Lost in Recommendation

A Design in the Life #8

Considering the concept of recommendations

How did you come to know about the books you recently purchased, the clothes you recently bought, the music you recently started listening to, or the movies you recently watched? It could be through magazines, advertisements, or online media, where you saw something that caught your interest. It could also be something you happened to hear or see while walking around town. Or perhaps it was a product recommended by an influencer or celebrity on social media, or something you saw on a TV program.

Once, while on a work trip to Oita Prefecture, I had some free time and decided to look for a bookstore within walking distance from the hotel where I was staying. I found a small bookstore called Bareishoten with a café attached, using Google Maps, and decided to visit it late at night.

Bareishoten bookstore cafe in Kaneike-cho Oita-shi, Oita prefecture

I was amazed when I saw the books lined up in that bookstore. I recognized many of the books there, and it felt as if I was looking at my own bookshelf at home. Well, that may be a bit exaggerated, but the bookshelves, seemingly carefully curated by the store owner, were filled with books grouped by themes. There were books that I already owned, books I had read multiple times, books I wanted to read, and books I was contemplating whether to buy or not. It was a well-stocked bookstore where all the books I desired were available. I enjoy bookstores with a cozy and relaxed atmosphere, but it was the first time I encountered a bookstore that resembled my bookshelf.

In today’s fast-paced communication era, I feel that we rely heavily on recommendations from others to discover, purchase, or try new things. When something is recommended by a friend with similar tastes, many people trust it almost unconditionally. However, I also believe that even if our preferences are similar, there is never a perfect match. It is in those slight differences within our similar preferences that we may find opportunities for discovering new “likes.”

The role of recommendations in digital services

What exactly are recommendations in the context of digital design and digital services? Recommendations involve suggesting optimal information, products, or services to users based on their preferences and behavioral history. Online retail sites often utilize a mechanism called “recommendations” to promote sales by showing users, “People who bought this also bought that.” Some may feel that mechanical recommendations still have a long way to go compared to human recommendations. The design of user experiences and recommendations are closely intertwined.

For example, it is said that approximately 75% of the viewing content on the video streaming service Netflix is generated through automated recommendations. In the vast e-commerce site Amazon, where you can now purchase various everyday items beyond books, it is reported that about 35% of sales are triggered by recommendations.

Amazon’s Recommendations

Amazon’s recommendations may initially seem complex, but they are based on the theory of “Item to Item” connections, managing only the relationships between products. In other words, it does not treat personal preferences as data. By recommending “People who bought this item also bought…” it increases persuasiveness. This is a good method for quickly extracting recommendations while handling a large number of products. Due to this algorithm, it is not uncommon for the item you recently purchased to appear in the recommendations, which can give a perplexing impression.

Reference: Amazon.com Recommendations Item-to-Item Collaborative Filtering

Netflix’s Recommendations

Netflix does not manage user profiles based on basic information such as age or gender. Instead, it categorizes its content into over 70,000 detailed genres. The reason behind this is that even among people who enjoy horror movies, there are various preferences regardless of age or gender. Additionally, to make users feel compelled to watch certain content, Netflix displays several different descriptions for each title, rather than just one.

Recommendations in Spotify

The music streaming service Spotify recommends personalized songs based on each individual’s listening style. The challenge with music recommendations lies in the fact that even for beloved songs, the desired music can vary depending on the surrounding environment, mood, and time of day.

Unlike movie or drama recommendations, where it is rare to watch the same film or episode twice, with music, people often want to listen to their favorite songs repeatedly. Spotify takes into account various parameters such as the number of times a song has been played, skipped, saved to a device, added to a playlist, discussed on social media, shared with friends online, and whether the user follows the artist. Additionally, songs are categorized based on genres, instruments, atmosphere, lyrics, themes, and language.

For example, based on parameters such as danceability, energy, and valence (positivity), Spotify recommends songs with low danceability and valence to someone listening to melancholic music, as long as the songs fall within the listener’s preferences.

These efforts by Spotify are summarized on the website “Spotify.Design,” which provides valuable information for those involved in digital services related to recommendations. The site showcases the research, development, and design behind Spotify’s services, giving insights into the thinking process. The site also features a playful design that allows users to see the faces and hear the voices of the people behind Spotify, conveying the message that the platform is truly created by people who love music.
The documentary-style drama on Netflix called The Playlist depicts the birth of Spotify is interesting. While the characters are portrayed by actors, they bear a striking resemblance to the actual founders. This series offers a glimpse into the challenges and passion during Spotify’s early days, which may change your impression of the platform.

Spotify Design Introducing Spotify’s design 
Official trailer for The Playlist

Reference: Music recommendation at Spotify

Other services such as YouTube for video streaming, UberEats for food delivery, TikTok for endless short videos, and Airbnb for renting someone’s home as accommodation, rely heavily on the mechanism of recommendations. We may not be consciously aware of this fact, but recommendations are at the core of many services.

When utilizing recommendations in various services, the following points are important for user experience:

Prioritize recommendations tailored to the user’s preferences

  Not prioritizing what the seller wants to sell

Clearly indicate the source of the information 

Explaining the background behind the recommendation increases its persuasiveness. 

(For example: “People who bought this item also bought…”)

Categorize recommendations 

 help users understand if recommendations align with their interests

Allow users to fine-tune the suggested content 

Since perfect recommendations are difficult, users can optimize by excluding items they don’t like or need

Update recommendations quickly and frequently 

If the recommended content remains unchanged, users may lose interest, so regular updates are necessary

Consider the direction of recommendations 

(For example, people who bought the first volume of a book are more likely to buy the second volume, but the reverse is rare)

The world you see

As a general metric for evaluating products and services, the Net Promoter Score (NPS) is often used. It asks the question, “Would you recommend this service or product to your close acquaintances? Please rate from 0 to 10.” The viewpoint of whether someone is willing to recommend a service or product, and whether they trust it, may be a more important and ultimate indicator than whether they personally like it.

According to Kazuo Ishiguro, the Nobel Prize-winning author in 2017, he focuses on the “vertical journey” to deeply understand people, rather than just traveling horizontally across regions. This means that instead of meeting people similar to oneself in similar cities, it is more important to travel to cities different from one’s own residence and get to know the people living there.

There is a word, “serendipity,” which represents unexpected discoveries, realizations, and encounters. For example, I have heard that if convenience stores only optimize and recommend products that sell well, it eliminates the potential for new discoveries and enjoyment, and overall sales may decline.

Are you surrounded by pleasant recommendations, where others recommend things without overwhelming your surroundings? In the pursuit of time efficiency, are you perhaps missing out on important white space and small details? By consciously visiting new places, gatherings, shops, and discovering new music, you may encounter something unexpected. You may also discover preferences, interests, skills, and diversity within yourself that have yet to be fully explored.

In today’s world where you can search for anything on search engines and get answers from AI chatbots, encountering something that cannot be found through recommendations may be the catalyst for important design breakthroughs.


In “A Design In The Life” series, we will provide hints on improving the resolution of the design experience from the perspectives of both designs in daily life and design in digital space. If you have a topic you would like us to cover, please let us know.

Written By

Yukio Andoh

Yukio is an UX Designer, UX Writer, Design Sprint Master. He has worked on a wide range of projects from web design, information appliances, smartphone applications, VR systems, giant stereoscopic dome theaters, digital signage, and media art. He loves movies and science fiction novels, and is buried in books in his everyday life.


Thanks for supporting Spectrum Tokyo ❤️

fest partner Recruit Co., Ltd. fest partner freee K.K.
fest partner DMM.com LLC fest partner TOYOTA Connected Corporation
fest partner Gaudiy, Inc. fest partner note,inc.
fest partner STORES, Inc. fest partner Ubie, Inc.
partners Design Matters

Spectrum Tokyoとの協業、協賛などはお問い合わせまで