Spotify in spotlight

Spotify has been put in the public eye quite often recently, largely thanks to Joe Rogan’s controversial narratives in his podcast and Neil Young’s threat to withdraw all his music as a counterstrike. On the business side, Spotify released the “Car thing”, a dedicated plugin device for playing Spotify, targeting vehicles without screens and Bluetooth. Feeling interested in Spotify’s trajectory, I read a book over the long weekend, The Spotify Play, written by two Swedish investigative journalists, Sven and Jonas.

That was a good read not only for understanding Spotify’s ambition in music and hardware, but being able to broadly scope to the entertainment industry, streaming war, VC funding, and the entire entrepreneurial ecosystem in Sweden and US. Here are some of what I learned and thought:

1- Try probabilistic methods when deterministic methods get you nowhere

The book primarily focuses on the strategy front of Spotify, but my engineering radar directs me to its scattered narratives about Spotify’s innovative engineering effort, particularly to reduce the buffering latency in its peer-to-peer-style transmission protocol. It all started from Spotify’s co-founder and CEO Daniel Ek’s mania on file-sharing and music, and his effort to build a legal streaming service, unlike Napster and Pirate Bay, but using their technology. However, peer-to-peer/BitTorrent protocol has a natural downside: it slices big files into small chunks and combines them when the host collects them all, which, for music streaming, means long buffering latency. To trigger an instant start, Spotify turns to a probabilistic solution: its hybrid protocol only slices music files into 4 pieces (MB-scale for mp3 format) and the chances of loading the first piece upfront are dramatically increased.

In addition, Spotify preloads anticipated songs to ensure a smooth transition even in a shabby network environment, and only direct streams from its high-end servers when users’ selection is unexpected. Balancing out the cost and UX, the probabilistic solution is cheap and ensures a good user experience for a majority of the audience.

2- Secret Sauce: Freemium, but Shuffle only

Besides the instant start, what differentiates Spotify and other streaming services like Beats Music, Pandora, and Tidal is its free tier, for which Spotify exhausted all its resources to negotiate and transferred up to its 85% of profits, 15% of the equity with a Schmuck insurance to settle with big record labels because music industry found hard to understand the new approach to distributing music “FOR FREE”. However, the way humans consume music determines we are more willing to rent, instead of buy, music when the media permits. How many readers have not listened to any songs you bought on iTunes, but start streaming music whenever you read / code?

On that note, Spotify rents music to users by streaming encrypted music to your device and keeping the key in the cloud. Thanks to the early adoption on Facebook page that channeled billions of users to Spotify, the free tier retained most of them, but here comes the question of how to convert them to paid users. One thing Spotify did right is down-to-earth user research that suggested the best user experience for free tier came when users have unlimited time, but only access a subset of the catalog played on a shuffled mode and no offline playing. This offering is strong enough to attract new users and weak enough that existing users would not downgrade. Such a perfect sweet spot for Spotify to grow its user base and retention rate!

This becomes one of the fundamental etiquette in Spotify that free tier users have access to all music with limits. Even after Bob Dylan and Taylor Swift withdrew all their music, Spotify didn’t yield at the expense of this etiquette, otherwise, the entire business model would fall apart. Data was telling the same story. Spotify only observed tiny subscription cancellations after those withdrawal events, and always successfully lured artists back based on their FOMO of a stable revenue stream.

3- Every software company is building hardware

Spotify has always been ambitious in building hardware that serves as the user’s entry point to its service. From 2011 to 2015 in the stealthy project Magneto, Spotify poured countless resources to build a Spotify TV hardware, by which it tried to replicate its success in music by getting licenses from TV stations and distributing to paid subscribers, but eventually failed apart. Unlike big labels in the music industry, the TV/video industry is way more fragmented and impossible to seal deals without skyrocketing the subscription fee ($45 only for TV in Sweden in 2014). Up to date, YouTube and Prime’s offering in the field is remarkable yet still lacking most live sports. A startup like Spotify was not in the shape to tackle back then.

Yet, Spotify is not the only company seeing the opportunity in the video streaming market, but suffering from the high cost and fragmentation. Currently, there are two ways out: The first is making your own content, like Netflix and HBO, that requires tremendous upfront capital, but never be restricted by copyright issues; the second is outsourcing content generation to community creators, like TikTok and Reels. Spotify tried both: it builds a Netflix-style service for podcasting, signing exclusive titles and producing their own, and predicts podcast leads to more paid users; it also briefly allowed musicians to upload their pieces and almost acquired a music community SoundCloud, but sadly yielded to the pressure from monopolizing record labels.

In an interview after Spotify’s IPO, Daniel Ek justified his effort to build Spotify TV and said it confirmed Spotify’s focus on audio and audio-only. But its hardware ambition never fades away. The newly released car thing is targeting old vehicles and, more importantly, cultivating users’ listening habits across all age spectrums. The plugin device is not supposed to stay long, since a 2018 federal law requires new cars to have a rear camera and a screen, let alone Tesla comes with a gigantic “iPad” and built-in Spotify app, but Daniel always believes acquiring users go first in the chain of acquisition, activation, and retention. “More listening, more paying.”

4- Carpe Diem

Grappling with record labels, lacking artists’ support, challenged by Goliathes like Apple and Google, Spotify turned towards and laser-focused on its millions of users. Unlike Tidal and Apple Music, who root much deeper in the music industry, marketing their expert-curated playlists, Spotify runs clustering algorithms (i.e. Nearest-neighbors and Collaborative Filtering) based on 1.5 billion user-curated playlists, both their titles and contents, to create Daily Mix and Discovery Weekly, in which users surprisingly found Spotify knows their taste better than themselves. Once proving it’s low-cost, effective, and scalable, Spotify cannot help extending this approach to presumably seize every single moment of users’ lives by music. Granular playlists for baby showers, puppy dates, and alike are generated in batches. Spotify is not streaming music, but delivering fleeting experiences.

Another side effect from AI-curated playlists, like Discovery Weekly, is Spotify has the power to replace talent scouting that traditionally has been done by record labels. The scale of power between record labels and Spotify gradually tilts towards the distributor as it consolidates upstream services according to users’ preferences.

5- Brilliant Swede

The read also provides me an opportunity to glance at the entrepreneurial ecosystem in Sweden, which gave birth to world-class companies and unicorns like Ericsson, IKEA, H&M, and Volvo. Nordic ever-long and damp winters foster Swede’s culture of deep thinking and passion towards math and science. Founding engineers of Spotify mostly come from well-known KTH and laid the engineering fundamentals in line with Spotify’s etiquette — no tolerance of buffering.

On the downside, stock options are taxed heavily in Sweden, which forces startups to hire more employees overseas. Nowhere near being as nimble as VC funds, the largest investing institutes in Sweden are pension funds. Politicians argue that great welfare encourages entrepreneurs to take risks, but others see it hampers creativity and innovation. Spotify’s holding company and its co-founder’s investment funds are registered in Cyprus, Malta, and Luxembourg, and its employee body in New York is larger than in Stockholm. All these facts lead to whether Spotify is still a Swedish company, despite it being born and brought off the ground in the hometown of the Nobel Prize.

Welfare should not be static. A dynamic mechanism adopting changes in social environment is the key to keeping a country being competitive.

Parting thoughts: I felt very inspired reading the story of Daniel and his colleagues building Spotify and moving in the power structure of record labels, tech giants, and artists, just because they firmly believe their business model would eventually work. Even getting involved in several acquisition bets by Google and Microsoft, Daniel probably never thought of selling Spotify, but used the chance to test its market value, even in the “Death Valley” when Spotify was in a grinding negotiation with record labels for mobile licenses. Finally, Spotify rode on the wave of the mobile revolution and thrived. In the next wave of AR devices to replace mobile, should we put our faith?




Cloud Enthusiast

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Kinsu Q&A — Chris Sharpe, Kinsu CEO

Fintech Lender LendUp Shuttered by Consumer Bureau

My Interview In Forbes Today Apr 19, 2021, 010:00am EST

Twitter Now Forbids Posting Photos Of Other People…And Other Small Business Tech News This Week

Cloudflare S-1 Analysis — Do they have an edge?

IoT In Supply Chain Management & Its Benefits

Anjani Kasliwal VP & MD , Skumars Nationwide & Nitin Kasliwal VC & MD Skumars Nationwide in…

Here’s What You Need to Know about the Restaurant Revitalization Fund

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Michael Zhang

Michael Zhang

Cloud Enthusiast

More from Medium

[How To Write Bullets That Sell…]

Why Indians love Physical Stores: 5 humble lessons for digital commerce

Achieve Community-Led Growth, One User at a Time

The Importance of Digital Studies in Every Career