What is the "algo", really?

End of lesson.

In all seriousness, there is a temptation to shroud the algo in layers of mystery and intrigue. But the best any social media platform can do is build an algorithm that dynamically responds to the needs of its users in real time. If an algo hardcoded interests and what topics to push and what topics to suppress, it could never be adaptive enough to survive the fast paced nature of today’s discourse.

Maybe primitive algos started out that way due to a lack of good enough machine learning and AI to properly understand the whims of the masses, but today, all major platforms have very good algorithms that use the best of machine learning and artificial intelligence to respond to user interests in real time.

Algos are simply proxies for what people are interested in. When facing poor performance, many people are tempted to believe they are shadow-banned. That they are particularly damned by the algo for whatever reason.

It’s almost never the case. It’s almost always the case that your content just wasn’t competitive in that particular day’s attention market.

Attention market is a good way to understand social media environments. Market means highly dynamic, and no day is like the one before it. And like any market environment, people who did well today may do poorly tomorrow, if they are not well-attuned to its movements.

It’s good to imagine yourself at some first-century bazaar trying to sell your product or message. People will be walking by, and your challenge is to stand in one place and somehow say something that makes people stop. You can easily imagine that the more you do this, day in and day out, the more you become attuned to what makes people stop. In the beginning you will fail miserably, as you will be soft-spoken, shy, and awkward, but slowly you begin to gain comfort and iterate on your message until finally, one person stops in front of your pulpit to listen to what you have to say.

You take notice of this and begin repeating this same message, and to your delight, you find that more and more people are starting to stop and listen. Thirty days later and you have the largest crowd in the bazaar. Bingo—you’ve achieved the mythical product market fit (PMF).

But then something annoying starts to happen. A newcomer appears in the bazaar a few booths down. At first you laugh at the poor schmuck, looking over the heads of your large crowd to pity his non-existent audience. “He’ll never catch up to me,” you think. But he simply begins to do the same thing you did early on, which is to iterate on different messages until it makes one person stop. The newcomer also pays attention to what you’re saying, and because he has less to lose, he takes your message and shifts it slightly to account for the new market, political, and economic issues. Slowly but surely, his message begins to be more competitive in the attention marketplace. And you start to lose share.

This is the overall environment of social media today. It’s a market, and every day, you must bid for people’s attention. What worked yesterday may not work today. Your job is to read the room. What are people interested in? What are the masses, and your particular niche, thinking about today? What major news event recently occurred that may affect what people are thinking about? If you can latch onto the hivemind of your audience, you will fare very well in the attention marketplace. If however you’re not well-tuned to the ongoings of your audience, your bid will be poorly received, and other people’s content will do better.

Reading the market isn’t some mystical talent, it’s honestly just a habit of consumption. You have to actually be on the platform you’re trying to conquer, every single day, as a viewer and not only as a poster. Scroll your own feed for fifteen minutes in the morning and pay attention to whatever you see repeating: which arguments keep resurfacing, which phrases people are borrowing off each other, what everyone in your niche is annoyed about this week, etc. You will start to gain a sense of what the market is currently paying for.

A study by Berger and Milkman looked at nearly 7,000 New York Times articles over three months to see which ones made the most-emailed list. What they found was that what was shared the most wasn’t necessarily the most useful or most important stuff, but rather content that produced high-arousal emotion. Things like awe, anger, and anxiety. Sadness, which is a low-arousal emotion, actually performed worse and was shared less, and in follow-up experiments they got the same result with amusement. So when you’re reading the market, you’re not only asking what the topic of the day is, you’re asking what people are generally worked up about.

So skip cursing out the algo. We all do it, and it’s fine to vent sometimes. But really, it’s people. It’s people being flighty. It’s people changing their interests.

But I’ll tell you one thing: if you can match your content to the ongoings of the world, you will do very, very well.