Social*

Adding a music-first social layer to Apple Music without breaking its simplicity

Overview

Music discovery is deeply social but almost none of it happens inside Apple Music. People find songs on TikTok, hear them from friends, and share them over text and Snapchat, leaving the app for the most engaging part of the experience.

I designed a set of four lightweight social features for Apple Music: friend listening activity, music statuses, in-app song recommendations, and "currently listening" visibility and validated the concept through two rounds of usability testing.

The project ran under a realistic product constraint framed as a stakeholder ask:

That framing shaped everything: this wasn't just designing a feature, it was building the evidence for a go/no-go decision.

"We believe a social feature would help our business grow. Could you find out and show us a prototype to see if it's worth building?"

Research →

The Gap Lives Outside the App

User interviews. I conducted 8 semi-structured interviews (ages 16–39) across Apple Music, Spotify, YouTube, and SoundCloud users, then synthesized ~50 observations into an affinity map (raw quotes, clustered into patterns, distilled into six key insights).

What mattered most:

Insight 1

Users frequently discover music through external platforms like TikTok, social media, and word of mouth, rather than directly within streaming apps.

Meet the Listeners

Maya Thompson

the discovery-driven listener

17

High School Student

Chicago Suburbs

"If I hear a song I like on TikTok, I'll go look it up and add it to my playlist."

Goals

• Quickly find songs she heard on social media

• Discover new music that matches her taste

• Stay updated on trending songs

Needs

• Easier ways to discover music her friends like

• More social discovery inside the app

• Lightweight sharing that feels natural

Pain Points

• Saving songs heard online takes multiple steps

• Rarely sees what friends are listening to in the app

• Music sharing happens mostly outside the app

• Didn't know friend activity features existed

Two personas emerged from the interviews because two distinct listeners kept appearing.

Maya represents the younger, discovery-hungry user whose music life runs through TikTok and in-person sharing. She actively wants to find what her friends are playing and does not know that social features already exist in her streaming app.

Daniel represents the established listener who values simplicity and has little patience for clutter. He will use social features only if they are optional, lightweight, and private by default.

Designing for both set the central tension of the project: enough social interaction to matter for Maya, restrained enough not to annoy Daniel. That tension resolved specific decisions. Daniel's need for privacy control is why visibility settings (Friends / Close Friends / Public) appear on every sharing surface rather than living buried in settings. Maya's preference for casual, in-the-moment sharing is why statuses and reactions stayed one tap. Where a design choice could have gone either way, the personas broke the tie.

Competitive analysis. I analyzed four competitors: Spotify and YouTube Music (direct), SoundCloud (secondary), and Instagram (indirect, because that's where music sharing actually happens). Spotify proved social demand exists but buries Friend Activity on desktop and leans algorithm-heavy; Instagram showed how music flows socially but forces users out of the streaming app entirely. The gap: no one offers lightweight, privacy-controlled, friend-based discovery inside the listening experience.

Defining the Feature Set

The research drew hard boundaries before I drew screens:

Build:

friend listening activity on Home · music status bubbles · song recommendations triggered by a friend's status · opt-in "currently listening" sharing with Friends / Close Friends / Public visibility.

Every feature also had to live inside Apple Music's existing design system for familiar patterns and native components - no new visual language.

Prioritizing What to Build First

Research surfaced more opportunities than one release should carry. I sorted every feature into three tiers based on how strongly the research supported it and how directly it advanced in-app social discovery.

Core/MVP

The MVP focuses on social music discovery, the strongest insight from the interviews. Participants consistently described finding music through friends, social media, and conversation, but those interactions almost always happened outside the streaming platform.

Friend Listening Activity with built-in preview was prioritized first because it directly supports passive discovery while removing the friction of exploring something new. Music Status Bubbles followed because they let users express interests and request recommendations without an in-app messaging system, which matched the research finding that users enjoy talking about music but prefer lightweight interactions. Visibility controls were treated as core rather than optional because several participants raised concerns about who could see their listening activity, and without that control the rest of the feature set is a non-starter.

Secondary and future features were scoped out of v1 to keep the release focused and realistic.

The most important decision in this roadmap is the last row. In-app messaging and public comments were deliberately excluded. Users consistently preferred lightweight, music-first interactions over conversation, so adding a messaging layer would have contradicted the clearest signal in the research. It was one of my favorite ideas, and cutting it was the right call.

Mapping the Flows

User flows for Currently Listening and Post a Music Status. I mapped user flows before wireframing, using standard notation: circles for start and end points, diamonds for decision points, rectangles for screens, and rounded shapes for user actions.

Two findings changed the design. First, the recommendation loop only closes when the recipient receives a notification and can preview the song, so the flow had to be designed from both sides rather than just the sender's. Second, the path from discovering a friend's song to saving it ran longer than it needed to, which pushed me to surface "Add to Library" directly in the preview card instead of requiring users to open Now Playing first.

Flow 1 - View Friend Activity

Round 1: Lo-fi wireframes, tested with 6 participants. I tested the logic before the polish: four task-based scenarios in 20-minute moderated sessions. The concept validated strongly (status bubbles were immediately understood; friend-based discovery felt natural), but testing exposed real problems: song previews didn't communicate they were playing, the "currently listening" entry point felt misplaced, and the recommendation flow lacked confirmation feedback.

Iterations - Designing for Mental Models

- Multiple entry points for the same action: recommend a song from a status bubble or the library; share your listening from Now Playing or your profile because users don't follow one rigid path

• A complete interactive Figma prototype covering all four feature flows, built entirely within Apple Music's design language

• Concept validated across 13 total test sessions in two rounds with users across competing platforms, not just Apple Music loyalists

• A defensible answer to the stakeholder constraint: yes, it's worth building as users want friend-based discovery, and the risks (clutter, privacy, social noise) are designable-around with opt-in controls and music-first interactions

The biggest lesson: a strong concept can still fail on discoverability. Users loved these features once they found them. The entire usability gap was in entry points and affordance, not in the ideas. That taught me to design for existing mental models first and novel patterns second.

It also sharpened a habit I now consider core to how I work: letting research overrule my own preferences. Cutting messaging (a feature I'd sketched and liked) because users clearly didn't want it was the most important design decision in the project.

Next steps: test long-term engagement beyond first use, explore contextual recommendation prompts, and A/B test entry-point placements for "currently listening" sharing.