 4 and August 11, 2023. Unsorted, they represent the order they were received by my eyeballs, whereas sorting gives us a bar chart from 0-100%. Gray indicates a post made by an account I intentionally follow. Shades of red indicate other posts such as an ad, suggested post, suggested groups, Facebook Marketplace, etc. Hovering over an out-of-network post will highlight all posts of a similar type. A thin white line indicates 50 posts (or 50%).

I didn't at all classify the quality of the content — a shared Wordle score is still a post from a "friend" just as much as a post announcing a new baby or someone sharing a link or a Facebook memory without additional comment.

A typical (median) day on Facebook shows me 28 ads in 100 posts. Scattered throughout were days when the content-producing algorithm felt broken, and I barely saw any ads. These outliers led to a slightly lower overall calculated average of 26% ad content. I also saw a significant dip in ads in my Facebook feed between July 16 and July 27. There may have been some type of algorithm change, but it also coincided eerily with research for this article, during which I did a number of Google searches for "Facebook ad saturation." Ads were so obviously scarce during this period that I asked my partner if I could take a 100-post scroll through her Facebook feed to check and I found 32 ads. Basically normal, if not slightly high.

On Instagram, I tested out snoozing suggested posts, and these periods are clearly visible in both the unsorted and sorted data visual. With suggested posts left on, roughly half of Instagram's content is an ad or suggested content. Snoozing suggested posts effectively reclaims a quarter of the feed for in-network posts, and in this mode Instagram typically has a regular, repeating pattern of three normal posts followed by an ad. Occasionally a post gets dropped from the pattern, but it was odd enough that it was surprising when it happened. Suggested reels cannot be snoozed, but they tailed off of my feed during this study. Maybe I didn't engage with them frequently enough.

In the unsorted view, you might notice some clumping of ads and suggested content near the top of the feeds. The first post on Facebook (and Instagram) is almost never an ad, whereas the second post almost always is. On lucky days when ads were sparse in the Facebook feed, they tend to be positioned near the top. This leads me to suspect that you might see more ads overall if you're looking at Facebook multiple times per day. Facebook friend activity posts gradually increase down the feed, perhaps because my 950 friends didn't produce enough direct content. If I had kept track, I'm sure I could have shown that Wordle scores are also pushed to the back of my feed, after I've exhausted more relevant posts.

My Worth to a Company, as a "Daily Active"

I know that social media companies need revenue to operate, and they're entitled to make a profit from their services. But what is the value of an individual person to a social media company? A number of them have led the way to show us how much they think one person seeing ads is worth. For example, Tumblr allows users to pay $40/year for "ad-free browsing" on their blogging platform. Netflix charges $102/year more than the $84 base price to avoid ads. Spotify Premium costs individuals $132/year and peppers its free version with jarringly loud audio ads to make sure that you know it's worth paying to get rid of them.

For Facebook, ad sales makes up 97.5% of its total revenue. In 2022, Facebook reported having 86,482 employees, ~2 billion "daily actives" (i.e., users), total revenue of $116 billion, and net profits of $23.2 billion. The math works out to $57 per year if every single daily active user bought in. For Instagram, with an estimated 300-500 million users and $51.4 billion in revenue, the value would be on par with Netflix, at ~$103/year.

I acknowledge these are overly simplistic calculations. Allowing app users to pay to opt out of advertising can get tricky — after all, the value of an ad is dependent on how big of an audience it can reach. Also, changes to pricing plans for active users can cause immediate backlash (just ask Netflix), whereas Meta can increase the cost of an individual ad by pennies, and no one other than an account manager will bat an eye.

Keeping It Light

There's a world of people out there who have a different relative concern about money than I do, and they don't mind spending it on content subscriptions, channel packages, blue check marks, and ad-free content. But many of us are also trending down.

I currently pay for Spotify, The New York Times ($221/year with ads, if you don't get it on sale), my local newspaper, and ad-free Netflix. My barber mentioned that he had dropped all his streaming channels except Amazon Prime. "Mostly because of the shipping when I buy stuff," he said, "but there's enough on there that I can find something to watch anyway."

My parents use Facebook a lot, and they frequently ask me why I don't post more often. Over the course of the last three months, I saw 13,058 updates from accounts that I care about, and that does have a lot of value to me. However, it's offset by getting 6,942 rocks thrown at my head. I'm going to try texting more often instead.