Why You're In An Echo Chamber

I’ve recently noticed that despite my best efforts, I’m fed the same content from the same people at the same cadence. It’s an echo chamber that is hard to escape. I thought about how to break out of this loop and came up with the following mental model and solution.

Engagement and Discovery

Engagement funds platforms. If sign-ups, sign-ins, views, likes, and comments are increasing, then the platform is making more money. So to make more money (as all companies want to), it’s important to increase engagement.

If I’ve made a successful new platform, engagement is easy. If for no other reason than the influx of new users day after day. Network effects are powerful, and as I onboard more and more users, engagement goes up as a byproduct.

As my network matures, though, sign ups will level off. The explosive growth slows, and with it, the natural boost in engagement I'm used to. So how do I keep the engagement growing? I invest in the engagement of my current users. I do this with an algorithm.

Engagement algorithms are simple. If you, the user, have engaged with a certain topic in the past, you are likely to engage in the future. So when a new piece of content is created on the platform that belongs to that topic, why not show it to you? You might even give it a thumbs up (or like, or heart).

Your Engagement Maxima

This selection for engagement places you, the user, in an engagement maxima. You are maximally engaged given the topics you have expressed interest in in the past.

But what happens after a few years of this? Eventually, the algorithm is your only source of topics. We have found ourselves in a positive feedback loop. The algorithm selects topics for you based on engagement. You engage in the topics. The algorithm selects based on past engagement, and so on. How would you ever see anything unrelated?

This thought experiment exposes a negative side effect of engagement algorithms: limited content novelty. Engagement algorithms are conservative with novelty. The risk of presenting you something new that does not engage you is overhead. At best you ignore it. At worst, the topic offends you or makes you uncomfortable.

The engagement maxima my algorithm has placed you in is great for my company. But you have diverse interests. In the global space of all topics, there are many out there that might resonate with you. The odds your engagement maxima is your global maxima is slim.

When you consume content, the algorithm deprives you of the opportunity to discover something new.

Your Niche

This effect is even stronger when creating content. Common advice is to "niche down." You want to become the person known for topic "X." You and your topic meld together.

Why do you think that is? My guess is that it pleases the algorithms. My algorithm has a much easier time recommending you for a topic if you don't diversify your content.

This is also good for you. You have a reputation as the person for "X." This leads to word of mouth referrals. Meanwhile, my algorithm sends you a steady stream of new referrals based on "your" topic. Soon enough, you have an audience large enough to live on.

The flywheel of only creating content in your niche and only consuming content in your engagement maxima is the source of your echo chamber. The algorithm rewards you for staying in your pigeonhole.

Breaking Out & Embracing Discovery

So what do you do if your sources of content are not as diverse as you would like them to be? There are two strategies that I know of to mix things up when you feel like you’re in a content rut.

The first is to follow true polymaths. Not people who talk about being polymaths. There are a lot of people who talk about their diverse reading habits. They might even share the mountain of books they are reading. But what you are actually searching for are folks who share a diverse set of interests across many domains. They also can't be afraid to share them.

These people are harder to spot because they don’t feed the algorithm. They aren’t known for any one thing, so you can't rely on the algorithm to surface them.

The other technique is climbing the follower tree. People follow more diverse topics than they contribute to. If you find someone who is genuinely interesting, look at who they are following, reading, and watching. Oftentimes you’ll find most people are more interesting than you think.

I'm currently working on something new to help solve content discovery. Reach out if you want early access and sign up for the newsletter below to follow along.