Joris Dumas - Product Designer

Passionate about user-centered design, data-driven decision-making, and cross-functional collaboration.

My mission is to elevate, connect, and empower people by solving their problems in an elegant way.

Read the story ↓ All projects

Chapter 01 / 04

Why

Every project always starts with the ‘why’, and if we can’t answer that question, we simply move on to the next one!

Define the business goal

Le Temps · App 2030, Kick-off

We decided to meet the needs of people who now use AI-powered chatbots to get personalized information.

We held a kick-off workshop to align on a shared direction:

Evolve the “Le Temps” offer and product to engage new users and retain current subscribers who use AI chatbots to get personalized information.
Monetize this experience.
Foster an agile, user-centric company culture.

App vision project kick-off workshop board

Le Temps · Conversion funnel

We created a funnel to track the metrics that matter, quarter by quarter.

A conversion funnel, from reach to subscriber habits, that pinpoints exactly where and why users drop off, so we can prioritize what to focus on.

Conversion funnel, from discovery to retention, with drop-off metrics at each stage

Chapter 02 / 04

How

Make the right product by framing the problem through a discovery process.

Understand

Le Temps · User research

We ran interviews and observations to better understand how users get informed

Interviews

First, we ran a survey to better understand how our subscribers and non-subscribers consume news daily, then held interviews with those willing to share more, to understand the reasons behind their actions.

27survey responses

28people interviewed

Observations

We conducted observation sessions by asking users to use “Le Temps” app or website as they would in everyday life. This helped us understand how they interact with our product, their needs, and frustrations.

10people on the app

7people on the web

Le Temps · Data tools

We put in place a data-driven approach

We explored the data to surface trends that mattered and used these insights to enrich our ideation workshop with real, actionable data.

Data tools and analytics dashboard, view, click and conversion charts

Le Temps · Personas

We defined our reader personas to better identify our target audience

We created reader personas using all the insights we collected through surveys, interviews, observations, and data. These personas reflect the users consuming our product.

The main subscriber

Info-driven loyalist

I'm a discerning, well-informed news consumer who prioritises reliable sources and paid content for quality and depth. I find articles via “Le Temps”'s homepage, news aggregators and newsletters, reading mostly in the morning, on my commute and in the evening. I retain more by reading than listening, and I subscribed for the title's reputation, I'd only leave if quality dropped or prices rose sharply.

Our target

The curious reader

I'm a news consumer who likes to stay informed while being entertained. I reach articles mostly through social media and news apps, and I appreciate video and audio formats as much as text. I subscribe to a newspaper because I value quality, trustworthy information, but my commitment follows the news cycle, so I sometimes unsubscribe and resubscribe depending on the period.

Hard to catch

Free & connected info consumer

I prefer free articles and video to stay informed, and I'm not willing to pay for news. I open a free article only to dig deeper, and mostly watch video on social platforms. I care about my privacy, so I turn off notifications and avoid sharing data. Traditional newspaper like “Le Temps” leave me indifferent, I'd subscribe only if several articles genuinely interested me.

Le Temps · App 2030, User journey

We imagined high-level concepts for new ways to consume news.

The goal of mapping the user journey was to identify what works and what doesn’t for a user using our app daily. It then enabled us to imagine new usage concepts to get informed more efficiently across different contexts.

Le Temps · App 2030, Design studio

We wanted users to access personalized information while reading, anywhere and at any time.

With competitors already building AI search into their apps to deliver personalized information, our goal wasn’t to imagine similar solutions to catch up, but to envision the next step: how could our app integrate AI so users get tailored information at any point while reading, not just through search? It was a bit challenging at first, as we had no idea what this new product, capable of meeting our ambition, would look like.

Benchmark of user needs
Benchmark of what our competitors offer
Define the strengths our competitors will never have
Warm-up brainstorming
Individual poster creation

Le Temps · App 2030, Converging ideas

We aligned on a shared vision of validated concepts to prototype

Workshop boards of app concept sketches and ideas

We reviewed each other's ideas, voted, critiqued, and decided on the core directions for the app's new vision.

The pillars, in order of priority:

1. The prompt under an article as a common thread, between editorial line and the reader’s choice to dig deeper

It’s the common thread of this new way of staying informed. Each article offers prompts written as questions to anticipate the next step: the question readers would ask if they kept digging, written for them and placed at the right moment. It makes them want to go further rather than close the page.

2. The last-six-hours brief as the gateway to the news

The homepage opens with an introductory sentence that puts the day’s news in perspective, then delivers what happened over the last six hours through the eyes of “Le Temps”. Not the news from every outlet, but our hierarchy, our selection, our reading of current events.

3. A new app, a new brand, for new uses

What we’re building is new and disruptive enough not to fit within the existing product. A standalone app, with fresh branding and a brand that breaks from our traditional one, gives us the freedom to try things without being bound by the conventions of “Le Temps”.

Chapter 03 / 04

What

Make the product right with prototypes.

Prototypes

Le Temps · App, Vision 2030

We imagined an app that lets users transform information to fit their mood and available time

As we saw during the workshop, the prototype’s key feature is ready-made prompts: they let users learn more about a topic without typing a single word. These prompts anticipate users’ choices by suggesting the right question to explore a topic further, if they wish.

Then, as our user research and personas showed, the homepage and articles can switch between three formats: reading, immersive (similar to a social media feed), and listening. Users can read on the couch, scroll in immersive mode for a few tram stops, or listen while cycling.

The six-hour summary, 3 modes

The ready-made prompt system

Ready-made prompt opened with the assistant’s answer

One article, 3 modes

Le Temps · Subscription funnel

We reimagined the offers page to speed up sign-ups

We reworked the wording of the offers to make them easier to understand, and redesigned the account creation page and the order confirmation page, making the entire subscription process simpler and quicker to complete.

One of the first iterations

One of the last iterations

Le Temps · App, Audio

We have created an audio playback feature for articles

This was a response to a need that was frequently expressed during our user research: checking the news whilst doing something else.

Audio prototype

Le Temps · Design system

We made an agile design system

We first built the design system in Figma, with a core file gathering the main components across all devices, then one file for app-specific components and another for the web.

Later, we made a mirror version in Claude Design so our prototypes follow the same rules. Both versions are updated together whenever a change is made.

The components from both design systems enable the design team to design faster and developers to implement with precision.

Figma

Claude Design

Chapter 04 / 04

Evaluate

It is better to be wrong in the right way than to be right in the wrong way.

User testing

Le Temps · App, Vision 2030

We ensured our AI-enhanced content pleased our readers

We made sure readers were comfortable with AI-enhanced content and understood the new personalized homepage and article page, with features like article summaries or asking our chatbot a question about the article.

Results from tests with 8 people

Contextual personalization: a success, thanks to an adaptive AI block.

Contextual personalization, AI summary block across app screens

Flexible content formats: article summaries and AI-augmented search were appreciated.

Flexible content formats, article summary and AI-augmented search

Engaging experience: the stats page and year-in-review drew mixed interest and some tracking concerns.

Engaging experience, yearly stats, reading challenge and badges

Le Temps · Subscription funnel

We increased the conversion rate by 20%

First iterations

We first aligned on a build-your-own offer: a base digital plan, with the option to add the print paper. We tested different prototypes with more than 100 users across 7 iterations and saw that a made-to-measure system wasn't understood and didn't work for users.

Build-your-own subscription offer, base plan plus optional print supplements

Last iterations

After many iterations, we returned to a more traditional model, three clear plans, from a digital base to digital plus print, and renamed them so they'd be easier to grasp.

The result is a fluid path, well-understood offers, and an easy account-creation and checkout flow.

Three clear subscription plans, Digital, Digital & Heidi.news, and print — Select an offer

Checkout summary with optional 1-year commitment and 5% discount — Switch from monthly to yearly plan

A/B tests

We ran A/B tests comparing the previous and the new version of the offers page on a 50/50 split of users. After two weeks, the new version had increased the conversion rate by 26%.

A/B test dashboard showing variant v2 at 3.93% conversion, a 26% improvement over the original

Data analysis · Audio playback

We doubled audio article plays

With the new audio player design, playing an article in audio became far more visible. Once Play is pressed, it's much easier to see the options, and you can now build playlists.

Audio article plays, daily chart roughly doubling after the new player launch

Survey · New app launch

An app our readers love

For the launch, we ran a survey: 280+ people shared what they thought of the new app. They liked the speed and fluidity, the new features like bookmarks, audio playback, dark mode, and the more modern design. The main critique: the homepage feels too busy and dense. This will be a new challenge we’ll have to solve.

62.9survey NPS

2.6 → 4.6App Store rating, 2024 to 2026

Case studies