How to enhance users' perception of personalisation?

My Role

UX/UI Design, User Research, User Testing, Prototyping

Team

3 Product Designers

Timeline

2 weeks, 2025

Device

iOS

Context

About deezr

Founded in 2007, Deezer is a major player in the music streaming industry. Available in 180 countries, it holds a 29% market share in France, making it Spotify’s main competitor.

In a market where platforms offer similar catalogs, actors need to find ways of differanciation. Personalisation and recommendation algorithms are key areas where Deezer aims to enhance the user experience.

problem

Alexis Czornomaz, the VP Product, and his team, came with the following challenge:

How can we enhance users' perception of personalisation?

Research

To gain deeper insights into users' expectations and perceptions of music recommendations, we conducted 5 interviews with daily users of music streaming apps.

  • Users: age 28-34yo, all living in Paris

  • Format: 40-minute sessions

objectives

1.

Understand how users perceive music recommendations and what they expect from them.

2.

Understand if and how users would like to personalize their own recommendations.

3.

Understand the level of information users want about how the algorithm works.

key insights

5/5

They value and engage with personalised content.

“Playlists related to an artist I love are ideal for me. I know I’ll find songs that are relevant.” – Bilal

5/5

They want to influence the algorithm.

“I’d like to remove certain music styles from my algorithmic recommendations like reggaeton.” – Bilal

5/5

They want the algorithm to suggest new discoveries.

“It would be great to experiment with new things, not just stay in the same bubble.” – Tanguy

4/5

They want the algorithm to suggest new discoveries.

“The homepage has too much information, everything is mixed together” – Marion

how might we

How might we help users trust and shape their recommendations by making the algorithm transparent and responsive?

IDEATION

inspiration

We explored a large variety of apps in two main areas:

  • Mood tracking features

  • Personalisation components, such as sliders, filters, and multi-select options

Apple makes it very simple and intuitive to update the iPhone settings, like here for the general display.

How We Feel has an original and fun way to define and track emotions.

Foursquare shows all categories you can pick while onboarding

selected ideas

We explored various ideas to address user pain points and selected 3 key solutions based on their perceived low cost and high impact.

Customise Discovery

Enable users to adjust the level of exposure to new music genres.

Streamline the Home page

Offer an alternative display option focused on user preferences.

Shape Flow around human needs

Allow users to choose a mood, possibly combined with a music genre.

Design

Custom your Home

Tailor your Home layout to highlight your top genres, artists, or features.

Fine-Tune Discoveries

Adjust your discovery level and pick genres. Explore more with a dedicated Discovery section.

Reimagined Flow

Pick a mood or genre to launch Flow, now enhanced with a dynamic, immersive animation.

User Testing

To validate our hypotheses, we conducted user tests with the same 5 participants:

objectives

1.

Make sure users understand the Flow from the Home screen, and setting it up feels smooth and enjoyable.

2.

Make sure users understand where and how to influence the recommendation algorithm

3.

Make sure users find the “light” Home relevant

Key insights

5/5

Users grasp Flow from the Home and like its two-step process.

This is the kind of thing I like—getting what I want based on my mood." – Marion

4/5

Users value control over discovery level and type.

"The experience is really cool, and I find what I like within the genres." – Johann

4/5

Users don’t immediately notice the changes.

"Is that it? Is this the new Home with more discoveries? I was expecting a bit more explanation." – Marion

4/5

Users struggle to understand the discovery slider

"Hum.. Does this impact the new songs I will listen to? Im not sure." – Marion

3/5

Users distrust the banner and don’t fully understand it.

"I always feel like banners at the top are ads." – Johann

3/5

The header icon was not understood.

"The icon in the nav looks more like a settings mode. I would have clicked on my profile." – Tanguy

final iteration

Before

After

Clarifying page access

  • Replacing the header icon with a more intuitive option to better represent music personalisation

  • Replacing the banner with an in-page CTA, prompting users to refine their preferences

Making changes more noticeable

Making changes more noticeable

  • Adding the text “Tuning your Home” in the loading screen

  • Moving the Discovery zone higher on the Home page

  • Introducing a motion effect to highlight changes with an animation

Before

After

Before

After

Clarifying questions

  • Providing additional context for both personalisation questions

  • Adding a level indicator on the slider to improve clarity

reflection

summary

To help users shape and trust their recommendations, we designed a two-part solution:

  • An enhanced Flow experience, featuring a dynamic wave animation on the Home page to increase emotional engagement while listening.

  • A music preferences page, allowing users to personalise their discovery level and Home layout—bringing more transparency and control into the recommendation engine.

Impact

As this was a proposed initiative for Deezer, we haven’t been able to measure its impact yet. However, we would closely monitor the following KPIs to assess success:

Flow metrics

  • Average Listening Time - Measures how much users are enjoying and staying in Flow.

  • Mood/Genre Filter Usage Rate - Tracks how many users are trying the “shape your Flow” feature.

Home & Discovery Customisation metrics

  • Customisation Activation Rate - Measures how many users are personalising their experience.

  • Discovery Engagement Rate - Tells you if users are actually exploring new music.

Engagement

Retention

Adoption

next steps

As a next step, I would:

  • Run usability tests on the latest iteration to validate both UX clarity and perceived value.

  • Align with internal teams to assess technical feasibility and prioritisation of the Home customisation feature, which involves multiple stakeholders.

If I Had More Time…

If I had more time, I would:

  • Deepen user research by interviewing a more diverse range of users (e.g. lean-back vs. lean-forward listeners, non-French speakers, rural users) to uncover edge cases and unmet needs.

  • Refine visual design and motion to craft a more delightful, polished experience — using engaging animations to enhance emotional connection and stand out from the competition.

Let's drive impact together!

Let's drive impact together!

Let's drive impact together!