To what extent do "personalized recommendations" shape consumer tastes, and to what extent do they merely cater to existing ones?

Okay, this is a really interesting question; it hits right at the core of our digital lives today. I’ll try to explain my thoughts in plain language.

To what extent does "personalized recommendation" shape consumer taste, and to what extent is it merely catering to it?

You can think of "personalized recommendation" like a server who knows you well and is quite clever.

The first time you go, you order a Mala Xiang Guo (spicy hot pot) and eat it happily, sweating a lot. The next time you return, this server remembers and eagerly asks, “Hey bro, want the Mala Xiang Guo again today? We have some new ingredients perfect for it!”

See, this is the first, and most obvious, layer of "personalized recommendation."

Firstly, it’s absolutely “catering” to you, and it’s master-level

It’s like that server only recommending things related to Mala Xiang Guo. The algorithm writes down all your past actions—what songs you played, movies you watched, things you bought—on its little notepad.

The mechanism is simple: You like A, so I recommend more A, plus things similar to A—A+, A-.
Examples are everywhere:
- You search for "camping tent" on Taobao, and for the next month, your homepage, "Guess You Like," even ads on other apps, are flooded with sleeping bags, moisture-proof pads, camping carts. It’s catering to your known interest in "camping."
- You swipe past a few cat videos on Douyin (TikTok), and the algorithm decides you're a "cat person," then bombards you with all kinds of cute, funny, comforting cat content.

The benefit of this "catering" is efficiency—you don’t have to work hard to find things you like. But the downside is clear—it traps you in a "filter bubble," making you think the whole world shares your love of camping and cats. Your world shrinks.

But its real power lies in subtly "shaping" you while it caters

Now, let’s go back to that clever server.

He watches you order Mala Xiang Guo every visit, day after day for two weeks. One day, he says: “Bro, I see you really like intensely spicy and numbing flavors. We have a new chef whose signature dish is ‘Chongqing Spicy Chicken’—the heat is amazing, an incredible kick. Want to give it a try? A lot of guests who love Mala Xiang Guo ended up loving this dish.”

See, he didn’t recommend a mild Cantonese dish (that would likely fail), but pushed something you "might" like, yet have never tried before. This is where "shaping" begins.

The algorithm does the same, shaping your taste in two main ways:

"People similar to you" also like this: This is the algorithm’s classic move. It notices that thousands of users who love bands A, B, and C (just like you) also mostly like another obscure band D you’ve never heard of. So, the algorithm quietly slips Band D's songs into your recommended playlist. You might listen… and love it. Your musical taste just widened (or was shaped) a little.
Guiding you into the "Long Tail": This concept might sound technical but is simple. Imagine a massive bookstore. Bestsellers are prominently displayed at the entrance—the "Head." Books that sell fewer copies but still have buyers—specialized texts, niche novels—sit on shelves deep inside; this is the "Long Tail." Previously browsing a bookstore, we’d mainly see the bestsellers at the front. But algorithms are different. Knowing you're a history buff (catering), it will guide you from popular titles like "The Ming Dynasty Stories" (Head) towards "History of the Byzantine Empire" (Long Tail). Without the algorithm, you might never discover this book or develop an interest in Byzantium. Your niche interests are thus shaped and deepened by the algorithm.

The most typical examples are Spotify and Douyin (TikTok). Spotify's "Discover Weekly" playlist consistently and precisely recommends new songs you haven't heard but instantly click "like" for. Douyin can seamlessly take you from a familiar topic into an entirely new realm, keeping you scrolling for hours and sparking interest in new subjects.

Conclusion: Catering is the means, shaping is the goal

So, to answer the question, I believe:

"Catering" is the foundation and means of "shaping," while "shaping" is the ultimate goal of the algorithm (both commercially and technically).

It first uses "catering" to gain your trust, making you think, "Wow, it gets me," so you become reliant.
Once you develop this dependency and trust, it begins "shaping" your taste. It recommends new products, new content, new interests.

Why does it do this? Because pure "catering" quickly becomes boring and has limited commercial value. Through "shaping," platforms can:

Sell you more things: Not just tents, but also spicy chicken.
Keep you engaged longer: Exploring new territory is always more captivating.
Promote new content or goods needing exposure: Like pushing a new artist's songs or clearing excess inventory.

To summarize

Catering (70%): This is the part you clearly feel. Based on your history, it provides comfort and convenience, making the algorithm seem considerate.
Shaping (30%): This is the subtle part. Based on big data and the behavior of people "like you," it guides you to explore the unknown, broadening the boundaries of your taste. This is where the algorithm's real power lies.

This ratio is dynamic. The longer your "relationship" with an App, the stronger its ability to "shape" you becomes.

So, next time you find yourself obsessed with a band you'd never heard of before, or buying a garment in a completely different style, stop and think: Did I discover this myself, or is it a path the algorithm carefully paved for me?

The answer likely involves both. Our relationship with algorithms resembles a symbiotic one—mutually influencing and co-evolving.