The Spotify Taste Profile hands listeners a dial, and artist marketing must learn to answer it

What the Spotify Taste Profile actually shipped#

The fox does not chase every rustle in the hedgerow. It stands still, sniffs the wind and works out which trail the traffic will take next. That patient prowl is the right posture for the Spotify Taste Profile, because on 23 September 2026 Spotify began rolling out a beta to US Premium listeners over 18, after a spell of testing in New Zealand, as Fast Company reported.

Here is the mechanism in plain words. A listener taps their profile picture and opens a tab that shows a short summary of how Spotify reads their taste across music, podcasts and audiobooks. Beneath it sits a prompt called Tell us more, where they type what they want more or less of, as Hypebot walked through. Those notes then steer the Home feed, and the Spotify algorithm behind spotify recommendations reads them alongside everything a listener already plays, skips and saves.

Spotify says changes typically appear within a few hours, and that a listener can edit or delete a note whenever they like. The idea is not new to the boardroom either: co-CEO Gustav Söderström unveiled it at SXSW on 13 March, and the US launch is the first time Spotify's largest market gets to try it.

• 13 Mar

  when Söderström showed the idea at SXSW

• 23 Sep

  when the US Premium beta began

• 80%+

  of listeners value personalisation, says Spotify

• 48 hrs

  the longest lag Spotify quotes for a separate kids-and-family exclusion

Now the limit, which is the whole argument of this piece. Spotify says the tool cannot follow an artist on someone's behalf and cannot enforce a rule such as always show or never show. A note is a nudge, not a lock. That distinction decides where an artist team should spend the next quarter.

The number Spotify leads with is also worth a moment. The company says more than 80 per cent of listeners rank personalisation among the things they value most, in the Spotify newsletter announcing the expansion. That is a statement of preference from the platform that built the Spotify Taste Profile, not an audited finding, so read it as motive rather than proof. The motive is clear enough: a listener who can correct the feed is a listener who trusts it.

Why a follow still outranks a typed note#

Start with the route that is guaranteed. Spotify's own artist guidance says that if you have a million followers, it will automatically add your new music to a million playlists on release week. That is a deterministic promise. The Spotify algorithm may guess, weigh and reorder, but a follow delivers a release to a named listener without any guessing at all.

Compare that with Release Radar. Spotify's support page says the order of songs there rests on factors like release date and how much Spotify thinks the listener will like the song, judged from their past listening. The Spotify Taste Profile adds a stated signal beside that history. Hypebot's reading is that a listener can now state an interest before their listening history reflects it, which matters most to the artist whose sound is new to that listener.

So the honest ranking runs like this. A follow is a promise. A saved song is a strong hint. A typed note is a soft steer that Spotify says it cannot enforce. Each has a job, and a tidy campaign asks for the follow at the moment a listener is most delighted, because nothing else in the toolkit is as dependable.

Discovery Mode is a number to test, not a number to trust#

Spotify's Discovery Mode page says artist teams who select a song within the first month of eligibility have seen on average a 121 per cent rise in listeners for that song in its first month. That is Spotify's own figure, about Spotify's own programme, and the line quoted gives no control group. It is a reason to run a test, never a reason to skip one.

The pool is sourced from Spotify's Loud & Clear 2026 highlights, and the discovery door is open every month. The Loud & Clear FAQ says an average subscriber listens to more than 200 different artists a month, and nearly half of those are artists they are hearing for the first time. Room for newcomers is not the problem. Being describable, findable and memorable inside that room is.

One practical consequence follows for anyone still asking how does the Spotify algorithm work. It works on signals, and a note is now a new, human-written signal among many. The dial changes who can nudge, not the basic craft of making music a listener wants to keep hearing.

Will the dial help or hurt an emerging artist?#

Nobody outside Spotify can say what the Spotify Taste Profile will do to newcomers with data yet, and the fox distrusts anyone who claims to. What exists today is argument. In March, Hypebot's commentary on the announcement worried that letting listeners filter their own tastes could make artist development more restrictive. The worry is fair: if a listener types more of what I know, the sideways lane between genres narrows.

The counter-argument is that steering can also widen a lane. A listener who types that they want more contemporary jazz, or music for an early run, is telling Spotify something their history could not. Both readings are plausible, and both are opinion until someone publishes a measurement.

Academic work offers a little calm. A RecSys 2024 user study of music recommenders found that two ways of promoting less popular items did not harm listener satisfaction. The sample was 40 people and the study was not about Spotify, so treat it as a hint that fairness and satisfaction can coexist, not as a forecast. A separate arXiv paper on controlling recommenders with natural language tested the idea on films rather than music, which is precedent rather than proof.

That listener's suspicion is a useful weather vane. Whatever Spotify intends, some listeners will read a feature like this as a trade of data for control, and some will simply enjoy the control. Artist teams do not need to settle that argument. They need to remember that the listener, not the platform, decides whether a note is worth typing.

Regulators have been circling the same idea for years. Article 27 of the EU's Digital Services Act requires online platforms that use recommender systems to set out the main parameters in plain and intelligible language, and the European Commission's overview notes that on very large platforms users can opt for non-personalised feeds. Whether and how those rules bind Spotify is not something this piece claims. The direction of travel, towards visible, adjustable recommendations, is clear enough.

Five moves for artist teams, starting this quarter#

None of the five needs a new tool or a new budget line. All of them survive whatever Spotify does next, because they rest on craft rather than on a feature that may change shape in a month.

The first move is the cheapest and the most neglected. A listener typing a note uses ordinary language, not industry taxonomy. If your best track is described in a press release as a genre-defying sonic journey, no listener will ever type those words. If it is described as slow, warm guitar music for a wet Sunday, someone might.

The second move is the plainest. Spotify's own guidance says a million followers means a million playlists on release week, so the follow remains the sturdiest asset an artist owns on the platform. Build the ask into the show, the credits and the first reply to every message, and count follows as a growth metric in its own right.

For a wider view of how listeners have already moved from browsing to asking, read folkfox's piece on conversational music discovery, and the note on AI music marketing and the licensed fan.

Measuring a dial you cannot see#

The awkward truth is that an artist team cannot see the notes listeners type. It can only see the downstream shadow in follows, saves and repeat listens. Measurement therefore has to move upstream to conversion and forward to a slow, honest monthly rhythm.

That last bullet deserves a beat. It shows that Spotify already lets listeners rub out parts of their history so they stop shaping Spotify Discover Weekly and its cousins. The Spotify Taste Profile extends the same idea from removal to request. The direction is consistent, and it points toward a listener who edits their own feed with a light hand.

Set the baseline before the dial reaches your audience, because a shift you cannot show is a shift a manager will not fund. Snapshot follower conversion, saves and source-of-streams shares for your last three releases, then repeat that sweep monthly. The fox does not guess the depth of a stream, it puts a paw in first.

Expect the shape to change before the size does once the Spotify Taste Profile reaches your audience. Follows and repeat listens may move quietly while headline stream counts barely twitch. That is not failure, it is the new floor, and measuring against the old one will talk a team out of work that is genuinely working.

If you would rather have the measurement rhythm built than described, that is the work of folkfox music industry marketing, backed by content marketing that describes a release in the words a listener would use, and SEO and GEO that keeps the catalogue findable wherever the question is asked. The same discipline runs through brand strategy, because a describable artist is a strategic decision before it is a caption.

Frequently asked questions#