For years, Spotify has tried to work out what you like by watching what you do.<\/p>\n\n\n\n
What you play.<\/p>\n\n\n\n
What you skip.<\/p>\n\n\n\n
What you save.<\/p>\n\n\n\n
What you search for.<\/p>\n\n\n\n
What you listen to repeatedly.<\/p>\n\n\n\n
Now Spotify is starting to ask a completely different question:<\/p>\n\n\n\n
What do you want to like?<\/p>\n\n\n\n
Spotify has launched its new Taste Profile beta for Premium users aged 18 and over in the United States, following an earlier test in New Zealand. The feature allows listeners to see how Spotify interprets their tastes and then directly tell the platform how they want those recommendations to change.<\/p>\n\n\n\n
It sounds like a relatively small recommendation feature.<\/p>\n\n\n\n
But it could represent a much bigger change in how music discovery works.<\/p>\n\n\n\n
Spotify is moving from passive personalization towards something closer to a user-programmable recommendation algorithm.<\/p>\n\n\n\n
And that could make music discovery considerably more interesting.<\/p>\n\n\n\n
Spotify’s Taste Profile is essentially the platform’s interpretation of what you are interested in.<\/p>\n\n\n\n
Spotify says actions such as listening, searching, skipping and saving music can influence this profile, which its recommendation systems then use when deciding what to show you.<\/p>\n\n\n\n
Previously, most of this happened behind the scenes.<\/p>\n\n\n\n
You listened.<\/p>\n\n\n\n
Spotify learned.<\/p>\n\n\n\n
Spotify recommended.<\/p>\n\n\n\n
Taste Profile now begins to expose part of that relationship to the listener.<\/p>\n\n\n\n
Premium users taking part in the beta can open their profile inside Spotify and see an AI-generated description of how Spotify currently understands their taste across music, podcasts and audiobooks.<\/p>\n\n\n\n
More importantly, they can correct it.<\/p>\n\n\n\n
Under the “Tell us more” section, users can type instructions describing how they would like their recommendations to change.<\/p>\n\n\n\n
You might tell Spotify:<\/p>\n\n\n\n
\u201cShow me more new indie rock.\u201d<\/p>\n\n\n\n
\u201cI want to discover more jazz.\u201d<\/p>\n\n\n\n
\u201cGive me more energetic music in the morning.\u201d<\/p>\n\n\n\n
\u201cI’ve had enough of lo-fi.\u201d<\/p>\n\n\n\n
\u201cI want to explore 1990s alternative rock.\u201d<\/p>\n\n\n\n
Spotify says those instructions act as guidance for the recommendations appearing on the Home feed, with changes typically beginning to appear after the profile updates.<\/p>\n\n\n\n
This is the interesting part.<\/p>\n\n\n\n
Recommendation algorithms have traditionally relied heavily on what we might call revealed taste.<\/p>\n\n\n\n
You reveal your preferences through your behaviour.<\/p>\n\n\n\n
If you listen to Fred again.. repeatedly, Spotify can reasonably assume you might like similar electronic music.<\/p>\n\n\n\n
If you constantly skip country songs, that creates another signal.<\/p>\n\n\n\n
The problem is that listening behaviour and actual preference aren’t always the same thing.<\/p>\n\n\n\n
Imagine you play relaxing piano music every night while you sleep.<\/p>\n\n\n\n
Spotify sees hours of piano listening.<\/p>\n\n\n\n
That doesn’t necessarily mean you want your daytime recommendations filled with piano music.<\/p>\n\n\n\n
You might also have children using your account.<\/p>\n\n\n\n
You may have played an album repeatedly because your partner likes it.<\/p>\n\n\n\n
Perhaps you’ve become obsessed with one particular artist for three weeks but don’t necessarily want Spotify to spend the next year recommending artists that sound exactly the same.<\/p>\n\n\n\n
Spotify has already been trying to solve these problems.<\/p>\n\n\n\n
Users can exclude individual tracks and playlists from their taste profiles, reducing the effect those listens have on recommendations. Spotify recently also introduced a setting allowing parents to prevent kids’ and family music from influencing their profiles.<\/p>\n\n\n\n
Taste Profile goes considerably further.<\/p>\n\n\n\n
Instead of simply removing bad signals, listeners can introduce entirely new ones.<\/p>\n\n\n\n
This introduces another fascinating concept.<\/p>\n\n\n\n
Your listening history tells Spotify who you were.<\/p>\n\n\n\n
A prompt can tell Spotify who you want to become.<\/p>\n\n\n\n
Imagine you’ve decided you want to understand jazz.<\/p>\n\n\n\n
Your Spotify history might contain virtually no jazz at all.<\/p>\n\n\n\n
Traditionally, Spotify’s algorithms have little evidence suggesting they should suddenly begin recommending Miles Davis, John Coltrane or newer jazz artists.<\/p>\n\n\n\n
With Taste Profile, you can explicitly tell Spotify that this is something you’re interested in exploring.<\/p>\n\n\n\n
That is a very different recommendation signal.<\/p>\n\n\n\n
Listening data says:<\/p>\n\n\n\n
\u201cThis person likes electronic music.\u201d<\/p>\n\n\n\n
Declared intent says:<\/p>\n\n\n\n
\u201cThis person likes electronic music, but wants to discover jazz.\u201d<\/p>\n\n\n\n
The second statement arguably gives a recommendation system much more room to work with.<\/p>\n\n\n\n
It moves music discovery away from simply extrapolating your past.<\/p>\n\n\n\n
One of the biggest weaknesses with algorithmic recommendations is that they can become self-reinforcing.<\/p>\n\n\n\n
You listen to something.<\/p>\n\n\n\n
Spotify recommends something similar.<\/p>\n\n\n\n
You listen to that recommendation.<\/p>\n\n\n\n
Spotify receives another signal that you like that style.<\/p>\n\n\n\n
It recommends even more of it.<\/p>\n\n\n\n
Eventually your recommendations can become incredibly accurate but also incredibly narrow.<\/p>\n\n\n\n
This is the personalization paradox.<\/p>\n\n\n\n
The better an algorithm becomes at understanding what you already like, the easier it becomes for discovery to remain inside the same musical neighbourhood.<\/p>\n\n\n\n
Taste Profile potentially gives listeners an escape hatch.<\/p>\n\n\n\n
Rather than waiting for Spotify to unexpectedly introduce something new, you can tell it that you specifically want more variety.<\/p>\n\n\n\n
Spotify says users can ask for more or less of genres, artists or particular vibes, or tell the service they want to explore something completely different.<\/p>\n\n\n\n
That gives the listener a degree of control over how adventurous the recommendation system should become.<\/p>\n\n\n\n
Taste Profile also makes more sense when viewed alongside Spotify’s other AI features.<\/p>\n\n\n\n
Spotify has been experimenting with Prompted Playlists, allowing listeners to describe the playlist they want using natural language.<\/p>\n\n\n\n
In July 2026, Spotify also announced conversational features allowing eligible Premium users to type or speak requests directly to Spotify, including instructions such as asking it to play artists they haven’t heard before.<\/p>\n\n\n\n
Taste Profile takes the same natural-language interface and applies it to personalization itself.<\/p>\n\n\n\n
That creates three increasingly distinct ways to interact with Spotify.<\/p>\n\n\n\n
You can manually search for music.<\/p>\n\n\n\n
You can allow Spotify’s algorithms to recommend music automatically.<\/p>\n\n\n\n
Or you can talk to Spotify and explain what you want.<\/p>\n\n\n\n
The third option could eventually become incredibly important.<\/p>\n\n\n\n
Instead of navigating Spotify, you increasingly describe what you want Spotify to do.<\/p>\n\n\n\n
We’ve become accustomed to recommendation algorithms being mysterious systems operating in the background.<\/p>\n\n\n\n
YouTube decides which video comes next.<\/p>\n\n\n\n
TikTok decides which clip appears in your feed.<\/p>\n\n\n\n
Instagram decides which posts you see.<\/p>\n\n\n\n
Spotify decides which music you might like.<\/p>\n\n\n\n
Users provide behaviour.<\/p>\n\n\n\n
Algorithms provide recommendations.<\/p>\n\n\n\n
Taste Profile introduces a slightly different relationship.<\/p>\n\n\n\n
The algorithm becomes something that can be instructed.<\/p>\n\n\n\n
You aren’t controlling every recommendation, and Spotify makes clear that Taste Profile cannot currently enforce absolute rules such as \u201calways show\u201d or \u201cnever show\u201d a particular type of content.<\/p>\n\n\n\n
But you are beginning to manage its direction.<\/p>\n\n\n\n
That feels less like a traditional recommendation engine and more like having a personal music curator that you can train.<\/p>\n\n\n\n
There is another interesting consequence.<\/p>\n\n\n\n
If listeners start explicitly telling Spotify what they want to discover, the signals driving music discovery could begin changing.<\/p>\n\n\n\n
Historically, an emerging artist might depend on listeners behaving in ways that cause recommendation algorithms to connect that artist with similar music.<\/p>\n\n\n\n
Taste Profile potentially adds another route.<\/p>\n\n\n\n
A listener could explicitly ask Spotify for:<\/p>\n\n\n\n
\u201cNew Australian indie rock.\u201d<\/p>\n\n\n\n
\u201cFemale electronic producers I’ve never heard before.\u201d<\/p>\n\n\n\n
\u201cModern artists influenced by 1970s funk.\u201d<\/p>\n\n\n\n
\u201cNew alternative bands with fewer than 100,000 monthly listeners.\u201d<\/p>\n\n\n\n
Spotify hasn’t said that all requests of this kind will work exactly this way, and Taste Profile currently operates as softer guidance rather than a rigid search filter.<\/p>\n\n\n\n
But the wider direction is interesting.<\/p>\n\n\n\n
Discovery could increasingly be driven by listener intent rather than solely by historical listening behaviour.<\/p>\n\n\n\n
For independent artists, that could eventually create new opportunities to reach listeners who are actively asking for something different.<\/p>\n\n\n\n
There is another reason natural language matters.<\/p>\n\n\n\n
Listening history is incredibly useful data, but it lacks context.<\/p>\n\n\n\n
Spotify knows you listened to a song 20 times.<\/p>\n\n\n\n
It doesn’t necessarily know why.<\/p>\n\n\n\n
Maybe you love the artist.<\/p>\n\n\n\n
Maybe you’re learning the guitar part.<\/p>\n\n\n\n
Maybe you’re creating a playlist for someone else.<\/p>\n\n\n\n
Maybe your child kept playing it.<\/p>\n\n\n\n
Maybe you accidentally left Spotify running.<\/p>\n\n\n\n
Language can provide the missing context.<\/p>\n\n\n\n
\u201cI love this artist, but don’t recommend similar artists.\u201d<\/p>\n\n\n\n
\u201cI’ve started getting into classical music.\u201d<\/p>\n\n\n\n
\u201cI want more unfamiliar music.\u201d<\/p>\n\n\n\n
\u201cI listen to ambient music while working, but I don’t want it affecting my normal recommendations.\u201d<\/p>\n\n\n\n
Suddenly Spotify receives information that listening data alone could struggle to reveal.<\/p>\n\n\n\n
There is also a major strategic advantage for Spotify.<\/p>\n\n\n\n
The more accurately Spotify understands your taste, the more personalized your account becomes.<\/p>\n\n\n\n
Years of listening history already create a substantial personalization advantage.<\/p>\n\n\n\n
Now imagine adding years of explicitly stated preferences.<\/p>\n\n\n\n
Spotify might eventually know that:<\/p>\n\n\n\n
You like energetic electronic music while exercising.<\/p>\n\n\n\n
You prefer acoustic music on Sunday mornings.<\/p>\n\n\n\n
You’ve recently started exploring jazz.<\/p>\n\n\n\n
You don’t want children’s music affecting your recommendations.<\/p>\n\n\n\n
You want more emerging artists.<\/p>\n\n\n\n
You enjoy technology podcasts but aren’t interested in business podcasts.<\/p>\n\n\n\n
That represents an increasingly detailed personal entertainment profile.<\/p>\n\n\n\n
Switching to another streaming service would mean starting much of that relationship again.<\/p>\n\n\n\n
Taste Profile therefore isn’t only about improving recommendations.<\/p>\n\n\n\n
It could also deepen Spotify’s personalization moat.<\/p>\n\n\n\n
Music streaming services broadly have access to much of the same catalogue.<\/p>\n\n\n\n
Spotify.<\/p>\n\n\n\n
Apple Music.<\/p>\n\n\n\n
YouTube Music.<\/p>\n\n\n\n
Amazon Music.<\/p>\n\n\n\n
The difference between them increasingly isn’t simply which songs they contain.<\/p>\n\n\n\n
It’s how effectively they can help you navigate tens of millions of them.<\/p>\n\n\n\n
That makes discovery one of the most important battlegrounds in music streaming.<\/p>\n\n\n\n
Spotify built much of its reputation around algorithmic personalization.<\/p>\n\n\n\n
Taste Profile suggests its next step may be turning that algorithm into something collaborative.<\/p>\n\n\n\n
Instead of Spotify silently watching what you listen to and trying to predict your next song, you can increasingly just tell it.<\/p>\n\n\n\n
That might sound like a subtle change.<\/p>\n\n\n\n
It isn’t.<\/p>\n\n\n\n
The streaming recommendation model is gradually shifting from:<\/p>\n\n\n\n
\u201cWe think you’ll like this.\u201d<\/p>\n\n\n\n
Towards:<\/p>\n\n\n\n
\u201cTell us what you’d like to discover.\u201d<\/p>\n\n\n\n
And that could fundamentally change the way we find music.<\/p>\n","protected":false},"excerpt":{"rendered":"
For years, Spotify has tried to work out what you like by watching what you do. What you play. What you skip. What you save. What you search for. What you listen to repeatedly. Now Spotify is starting to ask a completely different question: What do you want to like? Spotify has launched its new Taste Profile beta for Premium users aged 18 and over in the United States, following an earlier test in New Zealand. The feature allows listeners to see how Spotify interprets their tastes and then directly tell the platform how they want those recommendations to change….<\/p>\n","protected":false},"author":1,"featured_media":4630,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[1],"tags":[2842,1321,2307,190],"coauthors":[2],"class_list":["post-4629","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","tag-algorithm","tag-discovery","tag-music-discovery","tag-spotify"],"acf":[],"yoast_head":"\n