Music taste broadens in youth but narrows with age

Music feels personal at any age. Yet if your feed suggests artists that do not stick, you are not alone.

A new analysis of long-running listening histories finds that our taste gradually narrows with age. Younger people sample widely, but by midlife many of us return more often to familiar tracks.

Professor Alan Said and his team at the University of Gothenburg (GU) analyzed 15 years of data to show how listening behavior changes with age, linking it to shifts in variety and personalization.

The team drew on Last.fm histories that connect plays from multiple services into one profile. Because many users self-report age at signup, the dataset allows age-linked modeling across years.

The backbone for such research is the LFM-2b dataset. It includes over 2.014 billion listening events from more than 120,000 users between 2005 and 2020, plus track and user metadata that enable careful analysis.

Music taste changes with age

Psychology has long described a reminiscence bump in autobiographical memory for music from adolescence and early adulthood.

Research shows that recognition, preference, and emotion all peak for songs from those years, and even reveals a secondary peak for the music of one’s parents.

Consumer behavior work adds another angle. Classic studies note that favorite popular music often reflects tastes shaped in late adolescence and early adulthood, a pattern that helps explain why songs from those periods feel enduringly salient.

The new findings fit these patterns without reducing older listening to pure nostalgia. Many older listeners still try new artists, yet they circle back more frequently to the songs that mattered during youth.

Teenagers tend to share many favorites with peers. With age, overlap falls as tastes become more individualized.

Music journeys grow more personal

The authors describe a broadening from adolescence into early adulthood, then a contraction into a more personal profile later in life. That turn toward personal history does not stop new discovery, but it shifts the balance toward repeat plays of established favorites.

The age-linked analysis of more than 40,000 Last.fm users, over a 15-year period, counted 542 million plays across more than one millions songs. That is enough data to trace trends across the life course. These numbers explain why the conclusions carry weight.

The team identified a clear pattern among older adults. “Most 65-year-olds do not embark on a musical exploration journey,” noted Professor Said.

The same data show that middle-aged listeners keep engaging with new releases. They just return to earlier personal favorites more often than younger listeners do.

Age and music recommendations

The study highlights the importance of recommender systems that adapt to life stage. A single strategy for everyone risks missing what different groups actually want, especially when older profiles reflect unique, deeply personal histories.

Age-aware systems can handle exploration and familiarity differently at different times of life. For younger listeners, novelty may play a larger role, while for older listeners, models can weight long term favorites more heavily and still surface fresh material that fits established preferences.

Long window evaluation also matters for system design. Many benchmarks sample a short slice of time, but the longitudinal arc of listening reveals patterns that short-term tests cannot capture, including how variety and uniqueness shift across decades.

Datasets with age and multi-year histories make it possible to evaluate these choices properly. They also open doors for fairness and inclusivity work, since demographic signals can expose where models overfit to a narrow slice of users.

Building your lifelong soundtrack

Younger readers can embrace the broad sampling that comes naturally. Variety now builds the library you will draw on later, and it can sharpen your sense of which artists feel most like you.

If you are in midlife or older, leaning into your personal canon is not a problem. You can still try a measured stream of new tracks that mesh with your history, then save the ones that land so your recommendations keep learning.

The industrial takeaway is simple. Recommendation teams should test strategies that mix new and familiar differently by age, and they should measure success on multi-year horizons.

The human takeaway is just as clear. A personal music soundtrack tightens with age, but it remains alive, and careful design can help it stay that way without forcing change.

The study is published in Adjunct Proceedings of the 33rd ACM Conference on User Modeling, Adaptation and Personalization.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–