How did Spotify get so good?
I joined Spotify six years ago and have listened to it, pretty much, every day since. Recently, it's not just because of its seemingly unlimited choice of songs, artists and genres - Apple Music has that. No, what keeps me hanging around is that it knows me better than I know myself.
Waking up on a Monday can be difficult for a lot of people. Not me. Like 75 million other people around the world, I have something to look forward to:
‘Discover Weekly’ is a weekly automated music recommendation engine provided by Spotify... But it’s so much more than that. It’s more like a therapy session than a playlist, 30 songs that help me understand the real me. It’s been said, if you want to get to know someone, listen to their Spotify, not their Facebook wall, Twitter feed or Instagram story - I’m not sure if that has actually ‘been said’. If not, you read it here first.
The reason behind this (almost) scary understanding is, of course: data. But how can data understand the subjective realm of music? And how can it please people week after week?
Matthew Ogle, who heads up DW at Spotify said that they have “more technology than ever before” to “connect the dots between artist and listeners,” the example he gives is:
“if you’re the smallest, strangest musician in the world, doing something that only 20 people in the world will dig, we can now find those 20 people.”
The quality of the picks is incredibly good. How is it done? I’m afraid, it’s not an easy answer… in short, they match what you like to listen to with what other people like to hear to and then tease out commonalities and ‘near-separations’ in your music choices… sort of like the ‘Mutual Friends’ feature on Facebook. This algorithm is so complicated that it involves diagrams like this:
I found a flowchart in a great Quartz article that does a good job of representing how it all works:
Actually, that makes it look pretty simple…
Taking huge scores of data and using it to power smart automation isn’t new, I mean, that’s how Google has worked for years: “based on all the information available, we (Google) believe the most likely website you are looking for is this one, then the second most, the third” and so on…
This new era of data recommendation and automation is about to be introduced to the world of finance.
You can read what we wrote about Open Banking here
Data will play a huge role in helping people be better served by their money. The products - the bank accounts, savings accounts, insurance, mortgages, investments - will work for people, to increase the likelihood of individual's achieving their financial goals. Imagine never having to look for the best way to do something with your money. The right product pops up at the right time and it’s there, ready for you to use.
We’re not far from a world where, instead of only knowing where to send your statements, your bank will be more like this:
“... knows more about me than I do” - imagine that.
By understanding spending data, behavioural data and group data, the helpfulness of banks and financial institutions will increase ten fold. When a customer’s ‘discover weekly’ is a new service that helps them save, spend and invest money, everyone will be better off.
Stay tuned for our next posts on this subject: 'The next steps' and 'The operations'