How Spotify Uses Alternative Data

What are you listening to right now? Right now, as you read this? With streaming apps like Apple Music, Pandora and Spotify, the answer for most people is something rather than nothing. And alternative data probably helped. 

At Suburbia we often talk about how alternative data is used in finance to get an investing edge. We’re not the only ones – do a Google search for “alternative data” and almost all the top search results will focus on finance and hedge funds. But the enormous potential it wields for businesses and governments is often overlooked.

Even tech companies like Spotify are already using alternative data in some shape or form. While the music streaming giant already has a ton of data about its users and the music they listen to, it’s actually looking beyond traditional data sets to make its customer experience better. 

Before we talk about how Spotify uses alternative data, let’s first explore what the term really means.

What is alternative data, anyway?

Most organizations outside of finance might simply think of alternative data as “big data”. So why not just call it big data? 

Because big data is a term mainly used in reference to the amount of data, not the source. Alternative data emphasizes the fact that insights can be uncovered from places you weren’t even looking to begin with. It shines a light on “data exhaust”, the data that’s generated as a by-product of a company’s normal operation. It might not be core to the business, but it could provide elements of information useful to others. 

In addition, when companies think about big data and analytics, it’s often in relation to the data they have under their own roofs. The capability to process and make sense of your internal data is good. But why stop there when there is a wealth of data out there that could be mined and combined? 

In order to unlock the full power of data, businesses need to figure out how to use external data and integrate it into their analysis. It can come from a variety of sources, especially as more non-personal information is being made available as part of open data policies. For instance, the UK government publishes more than 40,000 datasets on data.gov.uk, with similar projects all across Europe. Those that can combine data from multiple sources can end up making better decisions.

To be innovative, think alternative

These days, we’re swimming in a sea of data so identifying what is useful can often feel like an attempt at boiling the ocean. 

There are many sources of alternative data, from geolocation and tracking to web data (scraping) and transaction information based on aggregated and anonymized credit/debit card data. 

Many may not see the value in alternative data until they see an actual example of how it can help their business.

Let’s go back to Spotify, which provides a strong example of how web scraping is used in a creative and resourceful way to improve the user experience. 

Many Spotify users love the curated playlists, like the Discover Weekly playlist. Some have even tweeted that the streaming service knows their taste better than their loved ones.

One of the tools Spotify and other music streaming services use to generate these playlists is similar to what ecommerce sites use to recommend items.

To make its predictive algorithm more precise, Spotify doesn’t rely on just one source of information or its internal data alone. It combines a few different approaches. One of these is natural language processing (NLP). You’ve probably heard of it before but never realized the important role it plays in Discover Weekly. Basically, Spotify is continuously scouring the web, analyzing blog posts, articles and other text about music. It looks at which terms are used in reference to artists and songs, and each term is weighted by relevance. This enables Spotify’s engine to determine which songs or artists are similar to each other.

That could even inadvertently throw out a lifeline to music journalism, a field that’s seemingly on the brink of extinction! Especially if it serves greater utility than just being an arbiter of taste – after all, Rolling Stone’s reviews may be helping to decide whether or not Fleetwood Mac makes it to your Discover Weekly playlist…or that of Spotify’s 217 million active users.

The use of alternative data helps to surface unique insights for businesses that they might have otherwise failed to find. Like technology, data can be a real driver of innovation and differentiation. If you’re not doing it differently, then you’re never going to break from the pack. 

If Spotify’s spiders are crawling this article to look for descriptive words, we’d like to finish with some awesome, excellent, amazing recommendations we put together the old-fashioned way: hand-picked from our team to yours.



Leave a Reply