Sociology
The Social Life of Algorithms in Shaping Consumer Preferences
Quick fact
Algorithms on major platforms like YouTube and Amazon are not just reflecting your preferences—they actively create them by using your clicks, watch time, and purchases to feed you increasingly specific content, which in turn shapes what you will want next.
Why this is interesting
Have you ever noticed that after you watch one cat video, your feed suddenly overflows with cats? Why do your screens seem to know you better than you know yourself?
Read the full explanation
Understanding The Social Life of Algorithms in Shaping Consumer Preferences
Think of an algorithm as a salesperson who never sleeps. When you browse online, every click is a hint about what you like. The algorithm collects these hints and, based on patterns from millions of users, builds a profile of you. It then shows you items or videos that are more likely to get your attention, because your attention is valuable to the platform. This creates a loop: you see more of what you've already shown interest in, which makes you click on those items more, which reinforces the algorithm's belief that this is what you want. Over time, your preferences become narrower and more tailored, but you might not even realize it's happening because the suggestions seem so natural.
A deeper explanation
The key is that algorithms are not neutral mirrors of your desires—they are active participants in the social construction of taste. They use predictive models to anticipate what you'll like, which means they constantly test and refine their recommendations based on your behavior. This creates a feedback loop where your choices are both guided by and guide the algorithm. This loop works because of a few mechanisms. First, algorithms use collaborative filtering: they compare your behavior with that of thousands of other users to predict what you might like next. Second, they optimize for engagement, not for your actual satisfaction. So they might show you sensational or polarizing content because it keeps you clicking. Third, they embed social biases because the training data reflects existing inequalities. For example, if a beauty product is shown more to women, the algorithm may infer that women prefer it, and then show it even more to women, amplifying gendered stereotypes. The social life of algorithms, then, is that they are embedded in social contexts and, in turn, shape social contexts. They do not simply respond to consumer preferences; they actively help create them, perpetuating a cycle that can have profound impacts on culture, commerce, and even identity.