Last.fm

There is a specific kind of loyalty that only comes from sunk cost, and Last.fm has been mining it since 2002. The pitch has barely changed in over twenty years: connect your music sources, and the service scrobbles — quietly logs every track played — into a permanent, searchable ledger of a listening life. Ten years in, that ledger becomes something no competitor can hand over. Spotify Wrapped is a marketing event. Last.fm is a personal archive.

Developed by Last.fm Ltd. (long since absorbed into the CBS/Paramount corporate machine after its 2007 acquisition), the service has outlived nearly every music-social experiment of its generation. MySpace pages, Turntable.fm, Rdio’s social feed, Last.fm’s own beloved streaming radio — all gone. What remains is the database, and the database turns out to be the whole business. The mobile app is the front door: it connects to Spotify, Apple Music and friends, accepts manual scrobble submissions, and renders history into charts spanning the last seven days to all-time.

Positioning-wise, Last.fm occupies an odd niche. It isn’t a streaming service. It isn’t really a social network anymore, not in the way the follower graph and taste-compatibility scores imply it wants to be. It’s infrastructure with a UI bolted on — a metadata layer and behavioral dataset that happens to have a consumer face. Rivals like Stats.fm and Volt have built slicker interfaces on top of Spotify’s API, but they can’t offer what Last.fm can: platform-agnostic history that survives a switch from Spotify to Apple Music to Tidal and back. That portability is the moat.

The market has noticed the gap between service and software. The iOS app sits at 2.3 out of 5 across 1,528 ratings, and the Android app rates worse. Those aren’t the numbers of an abandoned product. They’re the numbers of a product people refuse to abandon while actively resenting it. That distinction shapes this entire assessment.

• Scrobbling: Automatic logging of every track played through connected services (Spotify, Apple Music) or submitted manually. The core mechanic, and the reason the whole thing exists.
• Personal Charts: Top artists, albums and tracks across configurable ranges — last 7 days, 30 days, 90, 180, 12 months, all-time. Granular in a way Wrapped never is.
• Last.week & Last.year Reports: Auto-generated, shareable listening recaps. The originator of the format everyone else copied.
• Scrobble Timeline: Scroll back through any given day and see exactly what was played, in order. Weirdly emotional.
• Similar Artists & Tag Browsing: Recommendations drawn from the aggregate behavior of millions of users, plus genre and scene exploration via community tags.
• Social Layer: Follow friends, see near-real-time listening, compare taste compatibility scores.
• Artist & Album Pages: Biographies, global listener counts, tags — a functional music reference tool.
• Free Core, Paid Pro: Subscription unlocks deeper statistics and removes advertising.

The Loop That Should Be Invisible

Every passive-tracking product lives or dies on one property: it should never demand attention. The gym app that needs a tap to start counting steps is a failed gym app. The scrobbler that needs babysitting is a failed scrobbler.

Last.fm’s loop is elegantly conceived. Listen, and it logs. Weeks later the app offers a chart that says something true — that one album got played forty times in March, that “eclectic taste” is actually four artists in a trench coat. The insight is a byproduct of doing nothing. That’s the whole trick, and on paper, nothing else in music software works quite like it.

In practice, it does not reliably work.

The failure modes reported by users are not cosmetic. Tracks logged dozens of times within a few minutes, corrupting the very charts the app exists to produce. Entire albums silently missing. Extended stretches where tracking simply stops and goes unnoticed until someone checks — which, in a passive product, is the worst possible failure, because the app trains its users not to check. A dropped scrobble isn’t a bug that can be retried. It’s a hole in a record built over a decade. Manual submission exists, but asking users to hand-repair an automated archive is an admission the automation failed.

Compounding this: crashes and freezes on launch and during scrobble submission are widely reported. An app that falls over while writing to a listening history is an app that can’t be trusted with a listening history.

Where the Software Actually Earns Its Keep

Once data is in, the presentation layer is genuinely good. The chart views are clean, quick to read, and offer a range of time windows that put every streaming service’s annual recap to shame. Seven days versus all-time is a comparison that reveals something — the gap between who someone is right now and who they’ve been.

The scrobble timeline is the underrated feature. Scrolling back to a specific day and seeing the exact sequence played is closer to reading an old journal entry than checking a stat. It’s the emotional payoff for years of passive logging, and no algorithmic recap replicates it.

Discovery is more of a mixed proposition. Similar-artist recommendations are built from collaborative filtering over an enormous behavioral dataset, and they can surface things Spotify’s algorithm won’t — Spotify optimizes for retention within its catalog; Last.fm’s suggestions come from what humans actually queued next. Tag browsing is the last good implementation of folksonomy on the internet, letting people navigate by scene and micro-genre rather than by mood playlist. Both features are undersold in the mobile UI, buried behind navigation that treats stats as the main event.

The Social Layer Is a Ghost Town With Good Bones

Following friends and comparing taste compatibility works fine mechanically. The problem is population density. Last.fm’s social graph is a relic of an era when music discovery was interpersonal, and the app makes no serious attempt to rebuild it — no discovery of new users, no meaningful conversation surface, nothing to convert a compatibility score into an actual interaction. A compatibility score with no path to conversation is a dead end.

The Subscription Question

Charging for deeper stats and ad removal is defensible. Charging for it while the free tier’s core function is unreliable is a harder sell. The Pro tier asks users to pay for better analysis of data the app may not have captured correctly. Fix the pipeline, then monetize the dashboard.

The design is competent and characterless. Last.fm’s red-and-dark aesthetic has been sanded down over successive redesigns into something that resembles every other modern music utility — card layouts, generous album art, sans-serif everything. It’s legible. It’s fine. It has none of the personality of the community that built the service.

Information density is the real design failure. This is a product for people who love data, presented in an interface that spaces things out like a lifestyle app. The website surfaces more, faster. Anyone doing serious analysis of their own history will end up in a browser, which says everything about where the app sits in the product hierarchy.

Reliability is the actual problem, not the pixels. The reported crashes and freezes on both platforms cluster around the write path — precisely where reliability matters most — and the app reads as under-maintained relative to the web experience. The store ratings, especially on Android, say the same thing at scale.

Android takes the harder beating, with ratings well below the iOS app’s. Android’s aggressive background-process management is partly to blame — battery optimization can kill the listener service, scrobbling stops, and the gap goes undiscovered until later. Other scrobblers on the platform (Pano Scrobbler chief among them) handle this with persistent notifications and explicit battery-exemption onboarding. Last.fm’s official app largely doesn’t, and it pays for that in ratings.

iOS fares somewhat better at 2.3 out of 5, in part because Apple Music integration runs through a more constrained but more predictable API. The tradeoff: less flexibility, fewer third-party sources, and a stricter background execution model that makes real-time scrobbling from non-integrated apps effectively impossible.

Neither platform gets meaningful native polish. No widgets worth the home screen space, no Live Activity showing the current scrobble, no Wear OS or watchOS presence. For a product whose entire value is ambient, the absence of ambient surfaces is a strange oversight.

Last.fm is the rare product where the advice is to sign up for the service and then brace for the app used to access it. The service underneath is close to irreplaceable — twenty years of accumulated listening behavior, a tagging system no one has beaten, and a personal archive that gets more valuable every year it’s fed. That archive is the product’s real achievement.

The mobile client is not that product. It’s a thin, unreliable window onto something excellent, and the ratings tell the story with unusual clarity: a 2.3 out of 5 on iOS and worse on Android, from people who clearly keep using the thing. Nobody leaves ratings like that on software they’re indifferent to. They leave them on software they need and can’t rely on.

What’s most frustrating is how fixable this appears to be. Nobody’s asking Last.fm to reinvent music social networking. They’re asking it to write rows to a database without duplicating them, and to keep a background service alive on Android. That’s engineering work, not innovation work — and the fact that it’s gone unaddressed long enough to become the app’s defining characteristic says something uncomfortable about how much attention the mobile side gets.

We rate it a mixed proposition: essential infrastructure wrapped in software that’s hard to recommend for most people. Install it, configure it — then check on it, because it will not say when it stops listening.