The Culture of Recommendations Through Algorithms

Today the idea of "recommendation" sounds extremely connected to the internet, but if you think about how people discovered things before the internet, you'll see that it was mostly through recommendations by real people as well as through their own interest. It's not a complex concept, it's simple: a person likes something to the point of sharing it with a friend or with someone from his own group who likes that same topic.

It's important to distinguish between recommendation and influence (which is highly connected to advertising and power). They're not same, though it's easy to get confused. Think about the rise of television in the 40s and 50s, it was basically a radio with video, which is key to understand its power of influence since humans are highly visual creatures. When people saw something on the TV and liked it, what was happening wasn't something born out of recommendation, but out of influence, because it was impossible for the TV guys to know each person in the audience to the point of recommending things, therefore how could they recommend something to people they don't even know?

The ones creating TV shows are always trying to create a popular show, and to create a popular show, you need to influence people, and to influence people, you touch on something that most people are already aware of, or at least something they have a stake in, while bringing up new things to the table.

Even after TVs, the way something was recommended to you was through friends and groups. There wasn't any "algorithm" showing things you may like. But things began to change in the 90s when the internet started to become widespread. Suddenly, you had big groups online sharing things and talking about them. Amazon already had some sort of "filtering" that understood customer-behavior and purchase history, suggesting items you would probably buy. That was the beginning of the shift in the way people see recommendations. That was no longer influence, it was personalized to each user. Amazon became your friend showing things you may like.

But how can a company balance that with advertising? Advertising has been a shadow of influence forever. Now you had a system that knew customers almost like a friend, and advertisers knocking on your door offering money with a set of guidelines. Advertisers won. And they won because of algorithms.

Algorithms were then just a set of commands designed to solve a specific problem. But somewhere around the late 90s and early 00s, when social media came around, they noticed that they could know more about you than your friends do. That's when "recommendation" as we know today began to shine. The good part of it is that we now can discover things much more easily and much faster. It's like a super-friend in your pocket, recommending you things whenever you interact with him. Take Youtube, if you watch a video about psychology, next time you enter the site, you get 5 videos on the topic that you'll probably like and they're probably useful. The bad part is that they figured that the more time you spend on the site, the more ads they can show to you; not only that, if they showed you ads about things that you were already interested in, their business grew. That's how we got here. And the problem is that recommendation started to turn into influence again.

With that, in general, culture and entertainment are becoming (even more) a by-product of influence or money-focused recommendations. Small things have their space on the internet, generally according to the audience, that's a plus; but assuming the audience is always bigger than we expect, it's understandable why they can't become big unless algorithms make them big.