Bowers & Wilkins

Case study access

Back to previous

Digital experience  ·  Information Architecture

Product Controls or Streaming Hub? Resolving the B&W App's IA Tension

Scope

  • IA research
  • Behavioural analytics
  • User research
  • Navigation design

Overview

An evidence-based investigation into how users actually engage with the B&W Music App — and what that tells us about where the information hierarchy should sit. Four research phases — product reviews, card sorting, support tickets, dashboard analytics — one consistent finding.

Duration

6 months of data
(Dec 2025 – May 2026)

1.4M+ Control interactions / month
44:1 Controls-to-streaming support ratio
2.10★ App store avg vs 4.49★ hardware
4 Research phases, same answer

The Problem

A Product With Two Competing Purposes

The B&W Music App serves two distinct functions. The first is a product controller — managing ANC modes, EQ, device connections, firmware, and headphone settings across the Px7 S3, Px8 S2, and Pi8. The second is a streaming music hub — integrating services like Tidal, Amazon Music, Qobuz and Tunein, providing a home page with personalised recommendations.

These two purposes are not necessarily the same hierarchy. How they are weighted relative to each other has implications for navigation structure, home screen design, and onboarding priority. The question this case study investigates is: what does actual usage data tell us about which surface users engage with more — and what that suggests for where the app's IA places its emphasis?

Do B&W users engage with the app primarily to control their headphones, or to listen to music through it?

The answer shapes navigation hierarchy, home screen design, and onboarding priority — making it worth examining carefully before investing in either direction.

Information Architecture Behavioural Analytics User Research Navigation Design Feature Prioritisation

Approach

The Research Journey — Four Phases, Four Reasons

Rather than starting from the analytics, the investigation was built outside-in: first what customers say in the wild, then how they think, then what breaks for them, and only then what the behavioural data shows. Each phase was chosen for a specific reason, and each fed the next.

01 · Product reviews

Why: product sales channels are a rich source of unprompted feedback. They focus mainly on hardware — but occasionally touch on the app, and those mentions reveal where the app sits in the overall product experience. 2,374 retail reviews plus 3,220 app store reviews were analysed.

02 · Card sorting

Why: mental models. For an app that combines streaming and product controls, I needed to understand how natural categorisation works for users — to validate assumptions about the existing nav structure and uncover patterns. Planned and led the activity together with the UX designer.

03 · Customer support tickets

Why: support tickets surface the biggest pain points — people contact support because something matters enough to fix. 1,762 Px7 S3 tickets (AI-categorised into 164 categories) gave a measure of the overall weight of each topic.

04 · Dashboard analytics

Why: to understand the most-used parts of the app — product controls and streaming-service interaction counts. Ten behavioural signals across every major surface, over a 6-month window (Dec '25 – May '26).


Phase 01 · Product Reviews

Loved Hardware, Resented App

The journey started in the sales channels. Retail reviews focus mainly on the hardware — but where they touch the app, the tone shifts sharply. The B&W Music App averages 2.10★ across 3,220 direct app store reviews, compared to 4.49★ for the hardware across 2,374 retail product reviews. App store reviews skew toward users who had something to report, and the dataset spans 2021–2026 so some older complaints may already be addressed. But the gap set the agenda for everything that followed: the app, not the hardware, is where the experience breaks.

↑ 4.49★
Hardware rating — retail product reviews (Px7 S3, Px8 S2, Pi8) across Amazon, Best Buy, Shopee
85.8% positive · 2,374 reviews · Jun '25–May '26
↓ 2.10★
App rating — direct app store reviews on Google Play and the App Store
68.5% negative · 3,220 reviews · 2021–2026
54.6%
of all app store reviews are 1-star — the absolute figure should be read alongside the topic distribution
1,759 of 3,220 reviews
67.6%
of app reviews mention controls or connectivity keywords — vs only 14.8% mentioning streaming services
2,177 controls · 477 streaming
↓ 2.02★
Google Play average vs 2.22★ on App Store — Android users report slightly more friction, consistent with Bluetooth complaints
1,925 Google Play · 1,295 App Store
Px7 S3 review · 1★
"The B&W app is not compatible with the headphones so all the functionality — button customisation, sound control etc — cannot be accessed. I'm surprised the product was released this way."
App Store · Germany · 1★
"Bad app, great headphones."

Phase 02 · Card Sorting

Mapping Mental Models — Users Restructured the App

Reviews told us where the pain was; they couldn't tell us how users think. For an app that combines two fundamentally different jobs — controlling a physical product and browsing streaming services — the existing navigation embeds assumptions about how those jobs relate. I planned and led two card sorting studies together with the UX designer to test those assumptions. Study one was an open sort to surface the natural categories; study two (April 2026) was a hybrid sort with 15 participants (13 valid) designed to re-validate the first study's findings and stress-test whether a three-tab model — Home / Library / Settings — holds at scale. Participants could create their own categories, including "I don't need this."

Current structure
Home 7 items
ANC control · Device connection · Recently played · Curated by B&W · Discover · Browse services · Link services
Library favourites & playlists
Favourited albums, artists, tracks · Own playlists
Search own tab
Cross-service search · AI playlist generator
Settings 20 items
Touch controls · Device info · Help · ANC settings · EQ · Auto-play/pause · Speaker grouping · Software update · Streaming quality · Add device · Rename device · Support · Email preferences · App details · Account · Legal · Bluetooth devices · Factory reset · Remove product · Link services
Participants' categorisation
Home 4 items
Notification inbox · Paired bluetooth devices · ANC settings* · Link services
Library absorbed Search
Browse services & discover content · Search for things to listen to · Build playlists with AI
Settings 18 items
Touch customisation · Device info · Help · ANC settings* · EQ · Auto-play/pause · Speaker grouping · Software update · Streaming quality · Add device · Rename device · Support · Email preferences · App details · Account · Bluetooth devices · Factory reset · Remove product
Not needed new pile
Discover other products / explore the brand · Legal documents

* ANC settings split between Home (47%) and Settings (40%) in round one — the one card users wanted in two places at once.

Where the Mental Model Was Near-Unanimous — Placement Agreement
Account management → Settings
93%
Update software → Settings
93%
Library content → Library
93%
Touch customisation → Settings
87%
AI playlist building → Library
87%
Factory reset → Settings
87%
Email preferences → Settings
87%
Search → Library
67%
% of participants placing the card in that category, round one (n=15). Admin and device-management tasks sorted cleanly; content tasks consolidated into Library.

Search lost its tab

No participant kept Search as a top-level destination. 67% placed it in Library and 33% in Home — an action inside the music space, not a place of its own. The current four-tab structure dedicates 25% of the nav to something users see as a feature of the music area.

Home was decluttered

Participants kept Home minimal: notification inbox (67% in round two), paired devices (47%), ANC (47%), and service linking (47%). Editorial and discovery cards moved out — recommendations split between Home (53%) and Library (47%), and brand-discovery content went to "I don't need this."

A rejection pile emerged

Given the option, participants used "I don't need this" deliberately: brand exploration was flagged by 4 of 13 in study two ("I'd already decided what kind of headphones I want"), legal documents by 3 of 13 ("just kind of adds bloat"). Cross-sell content triggered active aversion — "if you put ads in an app like this, you're gonna lose me pretty quickly."

Frequency drives Home, not function

The dominant mental model behind the contested cards — ANC (Home 47% / Settings 40%), quick-adjust (60/40), EQ (40/33), paired devices (47/40): Home = anything I touch every day; Settings = anything I configure once and never reopen. High-frequency controls kept drifting out of Settings into Home regardless of what they "are."

Three tabs feels too few

Unprompted, 4 of 13 vocal participants in study two invented a fourth tab — "Help," "Device Settings," "My Account," an "About" section — to relieve an overloaded Settings. The current Settings holds 20 items; participants kept trying to break it apart.

Library = streaming surface

Participants mapped Library to Spotify / Apple Music conventions — playlists, search, recommendations, podcasts, browsing. "Things I already have" landed cleanly; "discover" sometimes leaked to Home. The takeaway: don't fight the streaming-app mental model users already carry.

Study 2 · ar1994
"Even in Apple settings, you find your details in settings — that's naturally where I'd go."
Study 2 · lameloball14
"Active noise cancellation — this is home."
Study 2 · lostb0y
"These are all things I'm gonna do once and never again. They don't need to be visible in the library or home."
Study 2 · pender92
"Build playlists and discover music using AI — that's library."
IA Implication — Five Takeaways

Study two confirmed study one's headline: Settings is overloaded, Home is under-used, and high-frequency controls (ANC, EQ, paired devices) belong on Home or in a quick-settings drawer. Three tabs is the actual problem — participants invented a fourth tab but named it differently each time ("Help," "Device Settings," "My Account"); what the fourth tab should actually be only became clear in the dashboard phase. Library should mirror streaming-app conventions rather than fight them. Cross-sell content damages trust and should sit behind explicit opt-in. And users sort by "do I touch this daily," not by category — frequency beats function. That gave us a structural hypothesis to test against behavioural data: if controls anchor the mental model, usage should split the same way. The support tickets and dashboard sized it.


Phase 03 · Support Tickets

Weighing the Pain: Controls Lead by a Wide Margin

Support tickets capture the biggest pain points — people contact support because something matters enough to fix. 1,762 Px7 S3 customer support tickets, AI-categorised into 164 issue categories, gave a measure of the overall weight of each topic. Product controls and app functionality account for 524 tickets (30%). Streaming-specific issues account for 12 (0.7%).

44:1
Controls-to-streaming ticket ratio across 1,762 Px7 S3 support contacts
Controls: 524 tickets (30%) · Streaming: 12 (0.7%)
105
ANC Performance tickets — the single largest technical issue category, ahead of Bluetooth (86) and Firmware (52)
ANC is both the most-used feature and the most-complained-about
83
Spatial Audio + LE Audio tickets — users asking when firmware will upgrade the controls surface, not when streaming will improve
63 Spatial Audio · 20 LE Audio / Auracast
Top Control-Related Issue Categories — Px7 S3 Support Tickets
ANC Performance
105
Bluetooth Connection
86
Spatial Audio (demand)
63
Wear Sensor
56
Firmware Update Fails
52
Can't connect to app
50
LE Audio / Auracast
20
EQ Settings
19
Streaming Services (all)
12
Dark = active problems  ·  Light = feature requests  ·  Lightest = streaming

Streaming Model Clarity: A Small but Instructive Signal

Within the 12 streaming tickets, one is worth noting for what it suggests about IA legibility. A user's tracks were playing for only 30 seconds. The support agent explained that the app does not offer streaming services — the user had cancelled their subscriptions and didn't understand the dependency. A separate ticket asks whether Spotify will be available "soon."

Support ticket — streaming model confusion
"Customer's songs only play for 30 seconds in the B&W Music App because they cancelled their streaming subscriptions. He was informed the App does not offer streaming services and as he is not subscribed to any of them he can't hear the whole songs."
Support ticket — Spotify expectation unmet
"I have just purchased the lovely Px7 S3 headphones. I was hoping to use Spotify as my streaming service. This does not seem to be available. Can I ask if this service is soon going to be available?"
IA Implication — Streaming Model Legibility

Surfacing the streaming model clearly at the point of entry — which services are supported, and that an active subscription is required — would reduce this category of confusion. Something as simple as "Connect your Tidal, Qobuz or Amazon Music account" communicates the dependency before users encounter it mid-session. Spotify's absence should be stated upfront, not discovered.


Phase 04 · Dashboard Analytics

Sizing the Domains: What the Behavioural Data Shows

The final phase put numbers on everything the previous three suggested. To understand the most-used parts of the app, ten behavioural signals were analysed across every major surface — product control interactions and streaming-service interaction counts — over a 6-month window. Four findings emerged.

Charts are redrawn from the analytics dashboard and are illustrative of scale and trend rather than exact values.

Dashboard · Finding 01

Product Controls Dominate — by a Factor of 3×

When all control-side interactions are summed, the app's product control surfaces generate over 1.4 million interactions per month. Total streaming engagement runs at roughly 500k per month. Controls account for around 74% of all recorded app interactions, with streaming accounting for the remaining 26%.

Product Controls Streaming / Music Hub
1.4M+ / month  ·  74%
~500k / month  ·  26%

Combined monthly interactions, Dec '25 – May '26 average. Control-side: ANC (QAP + Settings) + EQ + headphone connections. Streaming-side: browse plays + homepage interactions + search selections.

1M+
ANC mode changes per month across QAP and Settings — the single largest interaction category in the app
QAP: 640–720k · Settings: 370–450k
430–560k
EQ adjustments per month — bass and treble roughly equal, confirming users are actively shaping their sound
Declining trend Dec '25 → May '26
155–195k
Headphone connections per month via QAP alone — the primary device connection pathway
Settings-path: only ~28k / month
ANC Changes via QAP — Dec '25 to Jun '26
0 200k 400k Dec '25 Jan Feb Mar Apr May Jun '26 On Passthrough Off
640–720k interactions/month. "On" dominates at ~45–50%, passthrough ~25–30%, off ~20%. Users are actively cycling through ANC modes — not setting and forgetting.
ANC Changes via Settings — Dec '25 to Jun '26
0 250k 500k Dec '25 Jan Feb Mar Apr May Jun '26 ANC changes via Settings
370–450k additional interactions/month via the Settings path — two entry points totalling over 1M monthly adjustments.
EQ Changes In-App — Dec '25 to Jun '26
0 150k 300k Dec '25 Jan Feb Mar Apr May Jun '26 Bass Treble
430–560k EQ adjustments/month, split equally between bass and treble. The 6-month declining trend (−23%) may signal users hitting the ceiling of the current 5-band EQ's capabilities.
Headphone Connections — QAP
0 100k 200k Dec '25 Jan Feb Mar Apr May Jun '26 Connections via QAP
QAP connects: 155–195k/month. The Quick Access Panel is where users connect to their headphones — not Settings.
Headphone Connections — Settings
0 20k 40k Dec '25 Jan Feb Mar Apr May Jun '26 Connections via Settings
Settings connects: ~25–31k/month. Only 1/6th of QAP volume — used for initial device setup, not regular reconnection.
IA Implication

ANC controls, EQ, and device connection generate significantly more interactions than streaming. An IA that positions these behind a music-first home screen may be misaligned with how the majority of users are currently engaging with the app. The Quick Access Panel pattern — surfacing controls immediately — appears well-suited to the observed usage profile.


Dashboard · Finding 02

Streaming Is Used — But Not for Discovery

Streaming interactions are real and substantial at ~500k/month. But the nature of that engagement matters as much as the volume. Users are not using B&W's streaming hub to find new music — they are using it to continue what they were already listening to. The home page functions as a "resume" button, not a discovery surface.

Home Page Content Interactions — Dec '25 to Jun '26
0 100k 200k Dec '25 Jan Feb Mar Apr May Jun '26 recentplay discover-feature all other modules
recentplay and discover-feature account for the vast majority of home page interactions. All other modules — recommended-playlist, cuts, linkable-services, recommended-song, sources — are marginal. Users are returning to content they already know.
Track Plays From Music We Suggest on Home Page — Dec '25 to Jun '26
0 25k 50k Dec '25 Jan Feb Mar Apr May Jun '26 recentplay all recommendation modules
recentplay generates ~85–90% of all homepage suggestion plays (37–41k/month total). Recommended-playlist, recommended-song, cuts, and recommended-album together account for under 10%.
Track Plays From Music Service Browse Pages — Dec '25 to Jun '26
0 40k 80k Dec '25 Jan Feb Mar Apr May Jun '26 Amazon Music Tidal Tunein Others
Browse plays run at 150–185k/month. Amazon Music, Tidal and Tunein account for ~80% of plays. Users are accessing services they already subscribe to — not adopting the B&W hub as their primary listening environment.
IA Implication

The streaming surface's job in the IA appears to be providing a convenient path back to content users already know. Discovery modules receive comparatively little engagement — which may suggest deprioritising or repositioning them, though further research into their placement and visibility would be useful before drawing firm conclusions.


Dashboard · Finding 03

Two Distinct User Profiles Within the Streaming Surface

Search behaviour reveals that not all streaming users are alike. Browse plays and search selections show different service distributions — pointing to two separate user profiles with different needs and different IA requirements within the streaming layer.

Selecting an Item From Search Results — Share by Service, Dec '25 to Jun '26
Tidal 45% Qobuz 25% Amazon Music 15% Tunein 10% Other 5%
Tidal (~45%) and Qobuz (~25%) dominate search — a markedly different profile from browse plays where Amazon and Tunein lead. Search users are audiophile-oriented, looking for specific content on high-fidelity platforms.
🎧
The Returner
Majority — casual / habitual
Opens the app to reconnect a device, adjust ANC for a new environment, and get back to what they were listening to. Does not browse for new music inside B&W. Relies on recentplay and their own streaming app.
Primary surfaceQAP / Controls
Streaming behaviourrecentplay only
Preferred servicesAmazon / Tunein
🎼
The Audiophile Seeker
Minority — intentional / high-fidelity
Opens the app to find specific content on Tidal or Qobuz and fine-tune the listening experience. Uses EQ actively, searches for artists or albums, and expects the app to function as a genuine hi-fi control surface.
Primary surfaceEQ + Search
Streaming behaviourActive search
Preferred servicesTidal / Qobuz
IA Implication

The IA does not need to choose between these two profiles — but designing the default state for the Returner (fast access to controls + recentplay) while keeping the Audiophile Seeker's path to EQ and search short would serve both. A tiered navigation structure — controls first, curated streaming secondary — appears well-matched to the observed usage split.


Dashboard · Finding 04

A Sharp Usage Decline — and the Controls–Streaming Relationship

Overall app usage declined significantly between April and May 2026. The steepness of the drop — particularly on the Settings surface — suggests a specific event may be involved: a firmware update, a compatibility issue, or an app store release. What the data does show is the relationship between the two surfaces when usage drops: controls and streaming fall together, with controls leading.

App Usage by Surface — Dec '25 to Apr '26
0 1.5M 3M Dec '25 Jan '26 Feb '26 Mar '26 Apr '26 Settings Browse Player Library & Search
Settings dominates all surfaces at 2.4–3M interactions/month. Browse holds a distant second at 900k–1.25M. Player, Library, and Search remain near-zero by comparison. The hierarchy is clear: users live in Settings and Browse — everything else is marginal.
The Fourth Tab, Resolved

Card sorting showed participants reaching for a fourth tab without agreeing on what it was. The dashboard answered it: Settings is the highest-interacted surface in the app, and the bulk of that traffic is product management — device info, connections, renaming, adding and removing products. Users can own multiple speakers and headphones, just headphones, or just speakers — which points to a dedicated Devices tab: a product listing showing all owned products, separating "manage my products" from app-level settings.

↓ 73%
Settings surface decline Apr → May '26. Cause unconfirmed — possibly firmware, compatibility, or a release event
↓ 72%
Browse surface decline in the same period — both surfaces fall together, with controls leading the drop
Review Corroboration

Of 125 critical product reviews (1–3★) over the trailing 12 months, the top complaints are Bluetooth/connectivity (39), firmware failure (10), app compatibility (9), and customer service (13). Streaming-related complaints total 1 in the same window. Product reviews also show negativity rising to 29% in March 2026 — approximately one month before the usage decline in the data.


Recommendation

The IA Hierarchy the Data Supports

Taken together, the four research phases suggest a three-layer hierarchy, delivered through a four-tab navigation: Home / Library / Devices / Settings. Devices is the tab card sorting asked for and the dashboard named — a product listing for every owned speaker and headphone, pulling product management out of an overloaded Settings. Reviews, mental models, support weight and behavioural data all point consistently in the same direction, though the right balance would benefit from further user testing to validate.

Primary
Product Controls — ANC, EQ, Device Management
The default landing state. Quick Access Panel as the home surface: connected device status, ANC mode toggle, EQ shortcut, battery, firmware state. This is what 74% of app interactions are already doing — the IA should reflect it.
1.4M+ interactions/month Both user profiles Drives retention
Secondary
Continue Listening — recentplay + linked services
One tap from the primary layer. Surface the last-played content with a direct play action. Link to Amazon, Tidal, Tunein and Qobuz — the services users are actually using — rather than B&W's own editorial curation. This is what recentplay's ~85–90% share of homepage plays demands.
~37–41k plays/month from homepage Returner profile Low editorial investment needed
Tertiary
Browse & Discover — service browsing, search, editorial
Accessible but not prominent. Full service browse pages (Amazon, Tidal, Qobuz, Tunein), search with service filtering, and optional editorial recommendations. The Audiophile Seeker finds this valuable; the Returner ignores it. Critically: the entry point must make the dependency model explicit — "Connect your Tidal or Qobuz account" — so users understand it's a passthrough, not a B&W-owned service. Spotify's absence should be surfaced here, not discovered mid-session.
150–185k browse plays/month Audiophile Seeker profile Dependency model must be legible No Spotify — state it upfront
What This Is Not

This hierarchy does not deprioritise streaming as a feature — it deprioritises the current home screen's over-investment in discovery modules that users have already demonstrated they do not engage with. B&W's streaming integration is real and valuable; it just belongs in the second and third layer, not the first.


Concept · Wireframes

The Four-Tab IA, Sketched

Early wireframes translating the hierarchy into screens. Each surface is zoned by priority — what the research says users touch most sits highest. The bottom navigation carries the four tabs the studies converged on: Home / Library / Devices / Settings.

Home

1st priority

Notification inbox
Product status — battery · codec
ANC
EQ
Spatial
Multipoint
Product controls entry →

2nd priority

Resume — last played ▸
MSP entry
+ Link new MSP

3rd priority

B&W playlists
TBD
HomeLibraryDevicesSettings
Library

1st priority

Search
Recently played
Playlists (new)

2nd priority

Discover
MSP
Link new — Tidal · Qobuz · Amazon
(subscription required)

3rd priority

TBD
HomeLibraryDevicesSettings
Devices

1st priority

Headphones

2nd priority

Spaces / speakers

3rd priority

Product help & accessories — for owned products

4th priority

TBD
HomeLibraryDevicesSettings
Settings

App-level only

Account
Codec info
App info
Warranty
HomeLibraryDevicesSettings

Home Adapts to What You Own

Because ownership varies — multiple speakers and headphones, just headphones, or just speakers — Home is composed per ownership profile rather than fixed. The hierarchy stays the same; the first-priority zone changes.

Headphones only

The first section is all about the product itself: connection status, battery, ANC, EQ, Spatial Audio, and Multipoint, with a clear route into the full controls screen.Streaming comes next.

Speakers only

The structure stays the same, but the controls adapt to the product. Users see speaker status, EQ, and a link to additional controls. Features that don't apply, like ANC or Multipoint, simply aren't shown. Streaming remains second, with playlists and content third.

Headphones + speakers

For users with multiple products, the primary headphone card appears first, including ANC, Multipoint, and access to full controls. Additional products sit one swipe away in a card stack. Streaming remains second, followed by playlists and content.

Why it works

The layout is based on what people use most, not how features are grouped internally. The controls users interact with every day are always first. Streaming comes next, as the second most common activity, while playlists and editorial content sit further down the page. The result is a homepage that changes depending on what products a user owns, while keeping a familiar structure and prioritising the actions people use most often.

Three Decisions Layered on Top

A global quick-settings drawer

ANC split Home (47%) / Settings (40%) in card sorting — users want it in two places at once. Rather than picking a side, ANC / EQ / connection live in a pull-down drawer reachable from any tab. The dashboard already validates the pattern: the QAP path beats the Settings path 6:1 for connections.

Cross-sell that feels relevant

Both studies showed that "Discover other products" was perceived as promotional and reduced trust, so we removed shopping as a primary destination. Instead, accessory recommendations appear within the context of products users already own. Empty states in Devices provide a more relevant way to introduce the wider ecosystem. For example, someone who only owns headphones might see an invitation to add a speaker, but only when it's contextually relevant and entirely optional.

Setting expectations around streaming

The support tickets showed users discovering mid-session that the app streams nothing itself — 30-second clips after a cancelled subscription, Spotify expected but absent. The link-service tile names the supported services and the subscription requirement at the point of entry, not after. This approach helps users understand that the app is a streaming gateway rather than a content provider.

Validating the structure

The next step is to tree-test the proposed four-tab architecture against the current three-tab structure and compare behaviour after launch against the Dec '25–Jun '26 baseline. Success would look like key controls being accessed more naturally through Home and Devices rather than Settings, alongside improvements in overall app satisfaction and a narrower gap between app ratings and the hardware's 4.49★ rating.


Reflection

What I've learned so far

Users tell you what matters through behaviour Leading with music in a headphone app feels like the obvious choice for an audio brand. But the evidence consistently pointed elsewhere. Users primarily come to manage their headphones — adjust ANC, change settings, and control their listening experience. The information architecture should reflect that reality rather than assumptions about what users might engage with.
The answer rarely comes from a single metric.No individual signal was strong enough to justify a recommendation on its own. Controls usage showed the scale of engagement. Recent-play activity revealed what streaming meant in practice. The decline between April and May exposed a dependency between experiences. The search-versus-browse split highlighted two distinct behaviours hidden within the same dataset. The recommendation only became clear when all signals were considered together.
Different methods, same conclusion. Product reviews, support tickets, card sorting, and behavioural analytics approached the problem from different angles, but they consistently pointed in the same direction. Users engage with controls more frequently than streaming features, and they tend to think of headphone management and music consumption as separate activities. That consistency across methods provided confidence in the recommendation.
Information architecture shapes more than navigation. The recommendation affects far more than menu structure. It influences what appears on the home screen, where investment should be prioritised, and how the overall product experience is framed. If controls are the primary reason users open the app, the experience should be designed around that behaviour first, with content and discovery supporting it rather than leading it.