How Spotify Uses Big Data, AI, and Machine Learning to Drive Business Success
Spotify marked as the largest on-demand music service in the world that overcomes its peers such as Apple Music, Youtube Music, and Deezer. Spotify itself has a history of pushing technological boundaries and using big data, artificial intelligence, and machine learning to drive success.
This article explains why Spotify called one of the data-driven companies. When people all over the world are listening to music every minute of the day, an extraordinary amount of information needed to know what songs get the most play time, where the listeners are tuning in from, and what devices they are using to access the application. Thereby, Spotify uses big data in every part of the organization to drive decisions. Also, they use the information to train the algorithms and machines to listen to music and extrapolate insights that impact their business and the experience of listeners. One of the famous features on how Spotify generates the data is the "Discover Weekly" section. Every user gets a personalized playlist every week of music that they have not heard before on the service but will be something they expected to enjoy. Another example is Spotify for Artist app that launched in 2017. It provides mobile access to analytics where a musician can post their own playlist to promote their work, then, they can get the whole information which playlists are generating new fans to how many streams they are getting overall.
On the other hand, Spotify had acquired some technology firms to enhance services. For example, the acquisitions of Niland company where Spotify will use the API-based product and machine learning to provide its users with better search and recommendations to help them discover music they will like.
So, from this, we know that presence and innovation of big data, AI, and machine learning are important for an on-demand service company to run their business.
Kelly 陳莉時
D0731529
Most of the time, I used Spotify to listen to music. On the front page of Spotify, we can see there are some options, one of them is Discover which provide playlist made just for you, top recommendation for you, suggested for you based on some artist and etc. This means that the big data that used by the company already analyze my behavior while listening to some music. Giving out some music recommendations that might be interesting for some user to listen to. Spotify also use the information to train and predict the algorithms in order to build some experience to the customers and attract more and more to subscribe on their application.
ReplyDeleteYouTube also use big data and AI to gives some recommendation to user to watch even though the videos already posted from the ages ago. If you read the comment section, many of them sharing their experience saying that they didn’t search for the videos but it pops up on their recommendation yet they still clicked on it and some might watch until the end of the videos. This proves us that AI really doing its job on observing their user needs. Big data also help some old videos on YouTube being watched again by some users that don’t even know if the video even exist, thus this helps the YouTubers to earns viewers and subscribers.
Here, big data and AI are really helpful for most of the business even big or small to observe their customers.
Lusiana
ReplyDeleteD0731621
As someone who enjoyed listening music most of the times, this topic that Kelly chose is really interesting for me. Me myself also used Spotify before to listen to songs daily. After reading what Kelly wrote, I learned some new facts about how Spotify works. First, Spotify gather every users’ data to find which songs is currently popular and played the most across the world. Then from there they got to decide which songs to recommend their users. Next, they also use the information to train the algorithms and machines to listen to the music so they can provide better experience for their users. Another example of big data application in Spotify is their most well-known features which is the “discover weekly”. So basically, this thing allows every user to get recommendation of personalized playlist with songs that the user will most likely be interested to listen or enjoy. Lastly, Spotify also launched a Spotify for artist app in 2017 which enable musicians to post their own playlist to promote their work then they can analyze which playlist generate the most fans.
Overall, I think this blog is well summarized. Kelly did a really good job in covering up all the important points from the article. I really enjoyed reading her post because the words choices are also understandable and not too difficult so I can read it easily without getting confused. I got a lot of interesting facts about how big data also influence the music industry especially music streaming platform.
吳彩麗
ReplyDeleteJacqueline Vivi
D0732325
Comment #3
This comes from my personal experience on using Spotify music. One of the reasons that I prefer to use Spotify to other music streaming platform is because of its’ famous feature "Discover Weekly". Every user gets a personalized playlist every week of music that they have not heard before on the service but will be something they expected to enjoy. This is because Spotify uses big data and algorithm to track what kind of music genre and artist I like and follow.
I also love how Spotify uses big data to help artists grow. Spotify just launched the Spotify for Artists app that provides mobile access to analytics. Everything from which playlists are generating new fans to how many streams they are getting overall. This allows their teams to plan tours more effectively as they know what songs really attract new fans. Artists also can have more control over their presence on Spotify including selecting the “artist’s pick,” and they can update their bios and post playlists.
I also look forward to the realization of the AI-composed music. Although it is still not certain but they have already launched AI Duet earlier this year where listeners could create a duet with a computer. So I do not think that it is that impossible to make AI-composed music. All and all, I look forward to how Spotify can use big data to enhance and innovates their service more in the future. I hope that it can help more people from the listener, to the artists, and to Spotify as a company too.
I’m a Spotify user, every Monday, include me, more than 100 million Spotify users will receive a new version of the recommended playlist waiting for them. It contains 30 pieces of music that users have never heard of, but are likely to like. This feature is called Discover Weekly, and it has caused a lot of discussion.
ReplyDeleteSpotify has not actually developed a recommendation model that relies on a single algorithm-it refers to the methods used by other services, and integrates its own best strategy to build an engine called Discovery.
To create Discovery Weekly playlists, Spotify mainly uses recommendation systems:
1. Collaborative filtering model (similar to the one used by Last.fm), works by analyzing your behavior and the behavior of other users.
2. Natural Language Processing (NLP) model, which works by analyzing text.
The second recommendation model used by Spotify is the Natural Language Processing (NLP) model. As the name suggests, the data source for this model comes from metadata, news, blogs, comments, and various other texts that can be found on the Internet. The NLP model will generate the weights of these narrative words into vectors, which represent the attributes of songs, and compare similar songs at the same time.
Finally, these understandings of songs allow Spotify to analyze the similarities between different songs and push you to similar new songs that exist in the user's listening list. Combining the above two methods, the Discover Weekly playlist is formed!
Of course, these recommendation models are also linked to the entire ecosystem of Spotify, which contains a lot of data, uses a lot of Hadoop to aggregate recommendation results, and allows these models to run stably on a matrix of large amounts of data, countless network texts, and music files. Now, everyone can have their own new song list.