spotify big data case study

• Español – América Latina
• Português – Brasil

Spotify: The future of audio. Putting data to work, one listener at a time.

About Spotify

A Google Cloud customer since 2016, Spotify is the most popular global audio streaming subscription service with 248m users, including 113m subscribers, across 79 markets. Spotify is the largest driver of revenue to the music business today.

Spotify exemplifies the new era of scaling a business. It launched a music-streaming service in late 2008, surpassed 1 million customers in early 2011, and today offers 248 million monthly active users in 79 markets access to more than 50 million songs and podcasts.

That’s technology-driven hypergrowth by anyone’s standard. Equally striking, though, is the way Spotify has continued to innovate its offering, while adhering to the enduring principles for growing and sustaining a successful business: Pay attention to the customer. Find new ways to delight them. Use your comparative advantage, doubling down on the things you are best at, and find good partners to handle other work. Focus on scaling your culture even as you scale your technology.

Those old truths may be even more urgent in the digital age. Streaming audio is a competitive business, requiring fast product development, customer understanding, and powerful tools for things like recommendation, music discovery, and connecting people. Besides helping people find new music and podcasts, Spotify helps artists connect with fans and collaborate with each other.

Google Cloud is proud to support Spotify’s increasing diversification and success. In 2016 we worked together to move 1200 online services and data processing DAGs (directed acyclic graphs) as well as 20,000 daily job executions, affecting more than 100 Spotify teams, from Spotify’s data centers to the cloud. Today, Spotify’s customers listen to billions of daily plays of music and podcasts leveraging Google Cloud’s global network.

By employing automated, developer-friendly services on Google Cloud, Spotify’s teams could focus better on its core business, while gaining access to services, like data analytics, on which it could grow.

“Google Cloud removes a lot of the operational complexity from our ecosystem. That frees up time,” said Tyson Singer, vice president of technology and platform at Spotify. “We can iterate quicker on key needs, like data insights and machine learning. Having infrastructure managed for us, with the lower-value details taken away, streamlines our ability to concentrate on what’s important to our users and give them the experiences they know and love about Spotify.”

Spotify, not surprisingly, has a very engineering-driven culture, with almost half of its staff focused on building, launching, and maintaining its products. With major research and development offices in Boston, Gothenburg, London, New York, and Stockholm, the size of its workforce matches the global scale of its business. That requires a culture of collaboration and swift execution. In the fourth quarter of 2019, Spotify reported 271 million monthly users and 124 million Premium subscribers, a record, continuing its history of global growth.

Effective data use that preserves customer privacy even as the services scale is another core part of the process. Some of that increase is from a growing user base, but even more is from effective understanding of the customer experience on Spotify. The engineering brilliance that matches data-driven insights with improved customer experiences is increasingly easier and faster on the cloud.

Robust building blocks that exist on top of core data storage, computing, and network services help take away much of the backend hassle on the way to new product creation. Spotify’s technology leaders point to the particular importance of BigQuery, the Google Cloud data analysis tool, as well as Pub/Sub, for faster software application development. Dataflow, for real-time and historical data analysis, has also been particularly useful.

Much of that data goes towards solving the tricky issue of personalization in new ways. Data privacy is at the core of Spotify’s development activities as it seeks to offer music lovers new ways to find the sounds they love and connect with artists. Podcasting, a recent groundbreaking effort, relies even more on robust discovery to discern things like topics, creators, and user interest levels.

For artists, the ability to find and connect with fans, or work on new material with other musicians, is another dimension of data-driven discovery. Artists on Spotify have access to dashboards that let them gain knowledge about their fans and other artists, which helps them make better-informed decisions about everything from where to plan their upcoming tour to when to drop their next release.

Ultimately, it is great user experiences that powers a business. In the past year alone, the number of Spotify’s premium subscribers has grown by 29 percent . The company credits growth in new markets, as well as innovative new products, for the increase.

Underlying Spotify's growth is its commitment to experimentation and innovation. Being able to go faster and to more efficiently test a wide spectrum of new features and ideas means Spotify will be able to focus its DNA of creativity and excellence on even more innovative experiences for its happy listeners.

Tell us your challenge. We're here to help.

• Reach out to our team to see how Google Cloud can help your business.

• > Big Data

How Spotify is using Big Data to enhance customer experience

• Mallika Rangaiah
• Jan 07, 2021

We exist in a world where purchasing music has become a thing of the past and streaming tracks is the new popular trend, which seems to be here to stay. This has led to streaming platforms rising like a phoenix, from Apple Music to Pandora, Songza, and of course, the well-known Spotify . 

All these music streaming platforms have been adopting data gained through user interactions as an attempt to enhance their algorithms, boost their user experiences, target potential audiences via ads and to improve their business approaches and decisions. One factor in which Spotify is a pro in is in having knowledge of its customers. The platform incorporates proprietary algorithms for comprehending the music taste of the user and to steer them towards fresh genres, songs, and artists. 

The time back when we would pay to get to download music is long gone. From Songza which incorporated a team of “music experts” who would compile the playlists as per their preferences to Pandora which manually labeled a song’s characteristics allowing users to select the tags and narrow them down to create the playlists they preferred.  And then comes Spotify, a music streaming platform that incorporates artificial intelligence, machine learning techniques as well as big data for the purpose of serving a customized and exclusive listening experience. 

This proves that Spotify is largely a company propelled by data and it adopts data in each of its functions to determine decisions. By acquiring data points, the platform is making use of that information for preparing algorithms and machines to listen to music and generate insights that can benefit their business and play a role in enhancing the experience of their customers. 

You can also take a look at How Spotify uses machine learning models 

Why Spotify makes use of Big Data

1. developing personalized content.

A crucial approach Spotify applies to adopt the data generated by their users is to use it for developing content that every user will regard to be exclusive to their unique tastes. The goal is to ensure that a satisfactory experience is provided to the users so that they become long-lasting customers. This has been achieved by adopting various Artificial Intelligence and Machine Learning algorithms. 

For instance, an integral role in Spotify’s data collection is played by the platform’s “Discover” feature, which had initially been introduced in 2012. This feature emerged as a playlist of the songs released by the beloved artists of the user but slowly went on to develop into a type of recommendation engine, which suggested a collection of tracks as the user’s playlist completed, which were aligned along the lines of the songs which the playlist contained. 

Presently the “Discover Weekly” has emerged as one of Spotify’s biggest trump cards, compiled fully through a machine learning algorithm, it generates a personalized playlist that is exclusive to the listening activity of the user. The algorithm examines the playlists of other users to determine the similarities among tracks and then adopts that data for developing a fresh playlist that aligns with the prevailing track preferences of the user. Additionally, every user has a personal “taste profile” made of microgenres which plays a role in personalizing these playlists. You can learn about the process in detail here . 

For the purpose of being able to personalize these playlists, a great deal of attention had to be paid by the platform to both the tracks which the users stream as well as how they generally interact with every track. 

For example, if a track has been played by the subscriber but skipped in less than the initial 30 seconds, it is perceived by Spotify as an un-enthusiastic reaction and the song’s information is not incorporated while computing playlists. Yet when a song has been added by the user to their library or playlist and has been listened to fully, this is perceived by the platform as a positive reaction, giving them the confirmation that the song has agreed with the taste of the user, which in turn aids the algorithm in further developing the user’s overall taste profile.

Recommended blog -  YouTube is using Artificial Intelligence  

Why is Spotify making use of Big Data?

2. Digitizing the taste of the user

The daily taste profile of the listener is also incorporated in Spotify’s playlists named “Daily Mixes” . These playlists are different from the music genres which the user normally incline towards and are generally composed of songs which the user has added to their playlists or saved, or which have been created by the artists which the user has included in their present playlists or any fresh artists or albums which the user is unfamiliar with. 

These playlists are vast and dynamic, though they may have more accustomed music compared to the “Discover Weekly” playlists, Spotify can still add a few intriguing songs which the user is unfamiliar with as an effort to make the playlist more lively. 

Another example is that of the “Release Radar” playlist. It’s a weekly playlist that incorporates various fresh releases by the artists that every user follows, which is likewise in format to the main “Discover” playlist. If the listeners follow their beloved artists on Spotify the algorithm is able to generate a precise playlist with fresh song suggestions by that artist. The algorithm can also affix some additional fresh songs, making the playlist compelling.  

3. For Enhanced Marketing through targeted ads

While enhancing the experience of customers, Spotify has also been able to adopt a humongous section of data generated through its users for the purpose of updating their ad campaigns and targeting their customers in a more compelling manner. 

This is basically carried out by the platform examining the knowledge they have gained regarding their listeners and then adopting those insights to create ads that trickily aim at the platform’s target audience. 

For instance, one display ad which first ran in Williamsburg, New York set off an extended prevailing marketing campaign for Spotify where the platform adopted used listening history for developing humorous, targeted advertisements. 

The first ad on which was written, “Sorry, Not Sorry Williamsburg, Bieber’s hit trended highest in this zip code” had been popular, engaging, and amusing among the local audience which helped Spotify measure the possible impression adopting the data of the listeners for developing such personalized advertisement campaigns could create on the platform’s user engagement and their sales. 

This experience set off the trend of strategic and well-responsive advertisements which are adopted to publicize the platform even now. A few of the well-known campaigns include a series of holiday advertisements, as well as a series of meme-inspired advertisements. In 2019, Spotify started operating a global market campaign on the basis of the listening history of its users. In this campaign, humorous and meme-like advertisements were created for ascertaining potential customers. In our prevailing world of meme-culture, there has been an enthusiastic response by Spotify’s potential and target customers. 

Recommended blog - Review based Recommendations system

4. Continuously updating its system

In early 2018, the streaming platform had stated that their free users would no longer be required to solely shuffle through music on their application. Rather, their users were now allowed the liberty of exploring 15 of the platform’s well-known playlists which included the platform’s popular “Discover Weekly” as well as “RapCaviar” .  

The primary intention behind the platform’s decision was propelled by data. The access shift allows the platform to produce the data of an additional hundred million-plus users, which is largely useful with the company focused upon advancing its suggestion algorithms to serve its users with a satisfactory personalized experience. 

As an endeavor to make their massive amount of data available for their musicians as well as their managers, the platform introduced a Spotify for Artists application in which access to analytics is provided such as which playlists have been helping to generate new users and the number of streams they are receiving overall. 

The mobile application permits the musicians to gain access to the information via their tour bus and the geographic streaming data can be useful for musicians and their teams to sketch out tours more efficiently. It also allows the artists to gain more control over their Spotify presence such as for choosing the “artist’s pick” and also for tasks like updating their bios or posting playlists. 

Recommended blog - Top 10 Machine Learning Algorithms

5. Spotify Wrapped

This is Spotify’s year-end report, a tradition that supplies each user of the platform with an excuse to share their music taste on social media without any hesitation. Spotify Wrapped’s application of data is not just about simple analytics. 

The company has a custom fit their listening platform to incorporate ingrained ego-boosters for its users. Around the completion of the year, a report will be received by the users informing them if they belong in the leading 1% of say, a band’s most loyal followers or among the leading non-mainstream song listeners. 

Through Spotify Wrapped, the platform is basically serving its users with data on a silver platter, presenting it in a manner that would intrigue and entice them. And it definitely works. The artfully presented data succeeds in making the users feel recognized and validated and sparks their enthusiasm. Through this data, the platform is developing an experience as if narrating a tale using music data instead of words. 

“Seeing top songs on Spotify Wrapped is like seeing an old best friend that you lost touch with.” - Haley Weiss in The Atlantic

This form of natural engagement becomes an integral reason why Spotify has kept Spotify Wrapped to serve as a useful weapon in their long-term marketing design. The platform’s listeners fill their social media accounts with screenshots of their Spotify profile as well as linking their playlists, to let their friends and family know where they stand on the platform. 

In the modern world, where streaming now dominates over purchased music, the industry has been compelled to steer its fixation from record sales towards accumulating such data with the goal of unraveling the impression a particular song, artist or album is creating on the public. As the data also supplies a more profound insight into listening trends, audience markets, and other such sections, it presents an unceasing revolution for the people in the industry. 

Being a social and sharing experience, and through combining its application of data with a robust user experience custom made for social media, Spotify becomes an inadvertently self-marketable platform since the users promote their engagement on their own accord. 

Share Blog :

Be a part of our Instagram community

Trending blogs

5 Factors Influencing Consumer Behavior

Elasticity of Demand and its Types

An Overview of Descriptive Analysis

What is PESTLE Analysis? Everything you need to know about it

What is Managerial Economics? Definition, Types, Nature, Principles, and Scope

5 Factors Affecting the Price Elasticity of Demand (PED)

6 Major Branches of Artificial Intelligence (AI)

Scope of Managerial Economics

Dijkstra’s Algorithm: The Shortest Path Algorithm

Different Types of Research Methods

Latest Comments

magretpaul6

I recently recovered back about 145k worth of Usdt from greedy and scam broker with the help of Mr Koven Gray a binary recovery specialist, I am very happy reaching out to him for help, he gave me some words of encouragement and told me not to worry, few weeks later I was very surprise of getting my lost fund in my account after losing all hope, he is really a blessing to this generation, and this is why I’m going to recommend him to everyone out there ready to recover back their lost of stolen asset in binary option trade. Contact him now via email at [email protected] or WhatsApp +1 218 296 6064.

• Artificial Intelligence
• Business Intelligence
• Deep Learning
• Machine Learning

Connect With Us

• Get A Quote
• Data Migration
• Data Warehouse
• Data Virtualization
• Predictive Analytics
• Text Analytics
• IoT Analytics
• Customer 360
• Video Analytics
• Qlik Sense & QlikView
• Microsoft Power BI
• Elastic Solution
• Banking and Financial
• Manufacturing
• Transportation And Logistics
• Media And Entertainment
• Travel And Hospitality
• Oil And Gas
• Real Estate And E-commerce
• Food And Beverages

How Spotify is using Big Data for better customer experience

• August 27, 2021

Harnil Oza is a CEO of HData Systems - Data Science Company & Hyperlink InfoSystem a top mobile app development company in Canada, USA, UK, and India having a team of best app developers who deliver best mobile solutions mainly on Android and iOS platform and also listed as one of the top app development companies by leading research platform.

Recent Blog Post

• Boost Your E-commerce Sales Conversions with Generative AI
• The Merits of Optimizing Clinical Trials with Generative AI
• Developing a Mobile App with Face Recognition Technology
• How Generative AI Transforming Life Sciences
• Data Science
• Big Data Analytics
• Data Mining
• Data Visualization

Relevant Blogs

Top Big Data Technologies And Solutions to Watch

• March 24, 2023

How Big Data Can Help You Do Wonders In Your Business

• March 07, 2023

Common Data Lake Challenges and Key Components

• January 20, 2023

Powered By Hyperlink InfoSystem

Hyperlink InfoSystem is one of the leading software development companies based in India and has offices in USA, UK, UAE, France, and Canada. With 10+ years of experience in the industry, Hyperlink InfoSystem served more than 2,300 clients worldwide. The company has a team of 450+ highly skilled developers who works on any custom solutions using the latest technologies.

Get In Touch With Us

One World Trade Center, 285 Fulton Street suite 8500, New York, NY 10007, United States

Our Business Team Will Get Back to You Soon.

Quick Inquiry

Thank you...

Big Data Processing at Spotify: The Road to Scio (Part 1)

This is the first part of a 2 part blog series. In this series we will talk about  Scio , a Scala API for  Apache Beam  and  Google Cloud Dataflow , and how we built the majority of our new data pipelines on Google Cloud with Scio.

Scio > Ecclesiastical Latin IPA: /ˈʃi.o/, [ˈʃiː.o], [ˈʃi.i̯o] > Verb: I can, know, understand, have knowledge.

Introduction

Over the past couple of years, Spotify has been migrating our infrastructure from on premise to Google Cloud. One key consideration was Google’s unique offerings of high quality big data products, including Dataflow, BigQuery, Bigtable, Pub/Sub and many more.

Google released Cloud Dataflow in early 2015 ( VLDB paper ), as a cloud product based on  FlumeJava  and  MillWheel , two Google internal systems for batch and streaming data processing. Dataflow introduced a unified model to batch and streaming that consolidates ideas from these previous systems, and the Google later donated the model and SDK code to the Apache Software Foundation as Apache Beam. With Beam, an end user can build a pipeline using one of the SDKs (currently Java and Python), which gets executed by a runner for one of the supported distributed systems, including Apache Apex, Apache Flink, Apache Spark and Google Cloud Dataflow.

Scio is a high level Scala API for the Beam Java SDK created by Spotify to run both batch and streaming pipelines at scale. We run Scio mainly on the Google Cloud Dataflow runner, a fully managed service, and process data stored in various systems including most Google Cloud products, HDFS, Cassandra, Elasticsearch, PostgreSQL and more. We announced Scio at  GCPNEXT16  last March and it’s been gaining traction ever since. It is now the preferred data processing framework within Spotify and has gained many external users and open source contributors.

In this first post we will take a look at the history of big data at Spotify, the Beam unified batch and streaming model, and how Scio + Beam + Dataflow compares to the other tools we’ve been using. In the second post we will look at the basics of Scio, its unique features, and some concrete use cases at Spotify.

Big Data at Spotify

At Spotify we process a lot of data for various reasons, including business reporting, music recommendation, ad serving and artist insights. We serve billions of streams in 61 different markets and add thousands of new tracks to our catalogue every day. To handle this massive inflow of data, we have a ~2500 node on-premise Apache Hadoop cluster, one of the largest deployments in Europe, that runs more than 20K jobs a day.

Spotify started as a Python shop. We created  Luigi  for both job orchestration and Python MapReduce jobs via Hadoop streaming. As we matured in data processing, we began to use a lot of  Scalding  for batch processing. Scalding is a Scala API from Twitter that runs on  Cascading , a high-level Java library for Hadoop MapReduce. It allows us to write concise pipelines with significant performance improvement over Python. The type-safe functional paradigm also boosts our confidence in code quality and correctness. Discover Weekly, one of our very popular features, is powered by Scalding ( BDS2015 talk ). We also use  Apache Spark  for some  machine learning applications , leveraging its in-memory caching capability for iterative algorithms.

On the streaming side we’ve been using  Apache Storm  for a few years now to power real time use cases like new user recommendation, ads targeting and product metrics. Most pipelines are fairly simple, consuming events from  Apache Kafka , performing simple filtering, aggregation, metadata lookups, and saving output to  Apache Cassandra  or Elasticsearch. The Storm API is fairly low level which limited its application for complex pipelines. We’ve since  moved from Kafka to Google Cloud PubSub  for ease of operations and scaling.

Apart from batch and streaming data processing, we also do a lot of ad-hoc analysis using  Hive . Hive allows business analysts and product managers to analyze huge amounts of data easily with SQL-like queries. However Hive queries are translated into MapReduce jobs which incur a lot of IO overhead. On top of that we store most of our data in row-oriented  Avro  files which means any query, regardless of actual columns selected, requires a full scan of all input files. We migrated some core datasets to Apache  Parquet , a columnar storage format based on Google’s  Dremel paper . We’ve seen many processing jobs gaining 5-10x speed up when reading from Parquet. However support in both Hive and Scalding has some rough edges and limited its adoption. We’ve since moved to Google BigQuery for most ad-hoc query use cases and have experienced dramatic improvements in productivity. BigQuery integration in Scio is also one of its most popular features which we’ll cover in the second part.

Apache Beam is a new top level Apache project for unified batch and streaming data processing. It was known as Google Cloud Dataflow before Google donated the model and SDK code to the Apache Software Foundation. Before Beam the world of large scale data processing was divided into two approaches: batch and streaming. Batch systems, e.g. Hadoop map/reduce, Hive, treat data as immutable, discrete chunks, e.g. hourly or daily buckets, and process them as a single unit. Streaming systems, e.g. Storm, Samza, process continuous streams of events as soon as possible. There is prior work on unifying the two, like the  Lambda  and  Kappa  architectures, but none which address the different mechanics and semantics in batch and streaming systems.

Beam implements a new unified programming model for batch and streaming introduced in the  Dataflow  paper. In this model, batch is treated as a special case of streaming. Each element in the system has an implicit timestamp and window assignment. In streaming mode, the system consumes from unbounded (infinite and continuous) sources. Events are assigned timestamps at creation (event time) and windowed, e.g. fixed or sliding window. In traditional batch mode, elements are consumed from bounded (finite and discrete) sources and assigned to the same global window. Timestamps usually reflect the data being processed, i.e. hourly or daily bucket.

This model also abstracts parallel data processing as two primitive operations, parallel do (ParDo) and group by key (GBK). ParDo, as the name suggests, processes elements independently in parallel. It is the primitive behind map, flatMap, filter, etc. and behaves the same in either batch or streaming mode. GBK shuffles key-value pairs on a per window basis to collocate the same keys on the same workers and powers groupBy, join, cogroup, etc. In the streaming model, grouping happens as soon as elements in a window are ready for processing. In batch mode with single global window, all pairs are shuffled in the same step.

With this simple yet powerful abstraction, one can write truly unified batch and streaming pipelines in the same API. We can develop against sampled log files, parsing timestamps and assigning windows to log lines in batch mode, and later run the same pipeline in streaming mode using Pub/Sub input with little code change. Checkout Beam’s  mobile gaming examples  for a complete set of batch + streaming pipeline use cases.

We built Scio as a Scala API for Apache Beam’s Java SDK and took heavy inspiration from Scalding and Spark. Scala is the preferred programming language for data processing at Spotify for  three reasons :

• Good balance between productivity and performance. Pipeline code written in Scala are usually 20% the size of their Java counterparts while offering comparable performance and big improvement over Python.
• Access to large ecosystem of both infrastructure libraries in Java e.g. Hadoop, Beam, Avro, Parquet and high level numerical processing libraries in Scala like  Algebird  and  Breeze .
• Functional and type-safe code is easy to reason about, test and refactor. These are important factors in pipeline correctness and developer confidence.

In our experience, Scalding or Spark developers can usually pick up Scio without any training while those from Python or R background usually become productive within a few weeks and many enjoy the confidence boost from functional and type-safe programming.

So apart from the similar API, how does Scio compare to Scalding and Spark? Here are some observations from different perspectives.

Programming model

• Spark supports both batch and streaming, but in separate APIs. Spark supports in-memory caching and dynamic execution driven by the master node. These features make it great for iterative machine learning algorithms. On the other hand it’s also  hard to tune  at our scale.
• Scalding supports batch only and there’s no in-memory caching or iterative algorithm support.  Summingbird  is another Twitter project that supports batch + streaming using Scalding + Storm. But this also means operating two complex systems.
• Scio supports both batch and streaming in the same API. There’s no in-memory caching or iterative algorithm support like Spark but since we don’t use Scio mainly for ML it has not been a problem.

Operational modes

• With Spark, Scalding, Storm, etc. you generally need to operate the infrastructure and manage resources yourself, and at Spotify’s scale this usually means a full team. Deploying and running code often requires both knowledge of the programming model and the infrastructure you’re running it on. While there are services like Google Cloud DataProc and similar Hadoop-as-a-Service products, they still require some administrative know-how to run in a scalable and cost-effective manner.  Spydra  and Netflix’s  Genie  are some examples of additional tooling for such operation.
• Scio on Google Cloud Dataflow is fully managed, which means there is no operational overhead of setting up, tear down or maintaining a cluster. Dataflow supports auto-scaling and dynamic work rebalancing which makes the jobs more elastic in terms of resource utilization. A data engineer can write code and deploy from laptop to the cloud at scale without any operational experience.

Google Cloud Integration

• While there are Hadoop connectors for GCS, BigQuery, plus native clients for several other products, the integration of these with Scalding and Spark is nowhere near as seamless as that of Dataflow.
• This is where Dataflow shines. Being a Google project, it comes with connectors for most Google Cloud big data projects, including Cloud Storage, Pub/Sub, BigQuery, Bigtable, Datastore, and Spanner. One can easily build pipelines that leverage these products.

By moving to Scio, we are simplifying our inventory of libraries and systems to maintain. Instead of managing Hadoop, Scalding, Spark and Storm, we can now run the majority of our workloads with a single system, with little operational overhead. Replacing other components of our big data ecosystem with managed services, e.g. Kafka with Pub/Sub, Hive with BigQuery, Cassandra with Bigtable, further reduces our development and support cost.

This concludes the first part of this blog series. In the next part we’ll take a closer look at Scio and its use cases. Stay tuned.

Sign up for engineering updates

More From Forbes

The amazing ways spotify uses big data, ai and machine learning to drive business success.

• Share to Facebook
• Share to Twitter
• Share to Linkedin

Spotify , the largest on-demand music service in the world, has a history of pushing technological boundaries and using big data, artificial intelligence and machine learning to drive success. The digital music company with more than 100 million users has been busy this year enhancing its service and tech capabilities through several acquisitions. Industry watch dogs predict the company will launch an IPO in 2018 .

Shutterstock

Data: Powerful By-product of Streaming Music

When you have tens of millions of people listening to music every minute of the day, you have access to an extraordinary amount of intel that includes what songs get the most play time, to where listeners are tuning in from and even what device they are using to access the service. There’s no doubt Spotify is a data-driven company and it uses the data in every part of the organization to drive decisions. As the service continues to acquire data points, it’s using that information to train the algorithms and machines to listen to music and extrapolate insights that impact its business and the experience of listeners.

One example is the Discover Weekly feature on Spotify that reached 40 million people in its first year. Every user gets a personalized playlist every week from Spotify of music that they have not heard before on the service, but that will be something the listener is expected to enjoy—a modern-day version of a best friend creating a personalized mix tape.

Spotify for Artists

In an effort to make its mountains of data available to musicians and their managers, 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. Think Google Analytics for musicians. It was originally launched in a web version earlier this year, but the mobile app allows musicians to access the info from the tour bus and the geographic streaming data can be instrumental to musicians and their teams to plan tours more effectively. Artists also have more control over their presence on Spotify including selecting the “artist’s pick,” and they can update their bios and post playlists.

This is just the latest initiative from Spotify to make a concerted effort to empower artists and make them less skeptical of the company. Fans First is another Spotify program that uses data to find an artist’s most passionate fans and target them with special offers.

Spotify Acquires Technology Firms to Enhance Service

With the acquisition of Niland , the fourth acquisition for 2017, 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.

Earlier this year, Spotify acquired the blockchain startup Mediachain Labs to help develop solutions via a decentralized database to better connect artists and licensing agreements with the tracks on Spotify’s service. MightyTV , a content recommendation service, and audio detection startup Sonalytic were also acquired this year.

What’s Next for Spotify?

When news broke that Francois Pachet , a French scientist and expert on music composed by AI, joined the Spotify team to “focus on making tools to help artists in their creative process,” not everyone believed that’s ALL that he’d do. You can just imagine how a leader in AI (Artificial Intelligence) might use his expertise to turn the tables at Spotify to make AI-composed music that would push out artists and their labels. So far, Spotify denies that this will be the case even though this isn’t the first AI feature they launched—AI Duet released earlier this year where listeners could create a duet with a computer.

We can also expect the company to continue to humanize data in creative ways like it did when it used its vast amounts of data to launch a global ad campaign that highlighted some of the more bizarre user habits of 2016. Headlines included “Dear person who played ‘Sorry’ 42 times on Valentine’s Day, what did you do?” and “Dear 3,749 people who streamed ‘It's the End of the World as We Know It’ the day of the Brexit vote, hang in there.”

As Spotify learned in 2015, its community will respond if it feels like it’s taking too many liberties with data. After introducing large-scale changes to its privacy policy, users let the company know they were angry by cancelling subscriptions and taking to social media to express their dismay. This prompted Spotify CEO Daniel Ek  to apologize for unclear communication and made it clear any access to personal data would only occur with the permission of the individual.

We might not know today where Spotify will innovate next, but we will be watching. As innovators they will encounter learning experiences and even failures as they use big data, AI and machine learning to drive success. Those are experiences we can all learn from.

• Editorial Standards
• Reprints & Permissions

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

Enter the email address you signed up with and we'll email you a reset link.

• We're Hiring!
• Help Center

Big Data Analytics: A Spotify Case Study

2021, IJRASET

By developing products that are in line with consumer needs, anticipating their profitability and manufacturing them, Big Data has opened up a lot of possibilities for building customer loyalty and commercial business by proactively engaging and comprehensively streamlining offers across all customer touch points. The use of big data to determine the best, most efficient ways to engage and interact with their customers will be discussed in this paper. An insight into how Spotify intends to provide music lovers additional ways to find their favourite songs, interact with artists, and improve Spotify recommendations has been provided.

Related Papers

International Journal of Case Studies in Business, IT, and Education (IJCSBE)

Srinivas Publication

Netflix is one of the largest online streaming media providers. It began its operations in 1997.Founded by two tech entrepreneur Reed Hastings and Marc Randolph. The Company’s head office is in Los Gatos, California. Netflix’s initially started selling DVDs or provide them on a rental basis. Over the period with growth of internet users and the decline of DVD sales and rental services, it changed its business model to video on demand. From 2012 onwards, it started producing its original TV-series and movies. Netflix uses bigdata analytics to understand its customers base better. By using these data, they provide better service or product to the customer. Netflix collects huge amounts of data from a vast variety of subscriber base. It collects data such as the location of a user; content watched by the user, user interests, the data searched by the user, and the time at which user watched. Based on these parameters its algorithm gives a personalized recommendation based on the user interest. Netflix has constantly focused on changing business needs they have moved their business model from DVD rental to video on demand and currently producing original shows. In this paper we analyze various business strategies of Netflix. This paper also analyzes how Netflix with the help of bigdata analytics focused on improving the subscriber’s experience and how it helped to be more customer-centric and increased its user base. Based on the SWOT and PESTLE analysis we have provided some suggestion that can be incorporated by Netflix as business strategy.

Steven I Davis

Apramey Dube

International Journal of Advanced Research in Computer Science and Software Engineering

Philip Adebo

ABSTRACTBusiness has always desired to derive insights from big data in order to make better, smarter, data-driven decisions. Big data refers to data that are generated at high volume, high velocity, high variety, high veracity, and high value. It has fundamentally changed the way business companies operate, make decisions, and compete. It can create value for businesses. This paper provides a brief introduction to how big data is being used in businesses.

Ossi Ylijoki

John Olotewo

— the notion of big data holds a processing and technology background. In an era that has become noteworthy for the widespread availability and addition of larger data sets in exploiting the massive amounts of data comprised therein (LaValle et al. 2013). Its definition has undergone a considerable evolution as evidenced by its shifts away from its original connotations, which revolved around the control of data varieties, velocity, and volume. The term currently incorporates renewed attention, which emanates from open source technology combinations aimed at the storage and manipulation of data. It has come to include an enhanced importance in the contexts of business intelligence, as well as decision-making and value opportunity. While a significant degree of marketing, endeavors are retrospective in nature. The promise provided by the concept of big data lies in its capabilities, to facilitate predictions and decisions upon which it holds a basis (Gantz & Reinsel, 2011). This idea...

Proceedings of the 1st International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications

Mahmoud Hamed

Kindle Edition

This eBook takes readers on a journey through the complexities of Big Data. It covers fundamental concepts like the 4Vs, artificial intelligence integration, and popular applications. It explores ethical principles, potential misuses, and the evolving future of Big Data. Readers gain a clear understanding of terminology, the 5As of Big Data, and the dynamic relationship between data science and business intelligence. It doesn't just unravel the intricacies of Big Data-it serves as a practical guide for aspiring professionals, offering tips on education, skills development, and staying relevant in this rapidly evolving landscape. This eBook is a roadmap for anyone looking to thrive in the data-driven era, providing insights, strategies, and a comprehensive understanding of the multifaceted world of Big Data.

Business Horizons

Matthew J Mazzei

Journal of the Royal Statistical Society: Series A (Statistics in Society)

Colin Strong

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

RELATED PAPERS

Helena Nobre , Mrs Sasmita Mohanty

Journal of Innovation Management

Luis Rabelo

Abderahman Rejeb

Tamizh arasi

Goran Petkovic

Shahriar Akter

International Journal of Business Administration

Leandro Sabino

5th IASDR - International Association of Societies of Design Research

Pelin Arslan

International Journal of Economics and Business Research

Mohammed T Nuseir

International Journal of Operations & Production Management

John Aloysius

International Journal of Scientific Research and Review

Dr. Mudassir Khan

Exploring the Convergence of Big Data and the Internet of Things

Muhammad Anshari

The International Technology Management Review

Journal of Data Analysis and Information Processing

Ana Akemi Ikeda

Martha Stone

Lecture Notes in Computer Science

Johannes Bjelland

Big Data, Digital Marketing and Trendwatching

Agnieszka Weglinska , Maria Nikolova

Moses Okechukwu Onyesolu

Aviad Dagan

Oğuz Kuş , Ljupcho Efremov

vikram karimella

International Journal of Scientific Research in Science and Technology

International Journal of Scientific Research in Science and Technology IJSRST

RELATED TOPICS

•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

Spotify: Big Data Shows Big Results

Industry Advice Analytics

Spotify is currently the world’s most valuable music company , and for good reason. Now worth around $25 billion, the company has had a major impact on both the popularity of music streaming and the way the music industry uses the data these streaming services generate in impactful ways.

How Spotify Uses Big Data

Streaming music platforms are using data collected by consumer interaction in an effort to hone their algorithms, improve user experiences, target audiences with ads, and make overall better-informed business decisions . Spotify currently has over 108 million paying subscribers —and another 124 million free users—meaning there are billions of streams contributing daily to these processes, whether listeners are directly aware of it or not. Read on to explore some of the most common trends in big data use by Spotify and how that data is being used to improve streaming services today.

Download Our Free Guide to Breaking Into Analytics

A guide to what you need to know, from the industry’s most popular positions to today’s sought-after data skills.

DOWNLOAD NOW

1) To Enhance and Customize User Experiences

One of the most prominent ways Spotify uses the data generated by their customers is to help generate content that each user will consider in-line with their specific tastes. Although Spotify approaches this process from a variety of angles, the overarching goal is to provide a music-listening experience that is unique to each user, and that will inspire them to continue listening and discovering new music that they will be engaged with week after week. This is accomplished through the use of artificial intelligence and machine learning algorithms.

Creating Custom Content

One of the key players in data collection is Spotify’s “Discover” feature, which was first introduced in 2012. This feature began as a playlist of tracks released by a user’s favorite artists but soon evolved to become a recommendation engine of sorts, suggesting a set of songs at the end of a user’s playlists based on the existing songs within it.

Today, “Discover Weekly” has become one of Spotify’s most popular features; it creates a custom playlist unique to each listener’s activity which is curated entirely by a machine learning algorithm. The algorithm analyzes other users’ playlists to find commonalities between tracks, and then utilizes that information to develop a new playlist that fits in with the listener’s existing song preferences. Additionally, each user has a “taste profile” comprised of microgenres that help further customize these playlists. (This full process is outlined in the flow chart to the right.)

In order to successfully customize these playlists, Spotify had to start paying attention to not only what users listen to, but how they interact with each song. If a subscriber plays a track and changes it within the first 30 seconds, for instance, Spotify recognizes this as a “thumbs down” interaction, and won’t use that song’s data when calculating playlists. When a user adds a song to a playlist or library and listens to it in full, on the other hand, it tells the platform that the song positively aligned with their taste, a factor which then helps the algorithm further develop the user’s overall taste profile.

Digitization of Your Taste

A listener’s taste profile is also used in a Spotify function called “Daily Mixes.” These playlists are separated by the genres of music the user typically gravitate toward and are comprised of songs that:

• The user has saved or added to playlists.
• Are written by the same artists the user has in their current playlists.
• Are from new artists or albums the user doesn’t yet know.

These playlists are bottomless and ever-changing, and while they tend to have more familiar content than the “Discover Weekly” playlists, Spotify may still sprinkle in some interesting tracks you don’t know for variety.

Who’s On Your Radar?

“Release Radar” is a weekly playlist comprised of new releases from the artists each user follows, similar in format to the original “Discover” playlist. To get the most out of this playlist, it’s important that the listener actually “follows” their favorite artists on the platform, as this helps the algorithm to develop a more accurate playlist of new song recommendations from that artist. The algorithm may also add a few extra new songs just to keep the mix fresh.

2) To Better Market Their Product

Alongside improving customer experiences, Spotify is able to use the massive amount of data generated by its users to inform its own ad campaigns and better target consumers. At the most basic level, this is done by reviewing what they’ve learned about their listeners and using those insights to develop ads that strategically target their ideal audience.

The Power of Targeted Ads

One display ad —which first ran in Williamsburg, New York—sparked a long-running marketing campaign for Spotify in which the organization used listening history to develop funny, targeted ads.

This first ad—which read, “Sorry, Not Sorry Williamsburg, Bieber’s hit trended highest in this zip code”—was engaging, impactful, and humorous in the local market, considering it was displayed in a “hipster” area known for its notoriously high concentration of “music snobs.”

Through this experience, Spotify was able to gauge the potential impact that using listener data to develop customized ad campaigns could have on their sales and user engagement. This experience sparked a series of strategic and well-received ads that are utilized to market the streaming platform to this day. Some of the most popular campaigns included a set of holiday ads , a set of 2018 Goals ads, and a current set of “ Meme-Inspired ” ads. (Some examples of these campaigns are included below.)

A Constantly Improving System

In April of 2018, the music giant announced its free users would no longer be forced to only shuffle through music on the platform. Instead, these users were granted the freedom to explore 15 of Spotify’s most popular playlists, including “RapCaviar” and “Discover Weekly.” While the free users were excited by the opportunity to experience more of Spotify’s top content, the company had a more data-driven motive behind their decision. This shift in access now generates data on the listening habits of over 124 million more users than before, which is incredibly impactful as the organization works to hone their suggestion algorithms and give users the most customized experience possible.

In general, now that streaming far outranks music purchases, the industry has had to shift its focus from record sales to the collection of this kind of data in order to decipher how the public is responding to an artist, album, or song. Since this data also provides a deeper insight into listening trends, audience markets, and more, it is hopefully a sustainable change for those within the field. Either way, trends predict that Spotify will continue to be one of the largest music data sources for some time, and that data will continue to make for better business decisions across industries.

Interested in learning more about how data impacts almost every industry and organization today? Consider enrolling in the Master of Professional Studies in Analytics program at Northeastern , a top-40 university .

Subscribe below to receive future content from the Graduate Programs Blog.

About shayna joubert, related articles.

Is a Data Analytics Bootcamp Worth It?

What is Learning Analytics & How Can it Be Used?

How to Get Into Analytics: 5 Steps to Transition Careers

Did you know.

Nearly 50% of CIOs report having issues finding qualified candidates for advanced data roles (State of the CIO Report, 2020)

Graduate Programs in Analytics

Join the next generation of data-driven leaders.

Most Popular:

Tips for taking online classes: 8 strategies for success, public health careers: what can you do with an mph, 7 international business careers that are in high demand, edd vs. phd in education: what’s the difference, 7 must-have skills for data analysts, in-demand biotechnology careers shaping our future, the benefits of online learning: 8 advantages of online degrees, how to write a statement of purpose for graduate school, the best of our graduate blog—right to your inbox.

Stay up to date on our latest posts and university events. Plus receive relevant career tips and grad school advice.

By providing us with your email, you agree to the terms of our Privacy Policy and Terms of Service.

Keep Reading:

Top Higher Education Conferences To Attend in 2024

Grad School or Work? How To Balance Both

Is a Master’s in Computer Science Worth the Investment?

Should I Go to Grad School: 4 Questions To Consider

• Viblo Learning
• Viblo Partner
• Viblo Battle
• new Viblo Interview

Big Data Case Study: What makes Spotify successful

• Thêm vào series của tôi

Introduction

We live in a world where buying music has become obsolete, and streaming tunes is the new trendy trend that appears to be here to stay. As a result, streaming systems such as Apple Music, Pandora, Songza, and, of course, the well-known Spotify have risen like phoenixes.

Big data and data analytics have had a significant impact on enhancing user experience. The primary element has been Spotify's "Explore" function, which debuted in 2012 and provides users with fresh listening options. Ultimately, this evolved into a "Discover Weekly" function that sends users a tailored playlist of music they haven't heard before that should match their tastes. Spotify users spent over 2.3 billion hours playing "Discover Weekly" playlists in the first five years. This has not only added value for customers by helping them to find new music, but it has also enabled numerous musicians to break into foreign markets.

"Wrapped" is one of Spotify's most entertaining features. Every December, "Wrapped" compiles a list of users' favourite or most-listened-to songs/artists from the previous year. "Wrapped" also informs users whether they are in the top 1% of an artist's most devoted fans. This information is provided to all users in the form of a customised tale utilising data visualisation. By awarding badges to users who participate with "Wrapped," Spotify creatively fosters user engagement. A Tastemaker badge, for example, might be granted if a user's playlists garnered a large number of new followers. A Pioneer badge is awarded to a person who listens to hit music before anybody else.

Spotify and Big Data - The Lesser-Known Bond

All of these music streaming services have used data gathered from user interactions to better their algorithms, improve user experiences, target potential audiences through adverts, and improve their business approaches and decisions. One area where Spotify excels is in its understanding of its clients. The platform uses proprietary algorithms to understand the user's music preferences and direct them to new genres, songs, and artists.

The days of having to pay to download music are long gone. From Songza, which employed a staff of "music experts" to assemble playlists based on their tastes, to Pandora, which manually identified a song's qualities and allowed users to choose and filter the categories to make the playlists they desired. Then there's Spotify, a music streaming site that uses artificial intelligence, machine learning algorithms, and big data to provide a personalised and unique listening experience.

This demonstrates that Spotify is essentially a data-driven firm that uses data to make choices in all of its tasks. By accumulating data points, the platform is making use of that information for building algorithms and machines to listen to music and provide insights that may assist their business and play a part in enriching the experience of their clients.

Why does Spotify employ Big Data?

Spotify employs Big Data in the following ways:

Creating Customized Content - One critical method Spotify takes to adopting data supplied by its users is to utilise it to create material that each user will perceive as special to their individual likes. The objective is to guarantee that consumers have a positive experience so that they become repeat clients. This was accomplished through the use of several Artificial Intelligence and Machine Learning techniques. For example, Spotify's "Explore" function, which was first introduced in 2012, plays an important part in data collecting. This feature began as a playlist of music published by the user's favourite artists, but gradually evolved into a form of recommendation engine, which recommended a collection of tracks when the user's playlist completed, which were aligned along the lines of the songs in the playlist.

Now, "Discover Weekly" has emerged as one of Spotify's most powerful trump cards; built entirely using a machine learning algorithm, it provides a tailored playlist that is unique to the user's listening activities. The algorithm evaluates other users' playlists to find track similarities, and then uses that information to create a new playlist that fits with the user's current track choices. Furthermore, each user has a distinct "taste profile" made up of microgenres that helps to personalise these playlists.

Digitizing the user's taste - The listener's daily taste profile is also used in Spotify's "Daily Mixes" playlists. These playlists differ from the music genres that the user normally prefers and are typically made up of songs that the user has saved or added to their playlists, or that have been created by artists that the user has included in their current playlists, or any new artists or albums that the user is unfamiliar with. These playlists are large and dynamic; while they may contain more recognisable music than the "Discover Weekly" playlists, Spotify might still include a few fascinating tracks that the listener is unfamiliar with in order to make the playlist more active.

The "Release Radar" playlist is another example. It's a weekly playlist that includes various new releases by the artists that each user follows, in the same structure as the main "Discover" playlist. If users follow their favourite musicians on Spotify, the algorithm can create a customised playlist with new song choices from that artist. The algorithm can also add some new songs to the playlist, making it more interesting.

For Improved Marketing through targeted advertisements - While boosting the experience of consumers, Spotify has also been able to adopt a massive portion of data collected by its users for the purpose of upgrading their ad campaigns and targeting their customers in a more attractive manner. This is accomplished mostly by the platform studying the facts they have collected about its listeners and then using those insights to build adverts that subtly target the platform's intended demographic.

For example, a display ad that originally appeared in Williamsburg, New York, triggered an extensive global marketing effort for Spotify in which the site leveraged listening data to produce amusing, tailored commercials.

Continuously updating its system - In early 2018, the streaming platform had said that its free users will no longer be compelled to just shuffle through music on their application. Rather, its customers were now permitted the liberty of browsing 15 of the platform's well-known playlists which included the platform's renowned "Discover Weekly" as well as "RapCaviar". The fundamental motivation for the platform's choice was driven by data. The access change allows the platform to generate data from an additional hundred million or more users, which is particularly beneficial given that the firm is focusing on improving its suggestion algorithms in order to provide its consumers with a satisfying tailored experience.

In an effort to make their huge quantity of data available to its musicians and management, Spotify launched the Spotify for Musicians programme, which provides access to information such as which playlists have been helping to attract new users and the number of streams they are receiving overall.

The mobile application allows artists to access information through their tour bus, and the geographic streaming data can help musicians and their teams plan tours more effectively. It also gives artists more control over their Spotify presence, such as selecting the "artist's pick" and doing chores such as updating their profiles or publishing playlists.

Spotify Wrapped - This is Spotify's year-end report, a tradition that gives every platform user an occasion to freely share their music tastes on social media. The use of data by Spotify Wrapped goes beyond mere statistics. The corporation has tailored its listening platform to include embedded ego-boosters for its customers. By the end of the year, customers will receive a report notifying them if they are among the top 1% of, say, a band's most devoted fans or among the top non-mainstream song listeners.

Spotify Wrapped essentially serves data to its consumers on a silver platter, presenting it in a way that will interest and captivate them. And it certainly works. The visually appealing data succeeds in making people feel noticed and validated, and it piques their interest. With this data, the platform is creating an experience that is similar to narrating a story using music data instead of words.

In today's environment, when streaming has surpassed bought music, the music business has been forced to shift its focus from record sales to amassing such data in order to decipher the impact of a certain song, artist, or album on the general population. Because the data also provides a more in-depth understanding of listening habits, audience markets, and other similar areas, it represents an ongoing revolution for those in the business.

Spotify, as a social and sharing experience, and by combining its data application with a strong user experience tailored to social media, becomes an accidentally self-marketable platform since users promote their participation on their own.

Enroll today! Check out Skillslash's courses < Data Science Course In Delhi , and Data Science Course > today and get started on this exciting new venture.

All rights reserved

Không có mục lục

thg 11 30, 2023 2:19 CH

thg 11 30, 2023 7:05 SA

thg 11 29, 2023 3:54 SA

thg 3 29, 2023 2:55 CH

thg 3 27, 2023 11:13 SA

thg 4 4, 2022 9:11 SA

thg 4 9, 4:23 SA

thg 3 8, 8:32 SA

thg 12 14, 2023 9:41 SA

Spotify: A Case Study in Business Strategy and Value Compounding

This article is excerpted from a letter by Jake Rosser, Managing Partner of Coho Capital Management.

“If you focus on near-term growth above all else, you miss the most important question you should be asking: Will this business still be around a decade from now? Numbers alone won’t tell you the answer; instead, you must thing critically about the qualitative characteristics of your business.” –Peter Thiel

One of the most important considerations in consumer-facing technology investing is asking whether the product alleviates pain points, reduces friction or enhances convenience. Whether it is Amazon, Netflix or Peloton, all winning consumer platforms exhibit these attributes. With its best-in-class user experience (UX), along with class-leading music discovery and curation, so too does Spotify. It has all the makings of a company on its ways to platform dominance.

Spotify is the category leader in music streaming with 299 million subscribers across 92 countries. With 35% of the global music streaming market, the company has nearly twice the market share of Apple Music, at 19%. Spotify has compounded its leading position in recent years adding premium subscribers at twice the rate of Apple. While Spotify’s growth has been impressive we think the adoption of music streaming is still in early innings.

“Earshare is the new mindshare.” –Andreessen Horowitz

Music streaming has already produced an epochal shift in how people listen to music, with most of the adoption taking place through mobile phones. While historically mobile phones have provided the onramp to music streaming consumption, the next phase of growth will be driven by a plethora of emerging platforms including connected cars, gaming devices, workout equipment and smart speakers (owned by 53 million Americans). Music streaming is tailor-made for the emerging music everywhere lifestyle. The surfeit of products designed for music streaming enables one to listen to the same podcast or playlist while doing a morning workout, commuting to the office (headphones or connected car), working on your office PC and upon returning to home relaxing with a smart speaker. The transition across devices and activities is seamless allowing one to pick up where they left off no matter their activity. The integration of music everywhere into our lives exponentially increases the value of music streaming, moving it from a nice-to-have to a must-have service.

As device proliferation accelerates, Spotify’s position as the category leader makes it most likely to be designed into device presets. Just as Google Maps is pre-loaded on car dashboard screens so too is Spotify. This creates a virtuous feedback loop with scale leading to design integration, which in turn drives scale higher through new consumer trials. This is similar to what we have seen with Sirius’ dominance of satellite radio. Spotify is already available on 300 devices across 80 hardware brands. With that kind of hardware footprint, it becomes very difficult to dislodge incumbency.

While the market tends to categorize music streaming customers on a like per like basis, Spotify users are more passionate. Spotify listeners are twice as engaged as Apple Music users and three times as engaged as Amazon Music Unlimited users. Other than an operating system (which is perhaps the right way to think of Spotify – audio OS), there are few software programs or apps that generate the daily usage of Spotify. The average Spotify user spends 25 hours a month on the service, topping even Facebook at 19 hours a month and Instagram at 14 hours per month.

With each music streaming service offering up largely identical music libraries, conventional wisdom suggests there is little in the way of competitive differentiation making the music streaming platforms commodity businesses. This would be true if one did not care about user experience, search functionality or social integration. The very fact that the average Spotify user is on the platform an entire day a month suggests that ease of use and value of discovery are paramount. There is no doubt that Spotify’s platform is sticky with 70% of churned users returning within 45 days. This is indicative of its competitive differentiation, something we believe the market has not caught on to yet. Ultimately, we think the economic spoils from music subscription will be greater than video platforms as most users will only subscribe to one platform.

Spotify has rerated materially this year rising 70% due to excitement surrounding its podcasting strategy. While shares have recently caught a bid, they had languished for two years prior after going public at $169 per share in April of 2018. Much of the enthusiasm gap during this period can be traced to misunderstandings regarding Spotify’s business model and its future prospects. We will address these misgivings below and outline why we think Spotify is one of the best market opportunities over the next decade.

Labels: “You Can’t Always Get What You Want.” –Rolling Stones

“Aggregators consolidate demand to gain power of supply.” –Ben Thompson, Stratechery

The primary bear case on Spotify is that they will forever be in the music labels’ clutches, with 80% of current streaming hours supplied by four labels. Not only that, but at present Spotify benefits little from operating leverage with variable costs rising in tandem with streaming. Not surprisingly, it is much more lucrative to collect royalties (labels) then it is to pay royalties (Spotify).

To know where we are going, it is helpful to know where we’ve been. Prior to streaming, the music industry endured fifteen years of stagnation and decline with recorded music revenues dropping from $14.6 billion in 1999 to $6.7 billion in 2015 (a 68% drop in inflation adjusted terms). It was only once music streaming gained traction that the industry was able to return to growth. Last year, music streaming represented 80% of the music industry’s revenue.

Before music streaming, it made sense for music labels to collect the lion’s share of profits. After all, labels funded the retail network, oversaw the capital-intensive business of producing and distributing physical media and discovered and promoted stars. Many of those tasks have been rendered obsolete by music streaming. The retail network no longer exists and instead it is Spotify that is funding the build-out of music streaming. Streaming has all but eradicated physical media optimizing label cost structures and promotion is aided considerably by Spotify’s data tools. Despite the seismic shift in industry structure, industry profit pool participation is little changed. In fact, with labels capturing 65% of music publishing profits, one could rightfully accuse the labels of economic plundering.

“But if you try sometimes, well, you just might find, you get what you need.” –Rolling Stones

Current music streaming profit dynamics are unsustainable. It makes no sense for Spotify to continue to finance the build out of global music streaming while the music labels reap all the spoils. Ultimately, we think a sort of détente will prevail with label economic participation curtailed to better reflect their contemporary contribution to the eco-system. With streaming responsible for the resurgence of the music industry, the labels need Spotify for maximum distribution. While in theory the labels could pull their catalogues from Spotify to extract leverage, such a move would sabotage their relationships with music artists who would see their earnings drop precipitously. Rather than play hardball with Spotify, it makes more sense to give up a few points of margin in exchange for a thriving global music industry with double digit growth rates as far as the eye can see. Over time, we expect Spotify to capture the economics of the music industry value chain commensurate with its importance to the eco-system.

Apart from shifts in industry value creation, music streaming is upending how consumers discover new artists. With exploratory music streamers broadening their horizons, the music industry may well be less star-driven in the future. A digital distribution model has fewer gatekeepers than terrestrial radio and retail networks. This allows for an organic discovery process rather than a prescribed feting of the next big thing by label hype machines. While bandwagon effects can be amplified in digital environments, there is growing evidence that the enhanced discoverability of Spotify’s platform is making music listening more diffuse. To wit, “a couple of years ago…the top 90% of listening was about 16,000 artists, that’s now grown to 32,000 artists.” (now 43,000) – Spotify CFO Paul Aaron Vogel in 2019. The net effect should be broader market participation by independent artists, which would weaken label’s power over time. In addition, we expect enhanced discoverability to increase the globalization of music resulting in an erosion of US-centric labels’ market power and increased supply from non-domestic labels where supply tends to be more fragmented. In summary, the importance of search in surfacing music content is bound to diminish the music label’s hoarding of industry profits.

A David Among Goliaths

Apart from supplier power, the other mark against Spotify concerns its ability to ward off the competitive advances of tech behemoths Amazon and Apple. These concerns are not misguided as Apple has demonstrated its ability to take significant share, growing from a standing start in 2015 to 19% market share last year. A large active base of over one billion connected iPhones gives Apple a head-start in establishing a connection with non-music streaming customers and a significant advantage in Customer Acquisition Cost (CAC).

Amazon has also had success (13% market share) with a cut rate offering of $8.00 for Amazon Prime members. Like Apple, Amazon enjoys device advantages due to its Alexa smart speakers as well as Alexa design integration on non-Amazon hardware. Native design in smart speakers is a critical onramp for music subscribers as a request to play music defaults to Amazon Music.

Last, there is YouTube (5% market share), which benefits from its ubiquity on our screens.

As a competitive slate, this is a murderers’ row. All three dominate globally, possess deep pockets and enjoy low CAC. Moreover, each is happy to utilize music as a loss-leader to sell more phones (Apple), serve more ads (Google), or in Amazon’s case, deepen its commitment to its ecosystem. It would seem that Spotify’s die is cast.

Despite their advantages, each of Spotify’s competitors’ ability to scale globally is constrained. In Amazon’s case, it is a market issue. Amazon is in 45 markets relative to the 92 markets in which Spotify has planted its flag. Further, Spotify has done a better job of localizing music offerings than its American counterparts. For Apple, gains have been constrained by the company’s inability to gain significant traction outside of its iOS operating system, which currently hovers around 25% on a global basis. Last, while YouTube is everywhere, its model will always be subpar due to an artist payout ratio per stream only 1/6th that of Spotify – Digital Music News put it thusly, “once again, please don’t ever make a career out of your earnings on the popular video platform. Trust us, you’ll regret it.”

We want to focus our energies on what Spotify brings to the table and why we think it controls its destiny. It is worth nothing, that despite competitor inroads, Spotify remains the undisputed category leader. For observers it is difficult to digest, for in this case the market leader is a David rather than a Goliath.

Data Flywheel Compounds Advantages

“I’m just sitting here watching the wheels go round and round, I really love to watch them roll.” –John Lennon

In our 2016 annual letter we wrote about our attraction to self-reinforcing business models –the rare business where each transaction on its platform makes the business structurally stronger. Spotify is such a business. With more and more businesses harvesting data through AI, scale supremacy is critical. With digital platform businesses, the quantity of data fed to algorithms determines their efficacy. The data spun off by scaled platforms, particularly those with frequent consumer engagement (Facebook, Google, Instagram, Zillow) generate superior insights due to data sets which are an order of magnitude larger than competitor data sets. In Spotify’s case, it uses data insights gleaned from its users’ listening habits to improve its recommendations for daily and weekly playlists. With data training the algorithms, scaled businesses’ advantages compound at ever quickening rates. For example, at two times the size of Apple and twice the engagement, Spotify is collecting four times as much data as Apple. This enables the company to feed its recommendation engines more data compounding its advantages. Further, since Spotify has the most global reach, it is best able to cross-pollinate songs across borders leading to increased listening utility and enhanced discoverability.

While superior data collection provides Spotify a competitive advantage within music streaming, it also enables the company to sit astride emerging audio categories outside of music. Such categories include books, courses, meditations, sports, news, talk radio, podcasts, concerts and live events. Due to the variety of platforms, apps and exclusives, searching for many of these categories is unruly. By aggregating content and serving as a central depository of all things audio, Spotify can remove frictional search costs and become a one-stop-shop for audio content, a sort of Google for audio search. Given its scale and data flywheel there is a more than outside chance this becomes reality. The resulting total addressable market would be multiples larger than currently envisioned in a music streaming scenario. Spotify CEO Daniel Ek has been consistently clear that Spotify’s market is audio and he is going after earshare not music streaming share:

“The market we’re going after is audio. That adds up to two to three billion people around the world who want to consume some type of audio content on a daily or weekly basis. If we’re going to win that market, we’d have to be at least a third of it. We have somewhere between 10-15x of where we are now of opportunity left.” –Daniel Ek on Invest Like the Best podcast

Discovery and Curation – You Get Me Spotify, You Really Get Me

“We are in the discovery business… If discovery drives delight, and delight drives engagement, and engagement drives discovery, we believe Spotify wins and so do our users.” –Spotify F-1 “At the end of the day, margin flows to whoever owns demand creation. So demand creation is everything, both in terms of driving a virtuous cycle of engagement, conversion, retention, and lifetime value.” –Former Spotify CFO Barry McCarthy

Many have complained about the challenge of finding something to watch on Netflix, a platform which clearly fails in curation and algorithmic matching. With music, the challenge can be even more daunting. For example, Spotify has over 50 million songs on its platform, making a robust discovery and curation system even more important.

Discovery is Spotify’s superpower. As elucidated in its IPO filing documents (F-1), Spotify has always understood that its primary mission is to serve as a portal to music discovery. Playlists are the backbone of Spotify’s discovery focused UX with over four billion playlists on its platform. By providing best-in-class discovery and personalization tools, Spotify creates a virtuous flywheel of demand — with discovery driving engagement and engagement feeding data algorithms further improving personalization. The success of playlists in keeping listeners engaged is reflected in listener data with a third of listening time spent on Spotify generated playlists and a third of listening time spent on user generated playlists.

Spotify offers a playlist for every genre and every occasion with its editorial team constantly refining over 4,500 global playlists. The company’s most important playlist is Discovery Weekly, a new playlist delivered each week made just for you premised on your taste (or lack thereof). The product has been a monster hit generating 5.3 billion hours in listening since launching in 2015. By rearranging commoditized content in new ways with continual updating, Spotify is in its own way creating a form of original content. This should ultimately enable Spotify to better aggregate demand increasing its leverage over music suppliers.

The more Spotify users tailor their listening experience to their preferences the less likely they are to leave. Importantly, playlists cannot be shared across music platforms increasing customer retention. Switching costs typically denote a learning curve, but in Spotify’s case it’s premised on a personalization curve.

We all know the best entrée to music is often through an audiophile friend. With the largest base of users, coupled with the best integration in social media, Spotify offers the easiest way to find your friend’s playlist. With social at the center of playlist sharing, playlists are inherently scalable, building a stealth layer of network effects. Spotify already has the largest userbase and most engaged users naturally amplifying existing network effects.

Of course, Spotify’s competitors can produce playlists and curate content as well, but they are technology companies first whereas Spotify has passion for music in its cultural DNA. That spirit is embodied by Spotify’s RapCaviar, the most influential playlist in music. With over 13 million followers, RapCaviar breaks new stars, runs concert tours and moves culture. Just like New York’s Hot 97 used to confer star status on emerging hip hop artists so too does RapCaviar serve as a star maker for aspiring rappers of today.

Two-Sided Marketplace, Now with B Sides

“The problem is Spotify has data that we don’t have. They can see data before our labels can see it, so they have an opportunity to jump and make an investment on an artist that’s not a guess or based on gut, the way everywhere here in this room has to work – it’s based on hard knowledge and facts.” –Richard Burgess, CEO of the American Association of Independent Music

While the future balance of power between labels and Spotify will have outsized influence on Spotify’s future margin structure, we expect to see short-term initiatives provide margin relief as well. Chief among these are Spotify’s Two-Sided-Marketplace platform. As the nexus of global music distribution, Spotify collects a treasure trove of data. As such, it is uniquely positioned to deliver value to artists and record companies through richly featured data analytics. For artists, Spotify data can illustrate demand and preferences by geography and demographics. With behavioral data on 300 million users, artists can see which playlists are driving consumption and learn about their fans. For labels, Spotify can serve as a talent scout, dissecting listening data and offering insights into how to position and market artists.

The opportunity for record labels to utilize Spotify’s data is enormous. Music labels spend roughly $4 billion a year in artist advances, logistics and marketing costs. Historically, much of this spending has been spent on Led Zeppelin tour parties. The opportunity to revamp marketing dollars is vast and Spotify’s data is the key change agent toward optimizing label spend. There is a lot of soft middle ground here for the labels and Spotify to divvy up. Spotify’s Two-Sided Marketplace should allow a wholesale transfer of many of these marketing dollars to its coffers while simultaneously having a material impact on music label’s ROI. The distribution agreement reached between UMG and Spotify this month suggests closer cooperation between the two companies on the utilization of Spotify’s data – UMG commits to “deepen its leading role as an early adoption of future (marketing) products and provide valuable feedback to Spotify’s development team.”

Spotify also plans to utilize its Two-Sided Marketplace to allow for sponsored listings. Sponsored listings are a form of advertising in which labels or musicians can advertise a song to a user matched by Spotify’s algorithms. There are almost no incremental costs for Spotify as songs are merely inserted into existing playlists. Perhaps this is why Ek has stated that sponsored listings will have “software-like margins.”

On the artist side of the marketplace, Spotify has made a number of advances to burgeoning stars to trial direct relationships. According to media reports, Spotify has offered musicians that sign direct a 50% revenue share of music streams, higher than the 30% share offered by labels. While such efforts are a signal to labels of the disintermediation risk poised by music streaming platforms, we expect them to remain a small component of Spotify’s business. At present, Spotify does not have the resources to match the marketing firepower labels spend on roster stars. Nonetheless, it is an important arrow to have in its quiver as the industry evolves and is a signal for labels to play nice.

With the ability to license and promote artists, musicians may choose to increasingly go direct. Chance the Rapper is perhaps most famous for eschewing labels to go direct and yet his star has not dimmed without label promotion. As far back as 2012, Metallica realized its label, Warner Music Group, was not critical for reaching its fans or the buying public. As a result, the band ended its contract with the label and licensed its entire music catalogue to Spotify.

Podcasts: The New Talk Radio

Podcasting is a small market but growing rapidly. From its humble beginnings as a platform for audio bloggers to shout into the void, podcasting is now a $1.3 billion market growing at a 22% compound annual growth rate (CAGR). According to an annual survey commissioned by Edison Research and Triton Digital, one third of Americans listen to podcasts monthly with one quarter listening on a weekly basis. Podcast listeners are a deeply engaged bunch consuming six hours per week. Podcast listeners tend to be upwardly mobile with roughly half making over $75 thousand in annual income and one third having a graduate degree. On Spotify’s platform, engagement with podcasts rose 100% over year-over-year with an additional 20 million monthly average users listening to podcasts over the last six months.

In recent months, Spotify has accelerated its push into podcasting, announcing deals with Michelle Obama, DC Comics, Kim Kardashian and the Joe Rogan Experience. The increase in deal activity builds off the $600 million Spotify spent over the past year to acquire four podcasting companies including Gimlet (original content), The Ringer (pop culture and sports – potentially the ESPN of podcasting), Parcast (original content) and Anchor (distribution and monetizing of podcast content). Spotify is clearly angling for vertical integration to both publish and distribute content and has a real chance of becoming the preferred platform for podcast discovery.

Spotify’s exclusive with the Joe Rogan Experience podcast could be a game changer. In many ways, the Joe Rogan deal is akin to Sirius’ $500 million deal with Howard Stern in 2004. That deal completely changed the trajectory of satellite radio enabling Sirius to scale and drive operating leverage. It is worth noting that Sirius’ deal with Stern leveraged a base of 35 million US subscribers. In Spotify’s case, its deal with Rogan will be spread across 300 million existing Spotify users, not to mention Joe Rogan’s audience of 190 million monthly downloads, suggesting its potential for transformational impact could be even greater. Relative to consumption hours, podcasts are woefully under-monetized with radio generating four times as much revenue per hour.

Spotify’s embrace of podcasts is significant for two reasons; first, the flurry of activity underscores Spotify’s commitment to an audio first (inclusive of audio outside music such as courses, podcasts, meditations and books) market posture. The audio first mentality offers the potential to turn Spotify into the Google of audio search – where one begins their search for all things audio. Second, Spotify’s increasing investments in podcasts should lead to a shift in its cost structure with fixed costs replacing the variable costs paid to music labels. This is similar to Netflix’s shift from licensed content to original content. Like Netflix, Spotify’s move toward greater in-house content should drive operating leverage.

Aside from improved unit economics, podcasts also provide a point of competitive differentiation and thus improve conversion from free to paid while also increasing retention. Increased engagement with podcasts should ultimately result in increased pricing power.

It makes sense to spend heavily now as this is a business where scale begets more scale. As we have seen with Netflix, scale players can pay more for content due to their ability to spread content spend across a broader base of subscribers. On a global basis, this is a significant advantage and one Spotify should pursue aggressively. The more spent up front, the faster Spotify can make the flywheel spin – as long as engagement and new subscribers are rising in tandem. In Spotify’s case, there is the added benefit of lower cost supply due to fixed cost operating leverage on non-music content.

Management – Thinking Fast and Slow

We are big fans of management teams that ignore Wall Street. The first rule of winning is knowing what game you’re playing. For Spotify CEO Daniel Ek, it’s the long game. Since inception, Spotify has never wavered in its mission to be the best audio platform in the world and the best partner to artists and record labels. How a small Nordic-based upstart realigned the global music industry around its vision for music as a service while vanquishing the world’s most profitable company (Apple), the most widely used website service in the world (Google), and the world’s most powerful company (Amazon), all while being at the mercy of consolidated suppliers with money to burn will someday be a master class taught at the world’s best business schools. It is too soon to say Spotify has vanquished its competitors but thus far it has pressed its advantage and widened its lead.

“Music is everything we do all day, all night, and that clarity is the difference between the average and the really, really good.” –Daniel Ek

Mr. Ek has imbued Spotify with several cultural attributes that leave it well equipped to win the prize. First, and probably most important, has been focus. Ek understood early that music is a business about passion and creating a healthy ecosystem would require an artist’s mindset rather than that of a software engineer. That singularity of purpose informs its software. Apple Music is an add-on thrown in to drive revenues whereas Spotify feels like the music geek at your local record store guiding your browsing.

Second, Spotify has put the customer at the center of everything it does. Like Amazon, Netflix, and Costco, investment spend is geared toward elevating the user experience above all else. This is not done in the spirit of charity but in the recognition that customers have choice and scaled Internet platforms win the spoils. Daniel Ek put it best, “engagement drives usage, usage drives data insights, data insights drive a better user experience. A better user experience drives longer lifetime value.”

Mr. Ek has taken a patient approach in building out Spotify’s moat, realizing that he can better fortify the company with a long-term view. Just like Bezos with Amazon, Ek is happy to defer profitability in the pursuit of growth. Ek knows once Spotify achieves the scale he envisions there will be nothing anyone can do to dislodge its dominant perch. In the meantime, however, it appears foolish, just as Amazon’s 20-year march to profitability did.

It takes a unique mix of urgency and strategic planning to both focus on the long-term but relentlessly innovate in the short-term. Long-term thinking invites a plodding approach and an innovate or die approach often leads to sloppy decision making and capital allocation. Given the cognitive dissonance at the center of these two approaches it takes a master tactician to play along the continuum. Daniel Ek seems uniquely capable of playing at both ends. Spotify has continually out-innovated its peers while maintaining long-term discipline and a cost-conscious posture. Yet, it moves more quickly than anybody in the space. Mr. Ek understands scaled players win – “Success for us will be determined by our ability to move faster than everyone else in this space.”

We also like a CEO who puts his money where is mouth is. Late last year, Ek spent $16 million dollars to purchase 800 thousand Spotify warrants expiring in July of 2022. The warrants break even at $211, 56% higher than at Ek’s time of purchase. I can’t recall a CEO spending $16 million of their own money to buy warrants more than 50% out of the money. It’s a strong statement on Spotify’s future.

It is rare to find a CEO who can move fast and out-innovate competitors while at the same time remaining focused on the long game. Amazon CEO Jeff Bezos is one such example. By not playing to the whims of Wall Street, deferring profitability in the pursuit of widening its moat, and treating every day as day one, Amazon has built the most successful and enduring business the world as ever seen. It will be a long time before the world sees another Jeff Bezos but in looking at Daniel Ek’s track record thus far it is clear that he is cut from the same cloth.

Pricing Power – Through the Looking Glass

Unlimited on-demand streaming of a catalog of 50 million songs across devices and without commercials for $9.99 a month is one of the best deals around. Especially when you consider that Spotify has not changed its pricing since its US launch in 2011. Adjusting for inflation alone would equate to a price of $11.45 in today’s dollars.

Subscription services with increased engagement offer substantial consumer utility. While the price remains the same, increased engagement means lower costs for each unit of content consumed (songs for a service such as Spotify and shows or movies for a service like Netflix), a sort of personal operating leverage for consumers. This means marginal costs for extra music consumption is zero. In economics, this is known as a “consumer surplus,” which reflects the difference between the price consumers are paying for a service and the price they are willing to pay. Given the steadily growing engagement of Spotify consumers, it is our contention that the company benefits from a substantial consumer surplus which will be monetized in the future.

Many look at Spotify’s stagnant pricing and view it as proof of a commodity business. This is a dangerous assumption to make. Like Netflix, we believe Spotify has one of the longest pricing power runways in all of business. The decision to not flex pricing now is a conscious decision to hoover up as much market share as possible. As Spotify’s churn statistics indicate, it is difficult for Spotify customers to leave. Playlists, social integration, curation and user experience create enduring habits. The degree of personalization over time makes Spotify very sticky and positions it well to commoditize suppliers rather than the other way around. Given the winner take most nature of globally scaled internet businesses it makes sense to optimize for consumer lock-in now while consumer habits are still being formed. As long as engagement continues to inflect higher, monetization will come.

“The best decisions are the ones that are really instinctive and the most simple. You can use enormous amounts of data and find all kinds of clever ways of slicing and dissecting things. But at the end of the day, the simple decision tends to be the best…One of the simplest decisions you can make is to buy the category winner and wish that the whole category does well. Because if it does, so long as the category winner stays on top of the category, they will get the disproportionate amount of the gains. And that’s been completely true in markets since time immemorial.” –Chamath Palihapitiya

After lying dormant for two years (Spotify IPOed at $169 in April 2018), Spotify’s shares have finally caught a bid. The material rerating in shares was no doubt spurned by excitement surrounding recent podcast announcements and new label negotiations. Despite this year’s robust returns, we continue to believe that Spotify offers one of the best return profiles over the next decade.

To value Spotify, you must ask what this business looks like at scale. Both Goldman Sachs and Morgan Stanley assume the market for paid music streaming subscribers will grow to 1.2 billion by 2030. This seems more than reasonable given the world already has more than 2.7 billion global smart phones in use outside of China, so we will stick with it. We expect Spotify to continue to take market share and net out at 50% of the market. In terms of pricing, we assume Spotify will be able to grow its average revenue per user (ARPU) from $5 to $10. This may sound aggressive, but once Spotify migrates beyond the landgrab stage it will not hesitate to press on the pricing lever. Given the massive consumer surplus enjoyed by customers there is enough pricing runway to more than make up for lower ARPUs in developing markets. Further, ARPU will benefit from extra platform fees generated from Spotify’s Two-Sided Marketplace, ancillary revenue streams and increased share of podcasting in listening consumption. In aggregate, Spotify’s revenue jumps to $72 billion a year. At Spotify’s gross margin guidance of 35% (we assume 40% or higher due to business evolution and favorable label negotiations), the company would generate $25.2 billion in gross profits. Put another way, Spotify trades for roughly two times where we expect gross profits to net out in a decade. This is not inclusive of advertising profits, which given the emerging podcast platform could be significant. Nor does it consider call options on music streaming supplanting radio or Spotify becoming the destination for audio search.

Members , log in below to access the restricted content.

Username or email address  * Required

Password  * Required

Remember me Log in

Lost your password?

Not a member?

Thank you for your interest.  Please note that MOI Global is closed to new members at this time. If you would like to join the waiting list, complete the following form:

First name:

Email address:

City / state / country:

How did you hear about MOI Global? Select Member referral Staff This website Media / Internet Other

If referral, please provide member's name:

Which of the following applies most closely to you? Select Investment Firm - Head Investment Firm - Analyst or PM Institutional Allocator - Firm C-Level Executive Institutional Allocator - Firm Employee Family Office - Head or Co-Owner Family Office - Employee Non-Investment Firm - C-Level Executive Non-Investment Firm - Employee Other

How do you hope to add value to the community?

Keep compounding your investment wisdom:

• The Walt Disney Company: Streaming Service to Widen Moat
• Take-Two Interactive: Premium IP, High ROIC, and Growing FCF
• AB InBev: Scale-Advantaged Leader With Manageable Leverage

About The Author: Jake Rosser

Jake Rosser is the managing partner of Coho Capital Management. He founded Coho in 2007 after working in equity research positions at the value-oriented Auxier Focus Fund and sell-side firm Pacific Crest Securities. He was also a strategy consultant at Alliance Consulting Group. Jake holds an MBA from Tuck and lives in Portland, Oregon.

Digital Innovation and Transformation

Mba student perspectives.

• Assignments
• Assignment: Competing with Data

Spotify – How data is used to enhance your listening experience.

How Spotify uses Big Data to enhance its users’ experience. What are some potential problems that they might face with?

Spotify has long been focused on improving user experience as much as they can, and with their increasing focus on podcast investment, it is reasonable to assume that we will see even more enhancement of user experience.

Big data and data analytics has played a huge role in improving user experience. The core feature has been Spotify’s “Discover” feature, launched in 2012, which made new listening suggestions to users. Eventually, this became a “Discover Weekly” feature that provides users a personalized playlist every week composed of songs that the user has not heard and should align with the user’s taste. Within the first 5 years, Spotify users spent over 2.3 billion hours streaming “Discover Weekly” playlists. Not only has this created value for the users, by allowing them to discover new music, but it also allowed several artists to break into international markets.

Using Big Data to create value

One of Spotify’s funnest feature is “Wrapped”. Every December, “Wrapped” gives users a roundup of their favourite or most listened to song/artist of the entire year. “Wrapped” also lets users know if they were in the leading 1% of an artist’s most loyal listeners. This information is presented using data visualization to create a personalized story to all users. Spotify creatively encourages user engagement with “Wrapped” by issuing users with badges. For example, if a user’s playlists gained a number of new followers, they can be awarded with a Tastemaker badge. If a user listened to a hit song before anyone else, they are issued with a Pioneer badge.

Pathways to a Just Digital Future

Perhaps Spotify’s most impressive piece of engineering is its use of convolutional neural networks (CNN). Using CNN, Spotify analyzes raw audio data such as the song’s BPM, musical key, loudness, etc., to classify songs based on music type and further optimize its recommendation engine.

If Spotify can identify what specific musical attributes make a user like a particular song, their recommendation engine would reach completely new heights. However, I think that would be too much of a tall order.

What type of investments and process put that asset to use, and creates value for the company?

In January 2021, Spotify was granted a patent for speech recognition technology that would allow them to capture audio to more accurately identify the listener’s mood. More specifically, Spotify will be retrieving content metadata from the user’s voice and background noise. This metadata helps determine attributes such as age, gender, accent, which can help indicate the user’s emotional state. The next step is to classify the user’s emotional state which, presumably, helps make a better instantaneous recommendation.

Furthermore, the patent filing mentions another, more basic, approach of determining user emotion. Information such as intonation, stress, rhythm, are used to label user’s speech as “happy, angry, afraid, sad or neutral”.

The figure above demonstrates this process. It shows that once all of the metadata is extracted and analyzed, and previous requests, listening and rating history of the user and friends, new content is recommended.

What challenges and opportunities do you anticipate for the company in the near term?

In the near term, I suspect Spotify will receive less return from their investments into big data and analytics as well as recommendation algorithms. I think it is vital for Spotify to develop new products, which they have done, so far, given their investments into podcasts. However, even with podcasts, a market that is becoming more and more saturated, I believe there are big limitations to what big data can provide to Spotify. I think the roles of recommendations is arguably less important in the podcast world because podcast content is much longer than any single song, so users take longer to digest podcast content.

Other sources:

https://www.musicbusinessworldwide.com/spotifys-latest-invention-will-determine-your-emotional-state-from-your-speech-and-suggest-music-based-on-it/

https://www.musicbusinessworldwide.com/new-spotify-patent-sheds-more-light-on-potential-karaoke-mode-including-auto-tuned-vocals/

Student comments on Spotify – How data is used to enhance your listening experience.

Thank you for the post, Alan. I believe the patent for the voice recognition technology is both promising and concerning. From Spotify’s perspective, it can significantly help tailor the content suggested to users. Nevertheless, I believe it could give Spotify too much power to influence the emotional state of the user. Spotify could have an incentive to manipulate the user’s emotions to encourage them to continue listening to content on the platform. This could imply playing on a person’s depressive state to keep them listening to content that can further deteriorate that emotional state.

Sebastiano, this is a great point! There is always a fine line with these technologies, and you highlight an important way in which Spotify’s technology could be leveraged to exacerbate users’ emotional states.

From your perspective, how might you ensure Spotify’s responsible use of this technology? Do you think policy is ideal, or best practices? How would you like to see Spotify navigate its responsibility towards user mental health vs. maximizing streams?

Thank you for this interesting post, Alan! The improvements in Spotify’s recommendation algorithm have been impressive (anecdotally, I remember it almost being a meme how bad their recommendations were a few years back). I fully agree with your analysis that Spotify urgently needs to build new product/functionalities to distinguish themselves from their competitors as, in contrast to the movie/ tv show streaming ecosystem, all the music streaming providers offer the same music. Spotify Wrapped is a great success in that regard since they managed to turn into a cultural event where people actively share and send around screenshots of their Wrapped. Users of other platforms cannot participate in the fun and social interaction.

Alan, great minds think alike! What a wonderful piece!!

I’d love to hear more about your perspective regarding the limitations of big data with podcasts. I think big data is very impactful when it comes to podcasts. Most podcasts on Spotify contain ads within the podcast, which is a huge source of revenue for both Spotify and the podcasters themselves. Additionally, I think that since podcasts are a bigger commitment, there is less ‘multi-homing’ in the sense that most users may only listen to a handful of podcasts, compared to the thousands of songs. This might be a bit of a stretch, but I’d be interested in hearing your thoughts on this!

Hey Maxwell!

Thanks for your comment. When I was writing about the limitations of big data with podcasts I definitely made a few assumptions that are up to debate. My first one was that because podcasts are longer than individual songs, people won’t be listening to as many podcasts as songs, thus prediction would be made more difficult. Secondly, I assume that listeners are not as obsessed about constantly discovering new podcasts like they are with new songs since podcasts are a bigger commitment. I think that makes sense? I’m not so sure so I guess we’ll discuss in class tomorrow.

You make a great point about big data allowing Spotify to place more personalized ads, that is a huge driver of revenue for them.

Leave a comment Cancel reply

You must be logged in to post a comment.

Simon Kingsnorth

Customer data case study – spotify.

Home > Case Studies > Customer Data

Spotify launched in a handful of European countries on October 7, 2008 and has grown its active user base to become the leading music streaming platform, with over 354 million Monthly Active Users and 155 million Premium Subscribers by the fourth quarter of 2020.

Helping organizations around the world

Spotify’s focus on providing users with a continuously-evolving, personalized listening experience has made it the platform of choice for listeners. Its use of artificial intelligence and data analytics goes beyond simple search to learn the listening preferences of users, suggest new music and podcasts, build customized music playlists, in turn supporting a satisfying, novel and compelling user experience.

According to a 2020 blog post by Nicole Burrow, Senior Design Director at Spotify, the focus as the XD team has doubled over a two-year period has been building an internal design culture that includes growth paths for XD professionals, empathetic management and a robust hiring of skilled professionals and collaborative processes. Cliona O’Sullivan, Head of Design Ops at describes the integration of Design into leadership and business strategy, a hallmark of mature XD practices:

“Design leadership shapes and drives product design at Spotify, enabling our missions to create engaging, usable, and innovative products while maintaining joy throughout our process and output. Each member of the design leadership team sets the expectations of design within both their respective mission and collectively across the company.”

User research is also critical to Spotify’s XD program, with quantitative and qualitative data collection to understand the impact of new features on user experience.

Spotify’s success with listeners has resulted from a relentless focus on creating value for listeners: using technology as a mechanism to personalize and evolve their experience, thereby creating customer loyalty that results in MAU growth and retention. More recently, Spotify has also begun to use data analytics and visualization to support artist understanding of their own listener engagement and expand their creativity. This marks an expansion to a secondary user population, although the strategy of using its technology to heighten awareness, create satisfaction and motivate retention is similar to the approach that has been successful with listeners.

Key lessons

Spotify provides a view of a mature XD organization: its focus on Experience Design, measurable user outcomes and development of a culture that embraces a diverse, multidisciplinary team of XD professionals. As a result, Spotify has become a pace-setter for its industry in innovation, market leadership and product differentiation, all in a few short years.

Exploring How Spotify Uses Data Analytics Effectively Blog

Exploring How Spotify Uses Data Analytics Effectively

• by Yves Mulkers
• May 18, 2023

Spotify, a leader in the industry, has transformed how we access music through utilizing data analytics to give individualized experiences. In this blog post, we will delve into how Spotify uses data analytics to create a seamless user experience and support artists in understanding their fan base.

We'll begin by exploring Spotify's recommendation engine, which has evolved from simple playlist creation to offering highly personalized recommendations using natural language processing (NLP) and semantic search techniques. This ensures accurate content discovery for users based on their preferences and listening habits.

Next, we'll discuss the company's innovative approaches like Discover Weekly playlists and Wrapped insights that provide users with curated content tailored specifically for them. These features not only enhance user engagement but also foster loyalty towards the platform.

Furthermore, we will examine how Spotify empowers artists through its Fan Study initiative and 'Spotify for Artists' mobile app – both of which utilize data analytics to help musicians better understand their audience demographics as well as track real-time performance metrics.

Last but not least, our discussion will cover advanced machine learning algorithms employed by Spotify such as BaRT algorithm optimization and daily retraining processes that enable dynamic adaptation according to changing user behavior patterns over time.

In essence, this blog post aims at providing valuable insights into how Spotify uses data analytics effectively across multiple aspects of its business model – all in pursuit of delivering a superior music streaming experience for millions around the globe.

Table of Contents:

Spotify's recommendation engine, evolution from discover playlist creation to personalized recommendations, utilizing nlp and semantic search for accurate content discovery, discover weekly - weekly curated playlists aligned with user tastes, wrapped - annual roundup of most listened-to tracks by users, fan study - helping artists understand how fans discover them on spotify, 'spotify for artists' mobile app - real-time performance metrics at fingertips, bart algorithm optimizing content delivery based on stream duration, daily retraining process adapting dynamically toitem interaction data points, how does spotify use data analytics.

• Does Spotify use predictive analytics?

How is data analytics used in music?

Does using spotify use data.

One of the key aspects contributing to Spotify 's success is its recommendation engine, which uses semantic search in natural language processing (NLP) to provide users with relevant podcast suggestions based on their queries. This feature has evolved since its launch as "Discover" in 2012, thanks to AI-driven improvements.

In the early days of Spotify, the platform introduced a feature called "Discover", which allowed users to explore new music by generating playlists based on their listening history and preferences. However, this approach was limited due to its reliance on manual curation and user input for creating these playlists.

To overcome these limitations and enhance user experience further, Spotify shifted towards leveraging Big Data analytics , machine learning algorithms, and NLP techniques. These advancements enabled them not only to create more accurate content recommendations but also personalize them according to individual tastes.

Natural Language Processing plays a crucial role in understanding user intent behind their searches within Spotify's vast library of songs and podcasts. By analyzing text-based queries through advanced linguistic models like Latent Semantic Analysis (LSA) , the platform can identify patterns that help predict what kind of content would be most relevant for each listener.

• Semantic Search: Instead of relying solely on keyword matching or collaborative filtering, Spotify uses semantic search to understand the context and meaning behind user queries. This approach allows them to deliver more accurate results by considering factors such as genre, mood, artist similarity, and geographic streaming data .
• Social Media Integration: To further enhance content discovery, Spotify also integrates with popular social media platforms like Facebook and Twitter. By analyzing users' online behavior on these networks, they can gain valuable insights into their preferences and tastes - enabling even better recommendations.

In conclusion, Spotify's recommendation system has evolved significantly since its creation in 2012. Utilizing advanced analytics such as NLP and semantic search along with machine learning algorithms for personalization - the platform is now able to provide users with music recommendations that are tailored to their individual tastes. Through continuous innovation in Big Data analytics techniques like NLP and semantic search along with machine learning algorithms for personalization - the platform is now able to offer highly relevant music suggestions tailored specifically for each listener's unique taste.

Spotify's Recommendation Engine has revolutionized the way users discover and listen to music, providing a tailored experience for each individual user. Moving on from this engine, Personalized Playlists and Insights provide an even more personalized approach by giving users access to their own curated playlists as well as annual roundups of what they have listened to most in the past year.

Spotify's recommendation engine is a key contributor to its success, using big data analytics and NLP techniques for accurate content discovery. The platform shifted from manual curation to personalized recommendations through machine learning algorithms and semantic search, integrating with social media platforms like Facebook and Twitter for even better results.

Personalized Playlists and Insights

To cater to individual user preferences, Spotify employs machine learning algorithms that generate customized playlists like Discover Weekly and Wrapped. These features analyze each user's listening history and offer tailored music selections while providing insights into their favorite songs throughout the year.

Discover Weekly, one of Spotify's most popular personalized playlist offerings, is updated every Monday with a fresh selection of tracks based on users' unique listening habits. By leveraging big data analytics combined with geographic streaming data, this feature creates a highly-customized playlist that aligns closely with individual musical preferences.

• Analyzes past listening behavior to identify patterns in song choices.
• Incorporates social media activity for additional context on users' interests.
• Finds similarities between different artists or genres to create an eclectic mix of recommendations.

The end-of-year tradition known as Spotify Wrapped compiles a comprehensive summary of each listener's top songs, albums, artists, and podcasts from the previous twelve months. This engaging feature not only showcases personal favorites but also provides fascinating insights into global trends across various categories:

• User-specific statistics: Top songs, artists, and genres listened to by individual users throughout the year.
• Global trends: Most-streamed tracks, albums, and artists worldwide in a given year.
• New discoveries: A summary of new music or podcasts discovered by users during the year.

In addition to providing personalized content for listeners, these data-driven features also contribute valuable insights into user behavior that can be leveraged for future improvements within Spotify's platform. By continuously refining its algorithms based on real-time user feedback and engagement metrics, Spotify remains at the forefront of delivering highly relevant recommendations tailored specifically towards individual preferences.

Personalized Playlists and Insights have enabled Spotify to provide users with a tailored music experience that is continually adapting to their tastes. This has been further extended by the introduction of tools such as 'Spotify for Artists' mobile app, which provides artists real-time performance metrics at their fingertips in order to better understand how fans discover them on Spotify.

Spotify uses big data analytics and machine learning algorithms to provide personalized playlists like Discover Weekly and Wrapped, which offer tailored music selections while providing insights into users' favorite songs throughout the year. These features analyze past listening behavior, incorporate social media activity for additional context on users' interests, find similarities between different artists or genres to create an eclectic mix of recommendations, and contribute valuable insights into user behavior that can be leveraged for future improvements within Spotify's platform.

Supporting Artists through Data Analytics

In addition to enhancing the listener experience, Spotify also empowers artists by offering tools such as Fan Study for data-driven growth strategies. The platform recently launched a mobile app called 'Spotify for Artists,' granting them access to analytics related directly from their smartphones.

Fan Study is an initiative by Spotify that helps artists gain insights into how their music reaches listeners and which factors contribute to their success on the platform. By analyzing big data, geographic streaming data, and social media interactions, Fan Study provides valuable information about audience demographics, listening habits, playlist placements, and more. This enables musicians to make informed decisions when it comes to marketing campaigns or planning tours based on where they have the most significant fanbase.

The 'Spotify for Artists' mobile app allows creators easy access to essential performance metrics right from their smartphones. With this application in hand, musicians can monitor daily stream counts of individual songs or albums while tracking overall popularity trends over time. Additionally, they can receive notifications regarding new playlist additions or milestones achieved within the platform.

• Detailed statistics: The app offers comprehensive breakdowns of streams per song or album with demographic details like age group distribution and gender ratio among listeners.
• Audience insights: Artists can understand their listeners' preferences and behaviors, including the top countries where their music is streamed or saved.
• Real-time notifications: Stay updated with playlist additions, follower growth, and other significant milestones on Spotify.

By leveraging data analytics in these ways, Spotify not only creates a more personalized experience for its users but also supports artists by providing them with valuable insights to grow their careers strategically.

Data analytics can be used to provide invaluable insights into fan behavior and preferences, helping artists better understand their audience. Advanced machine learning algorithms are being developed to optimize content delivery and adapt dynamically according to user trends in order to maximize streaming success.

Spotify uses data analytics to support artists by offering tools like Fan Study and 'Spotify for Artists' mobile app, which provide valuable insights into audience demographics, listening habits, playlist placements, and more. By leveraging big data analysis and social media interactions, musicians can make informed decisions when it comes to marketing campaigns or planning tours based on where they have the most significant fanbase.

Advanced Machine Learning Algorithms

Spotify's commitment to delivering a personalized and engaging user experience is evident in their development of advanced machine learning algorithms. One such algorithm, known as BaRT , focuses on optimizing real-time music recommendations by considering streams longer than 30 seconds. This approach ensures that the platform delivers relevant content tailored specifically towards individual listeners' preferences.

The BaRT (Big Data Real-Time) algorithm is designed to prioritize songs with longer stream durations, indicating higher engagement from users. By focusing on these high-engagement tracks, Spotify can provide more accurate and appealing suggestions for its audience. The use of big data allows the platform to analyze vast amounts of information quickly and efficiently, further enhancing its recommendation capabilities.

• Data Collection: Spotify collects various interaction data points daily from users worldwide - including listening habits, playlist additions or removals, likes/dislikes, geographic streaming data (location-based preferences), and social media activity - which are then used to retrain the BaRT algorithm.
• Daily Retraining: With this wealth of fresh data at hand every day, Spotify continuously updates its ML model through a process called "online learning." This method enables the platform's recommendation engine to adapt dynamically according toItem changes in user behavior or emerging trends within specific regions or demographics.
• Fine-Tuning Recommendations: As a result of this daily retraining process, Spotify's recommendations become increasingly accurate and personalized over time. The platform can cater to individual preferences more effectively, ensuring that users continue to discover new music and podcasts they genuinely enjoy.

Incorporating advanced machine learning algorithms like BaRT into their recommendation engine has allowed Spotify to remain at the forefront of content personalization in the digital music industry. By continuously adapting its suggestions based on user interactions and geographic streaming data, the platform offers an unparalleled listening experience for millions of subscribers worldwide.

Spotify uses advanced machine learning algorithms like BaRT to optimize real-time music recommendations and prioritize songs with longer stream durations. By collecting various interaction data points daily from users worldwide, Spotify continuously updates its ML model through a process called "online learning," fine-tuning recommendations and offering an unparalleled listening experience for millions of subscribers worldwide.

Frequently Asked Questions How Spotify Uses Data Analytics

Spotify uses data analytics to create personalized playlists, recommendations, and optimize content delivery. By leveraging user interaction data points, machine learning algorithms like Discover Weekly and BaRT are employed for real-time music recommendation optimization. Additionally, the platform utilizes the Luigi Python framework to manage massive amounts of user data.

Does Spotify use predictive analytics ?

Yes, Spotify employs predictive analytics in various aspects such as forecasting Grammy Awards winners based on streaming trends analysis and creating targeted ad campaigns inspired by actual listener habits. These predictions help improve customer experience and drive business strategies.

Data analytics in music involves analyzing listening patterns, preferences, demographics, and other factors to provide personalized experiences for users. It helps platforms like Spotify , artists with insights into their fan base through tools like Fan Study or 'Spotify for Artists', enabling them to grow effectively.

Using Spotify consumes internet bandwidth or mobile cellular data when streaming songs online unless you have downloaded them for offline playback within the app. The amount of consumed data depends on audio quality settings: higher quality streams consume more bandwidth than lower-quality ones.

Spotify has demonstrated how data analytics can be employed to create tailored experiences and promote business development.  Their use of semantic search in NLP, machine learning algorithms, and real-time optimization techniques have resulted in highly accurate recommendations and playlists for users.

In addition to enhancing user experience, Spotify also uses their vast amount of user data to support artists through fan study insights and the 'Spotify for Artists' mobile app offering real-time analytics. They even predict Grammy Award winners based on streaming trends analysis.

If you want to learn more about how companies like Spotify are using data analytics to gain a competitive edge, visit 7wData today!

[Social9_Share class=”s9-widget-wrapper”]

• Latest Posts

Yves Mulkers

Data strategist at 7wdata.

Latest posts by Yves Mulkers (see all)

• Revolutionizing Urban Mobility: Public Transport Data Analytics - 27 August 2024
• Exploring Generative AI Examples: A Path to Innovation - 30 April 2024
• Maximizing Growth with BI Data Analytics: A Strategic Guide - 28 April 2024

One thought on “ Exploring How Spotify Uses Data Analytics Effectively ”

Loved reading this blog, another way of checking spotify stats is via myspotipie

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

This site uses Akismet to reduce spam. Learn how your comment data is processed .

Upcoming Events

From data silos to unified insights: the importance of connectivity in bi.

28 August 2024

16:00 CET – 17:00 CET

MarkLogic Server 12 Early Access: Harness Generative AI with State-of-the-Art Search

29 August 2024

5 PM CET – 6 PM CET

Unlocking Valuable Content for Next-Generation Applications

10 September 2024

MarkLogic World USA

24 – 25 September 2024

You Might Be Interested In

The role of ai in assisting customer experience.

From being the plots of sci-fi thrillers to being seen as threats by the working populace, Artificial Intelligence (AI) has …

How Leading Organizations are Leveraging Big Data and Analytic

“Data will talk to you if you’re willing to listen”— Jim Bergeson. Few can dispute that. However, the challenge comes …

3 Major Barriers Brands Face in Leveraging Real-Time Analytics with Third-Party Vendors

First-party data is the new imperative for every consumer-facing brand. A brand’s ability to ingest and interpret demographic and behavioral …

Recent Jobs

Senior Analyst, Global Fulfillment Strategy

Hybrid (Columbus, OH, USA)

25 Jun, 2024

Senior Engineer II

Hybrid (Seattle, WA, USA)

Director – Cyber Security Operations

Data Analyst: Support NYS Attorney General Investigations

Hybrid (New York, NY, USA)

Do You Want to Share Your Story?

Bring your insights on Data, Visualization, Innovation or Business Agility to our community. Let them learn from your experience.

Privacy Overview

Get the 3 steps, to drive analytics adoption and manage change.

Get Access to Event Discounts

Switch your 7wdata account from subscriber to event discount member by clicking the button below and get access to event discounts. learn & grow together with us in a more profitable way, create a 7wdata account and get access to event discounts. learn & grow together with us in a more profitable way, don't miss out, stay in touch and receive in depth articles, guides, news & commentary of all things data..

• DOI: 10.22214/ijraset.2021.38702
• Corpus ID: 240480310

Big Data Analytics: A Spotify Case Study

• Suraj Ingle
• Published in International Journal for… 31 October 2021
• Business, Computer Science

Related Papers

Showing 1 through 3 of 0 Related Papers