fizy, Turkcell, 2021

Bringing the fizy web player back

Everyone at fizy knew the web player needed work, and everyone agreed to fix it later. Later kept moving, and it slowly turned into a different product from the mobile app. So I got the team behind a smaller plan: everyone made a little room in their sprints, and we started the redesign piece by piece, until both felt like one fizy again.

Role
Senior Product Designer
Team
Two lead product managers, a product owner and a frontend developer
Timeline
2021, about six months
Scope
Web player, next to the mobile app
The redesigned fizy web player: artist page, discovery homepage and album page
My role
The redesign was my proposal. I led the design from the first audit to the last screen and built its design system.
Problem
One action had several button styles, colors followed no rules and four font families were in use.
Impact
In six months, the web player moved onto the mobile app’s design language.

Overview

The web player had fewer users than the mobile app, so the effort always went to mobile. But during the pandemic more people listened on their computers, and with no desktop app, the web player was fizy’s only place there.

Problem

Always fixed later

The web player was full of inconsistencies: one action had several button styles, colors followed no rules and four font families were in use. It looked and worked differently from the mobile app, so fizy didn’t feel like one product. And every new feature had to be designed and built twice.

Audit of the old web player: one action shown in many button styles, and four fonts in use: Brandon Text, Gotham Rounded, Nunito and Poppins

Approach and rationale

Small steps, agreed with the team

Start from what is live

The design files were out of date, so I mapped the live product screen by screen.

Make the problems visible

A shared list of inconsistencies was easier to discuss than opinions.

Prioritize with the team

Product, developers and I agreed together on what to fix first.

Reuse what works, keep it small

The web player took on the mobile app’s design language, with no new features, so the work fit into each sprint.

Execution

Mapping the product first

The user flows and the full page list showed which patterns repeat, so fixing one pattern fixed it everywhere.

FigJam board with the page list, the as is flow, the design audit and the component architecture

Decisions 01

Homepage

The web player had been designed for screens far smaller than the ones people actually used. We checked the screen sizes in our listener data and set the new layout at 1920 by 1080, then redesigned every element on the page to bring it in line with the mobile app.

Homepage before the redesign Homepage after the redesign
  1. Every item was far too large for the screen, with no clear hierarchy
  2. Different components didn’t look different from each other
  3. The menu and labels didn’t match the mobile app
  4. The homepage didn’t continue the experience people knew from mobile
  1. A modular homepage, the same as on mobile, so listeners switching devices find their way at once
  2. A hero carousel for campaigns and new releases, so new music is the first thing people see
  3. Each content type gets its own card, so playlists, albums and podcasts are easy to tell apart
  4. Rows built for exploring, so people find more to play beyond what they already know

Decisions 02

Album page

Side by side, the album and playlist pages didn’t match: layouts, button sets and actions all differed. Once people started listening, they couldn’t predict what the next page would look like, so every listening page now shares one structure.

Album page before the redesign Album page after the redesign
  1. Too many buttons with no hierarchy, so people didn’t know where to start
  2. A full width banner pushed the tracks down, out of sight
  3. The hero and the player took so much space that the music was hard to reach
  4. Dense track rows made picking a song slow
  1. Play, shuffle, add and more options in one card, so people know what to do first
  2. Clear differences between elements, so the eye goes to the music, not the chrome
  3. A track list with clear spacing and durations, quick to scan and pick from
  4. The same album page as on mobile, so nothing has to be learned twice

Decisions 03

Artist page

The artist page now follows the same structure as the playlist pages and the mobile app. Popular songs, albums, playlists and similar artists are arranged to lead people from discovering to listening.

Artist page before the redesign Artist page after the redesign
  1. The same visual problems as the album and playlist pages
  2. Follow and play were easy to miss, so the main actions went unused
  3. Albums, playlists and videos stacked without order, with no clear place to start
  1. Every action sits in one card, the same as on albums and playlists, so the page feels familiar
  2. Popular songs up front, so people can start listening in one click
  3. The same structure as the mobile app, so moving between devices feels seamless

Decisions 04

Search

Search used to show a plain list of past queries. Now previous searches come with artwork, and popular searches fill the screen before people start typing.

Search before the redesign Search after the redesign
  1. A plain list of past searches in large text, slow to scan
  2. No artwork, so results were hard to recognise at a glance
  3. Nothing to discover before typing, so search felt like a dead end
  1. Previous searches with artwork and type, so people recognise what they want before reading
  2. Popular searches before typing, so there is always something to play
  3. History in one place, easy to clear when people want a fresh start

Decisions 05

One system behind every screen

I started the system on my own initiative, so the web player and the mobile app could share one language instead of drifting apart again. I set the foundations, defined every component’s states together with the frontend developer and documented them, so screens could be built straight from the system.

Pages from the fizy web design system: colors, actions, cards and side menu navigation
  • Foundations

    A shared color set, one typeface, border widths, corner radii and a 1920 by 1080 layout.

  • Actions

    Buttons, icon buttons and text buttons with all their states, designed around every need in the product.

  • Cards

    Most of fizy is made of cards, so every card type and its states became part of the system.

  • Navigation

    Tabs, menus and popup menus, next to the side menu.

  • Player

    One play bar with every playback state, shared with the mobile app.

  • Iconography

    One icon set in fixed sizes, matched to the mobile app.

  • Frameworks

    Inputs, dropdowns and the other elements behind forms, lists and selection pages.

  • Banners

    A rule set for automated banners whose colors and layout follow the photo.

Outcome

One web player, one design language

The web player shipped on the same design language as the mobile app, and fizy finally felt like one product on every screen. The playlist page shows the change best: the same content, now in the mobile app’s layout.

Playlist page before the redesign Playlist page after the redesign
  1. The web player had its own layout, colours and type
  2. Next to the mobile app, it looked like a different product
  3. Each screen was designed and built on its own
  1. The same playlist page as on mobile, so listeners feel at home on any device
  2. One type family and one colour set, so actions and states read the same everywhere
  3. Built from shared components, so new features reach web and mobile together

Learnings

What I’d do differently

We never agreed on how to measure the redesign, so I can show what shipped but not how listening on the web changed. Next time I’d set success metrics with the product team before the work starts, and begin the documentation sooner, so developers could build from the system from the very first sprint.

Hiring a senior product designer,
or want to talk design systems?