WEEK.
From artist analytics to a shareable story.
Continued design work on AMP Highlights, shaping a shipped sharing flow that helped artists and managers package verified performance stats without stitching together screenshots by hand.
- Role
- UX/UI Designer
- Timeline
- Aug 2022 - Feb 2023
- Platform
- Analytics product across web and mobile surfaces
- Nova ReyesOct 7–13, 2022Oct 7–13A BIG
WEEK.Total global streams24.3M+11%vs prior week8 completed weeksPrior: 21.9MAMPHighlights - Nova ReyesOct 7–13, 2022Oct 7–13Nova Reyes / SingleTRACK
HIGHLIGHT.↗Most streamed trackGLASS HOURS6.1Mstreams last weekTrack · all countries↑ 18%AMPHighlights - Nova ReyesOct 7–13, 2022Oct 7–13ON THE
PLAYLISTS.Playlist placements · global4↑ 3Glass HoursLate Light · 2.1M followers9NEWLow TideSlow Burn Pop · 860K followers15↓ 2Glass HoursIndie Drive · 640K followersThree placements across three playlists.
Movement relative to the prior week.AMPHighlights
Select a package, then see what will be shared
AMP is Warner Music Group's analytics product for artists and managers. Users were already sharing performance data publicly by screenshotting views, cropping them by hand, and posting them outside the product.
The work was to decide what belonged in a shareable package, how much of it should travel together, and how to make the result feel at home in the places artists already engaged fans.
Stakeholder interviews, benchmarking, and user validation pointed to Instagram Stories as the primary destination, which shifted the problem from generic export to story-first packaging. Competitive review of year-in-review and social share patterns added constraints: fast publishing, minimal save friction, and multi-card bundles instead of one isolated stat.
How the package took shape
I facilitated a site-map and user-flow workshop with Product Owners and design leadership to map how Highlights related to the rest of AMP. That work shaped the information architecture and where the Highlights page sat after launch.
The early problem was structural: how cards should stack, what hierarchy they should follow, and how the flow would translate across devices. I used block wireframing in Axure to move quickly through layout options while staying grounded in implementation constraints from Flutter and the existing product.
I cross-referenced the Product Owner's metrics spreadsheet with usability feedback to decide which metrics to group and how those groups should read when shared publicly. That produced category logic across items like Top Countries, Top Songs, and Top Demographics, plus color and icon rules that made the set easier to scan and assemble. Prototypes were tested in responsive flex containers and built into a clickable flow for the next validation round.
- 01 / Inputs
Start with the information.
The Product Owner’s metric inventory, cross-referenced with usability feedback, decided how analytics should be grouped for sharing.
Category examples recorded in the caseTop SongsTop CountriesTop DemographicsProduct Owner · Metric inputs - 02 / Structure
Make the package reviewable.
Axure block wireframes focused discussion on card order and hierarchy. Responsive, clickable prototypes explored the flow within product and Flutter constraints.
Structure schematicArtist + periodPerformanceUnits + scope + comparisonProduct Owners + design leadership · IA workshop
Engineering · Layout feasibility - 03 / Validation
Refine what people can share.
Four moderated remote tests surfaced category control, multi-card story expectations and different KPI priorities by role.
Final sharing directionChoose → Preview → Download12345UX Research · Validation partnership
Research and validation
The feature was shaped with a dedicated research team. We built a research roadmap that combined targeted surveys with rolling research sprints to keep testing what users wanted to share and how AMP fit their working habits.
“I like the carousel idea, but I'm not sure I would want to share these categories. Is there a way to customize or change what's being shared?”
“On Spotify, you can share an entire story, which includes several cards and statistics. Is there a way you can help us do this?”
In four moderated remote tests, users liked the carousel idea, but the initial categories and data accuracy felt too rigid. Some wanted more control over what got shared. Others referenced Spotify-style story packages that bundled several related cards. Relevance also changed by role: artists, managers, and label-side users cared about different KPIs, so one undifferentiated package created noise.
The feedback changed the direction in three ways:
- clearer labeling and aggregation rules for cards
- stronger role-based filtering of what belonged in a package
- less reliance on preselected assumptions about which KPIs mattered most
A card system with enough range
The work included metric grouping, color and icon rules, responsive prototypes, and implementation work with engineering. The contemporary figures below make those information rules inspectable with a reusable visual system.
In this presentation, a stream milestone, a track release and a playlist ranking share context rules rather than one repeated composition. Typography and code-rendered charts carry the information; artist photography and album artwork are optional.
WEEK.
- What travels with every card
- Artist
- Nova ReyesOn the rail of every card.
- Reporting period
- Oct 7–13, 2022One completed week, stated in full.
- Metric
- Total global streamsNamed above the number, not implied by it.
- Unit and comparison
- 24.3M · +11% vs prior weekThe chart’s prior week is derived from the stated change.
- Scope
- All countriesA country filter adds a scope chip to the card.
HIGHLIGHT.↗
HIGHLIGHT.↗
Most streamed track · Release layoutTitle over 14 characters: set in Geist instead of the wide display face, so a long title wraps cleanly instead of shrinking to fit.
HIGHLIGHT.↗
HIGHLIGHT.↗
Most recent track · Release layoutReleased this week: there is no prior period, so the change chip is omitted rather than shown as zero. Long title falls back to Geist.
WEEK.
WEEK.
Total global streams · Milestone layoutDown week: the headline goes neutral and the change chip loses its highlight. The card never celebrates a drop, and the line still shows it.
All six content choices, in source order
WEEK.
LISTENING.
Unique listeners · all countries
HIGHLIGHT.↗
+6% vs prior week
HIGHLIGHT.↗
PLAYLISTS.
Movement relative to the prior week.
Up to 5 of these six can be selected for a reel. Six choices use five layout families.
Keep the data context visible at both widths
Highlights packaged stats from AMP's analytics views, which ran on desktop and mobile widths.
The comparison below keeps one artist, market and week in view while the layout reflows, so the responsive behavior can be read without resizing the site.
A future direction for analysis and sharing
Concept · Did not ship
A future concept explored showing Highlights cards directly within AMP's data pages, reducing the separation between analysis and sharing. It did not ship.
Highlights
Pick the moments.
Keep the context.
Highlights cards shown beside the data they summarize.
WEEK.
LISTENING.
Unique listeners · all countries
HIGHLIGHT.↗
+6% vs prior week
HIGHLIGHT.↗
PLAYLISTS.
Movement relative to the prior week.
Delivery
The final design let users share up to five cards in a single story post, with the data rearranged around role-based priorities and redundant information cut. I worked with engineering on component and flexbox considerations so the design held up in implementation.
Highlights shipped as an in-product social sharing workflow in AMP, replacing manual screenshot-and-crop with a story-ready package of verified stats. The research gave the team a clearer basis for which analytics to share, how to group them, and how role should shape the package in future iterations.
Post-launch, the team tracked engagement and time-on-page analytics in Amplitude and organized future enhancements in Jira.
Supporting evidence
The original card board, site map and AMP source context sit behind the reconstruction. The primary figures above are labeled reconstructions; these originals are supporting receipts, not substitutes for readable interaction evidence.
Have a complex
product problem?
If you’re hiring for senior, hands-on product design, I’d like to hear about it.
brianmartinux@gmail.com