Your Personal News AI

Lumi News

Using data science, machine learning and natural language processing to recommend the best news, stories and videos based on the things you read, stories you like, the topics and publishers you follow, the people you follow and from your friends across social networks.

gesture based reading experience, using data science that learns from the things you enjoy.

“’s founders relaunch Lumi as a news curation app powered by swipes and social networks”


There are too many places to search for the news you’re interested in reading.

People find it difficult to hear about news from friends and influencers.

News apps are still based on web based experience.

Working process

Changing a culture of engineering to incorporate design at the centre of the business, using prototypes to test concepts and validate assumptions before taking this into development.

Working with developers to establish the initial basic MVP and work in design sprints / development sprints as an efficient way to work closely with developers and execute on the product.

To complete the feedback loop we learn from MVP data to help make informed design decisions and make further improvements. By creating a solid MVP base we were able to continue expanding the app with new features and improvements to find product market fit.

Working with developers to establish the MVP

  • MVP prioritisation and work into stories
  • Working to weekly development sprints
  • Agile development methodology

React to the things you read

Create discussion from the things you like to read and from the people you follow on social media.

Social reading
Reactions from friends
Following friends article
React to story

View original story

Some publishers weren’t happy with some of their content being extracted from the page, so in some instances we showed the original article in a web view to avoid difficulties.

Removing the signup barrier and a better onboarding experience

One major objective for us was to increase app usage and signup rate. One way to improve and move the needle on this was to remove the login barrier upfront and spread social signup across the app experience. This improvement alone increased signups from 62% up to 97%.

Once users opened the app they were instantly recommended stories. And because the app learns instantly from many signals, we were able to build a better recommendation model of a users interests through their reading behaviours.

Introducing stories from friends and social metrics increased reading time

To increase reading time we added social metics (Facebook shares / likes / prioritising friends likes /comments) to encourage users to read more – We saw a 20% uplift in reading time due to the introduction of social reinforcing elements.

To improve further engagement we enabled users to add topics and publishers to enhance their news feed and news recommendation. This improvement increased app engagement and also enabled users to tailor content from the very beginning.

Finding the right navigation

Hamburger navigation
The problem we found with the hamburger menu was users did not navigate to the menu to discover other sections.

Tab bar
The next iteration was to change the nav bar to encourage more activity and discovery across the app.

Hamburger navigation
Tab bar navigation

User interface

Creating a sense of ease, a sense of magic through the user interface is often over looked. Animation details like the cards loading on opening the app, help to give an immediate sense of how to navigate the app, and by using an elastic effect snapping the cards back into place shows further usability improvements.

“I worked with David whilst being VP Engineering at Lumi. David introduced design best practice and prototyping into our existing development process. He fundamentally reimagined the Lumi user experience and worked with engineers, product managers and directly with the founders in executing on that vision.”

Neil Johnson

© 2020 David Standen