Adding a collaborative referencing system for readers and writers.
Medium
DURATION
3 months | 100+ hours
CLIENT
Student project @ DesignLab
TOOLS
Figma, FigJam, Maze
UX RESEARCH
UX DESIGN
UI DESIGN
USABILITY TESTING
BACKGROUND AND BRIEF
Design a feature that helps make Medium a more reliable source of information.
Medium is a publishing/blogging platform that hosts a hybrid collection of amateur and professional writers and publications.
Even though the reader base increases continuously due to search traffic, the current platform does not provide the means for users to feel confident about the quality of the information they provide, leaving many readers skeptical.
This project attempts to identify the main trust barriers and validation mechanisms and design a feature that attends to this limitation, improves overall user experience and confidence, and is coherent with their business goals and competitive market.
DISCOVERY AND RESEARCH
How does Medium curate and distribute its articles?
Medium has an internal network of article curation and distribution. Although they are not explicit about the entirety of the process, theories have been discussed by professional Medium writers.
Based on their internal guidelines, Medium uses human curators to review stories to be distributed to their readers through the homepage dashboard and suggestions. The company also uses AI (algorithms) to curate and distribute stories. It is in the writer's interest to get their article curated since it will boost its reach and recognition.
It is important to note that Medium suggests stories to their readers based on current relevance, not date. They believe this encourages writers to make bigger investments in their articles since they will amortize the investment over months, not days.
The platform's writers can post articles independently or through a publication to increase its reach. Moreover, readers may follow not only writers and publications but also #topics. It is through topics that readers will most likely find a new article, writer, or publication.
Medium dashboard
EMPATHY RESEARCH
Who are Medium's users?
Medium has approximately 54M registered users (writers and readers-only) and 700K paying members.​ According to Similarweb, most of Medium's traffic comes from organic search engines (61%), followed by direct referrals (25%).
A significant portion of occasional Medium users is not registered on the platform, arriving when researching a specific topic online.
In February 2022, Medium's website received 163 million visitors.
I interviewed a sample of 5 Medium users, among members, writers, and occasional readers. My main findings included:​
-
Most use Medium for professional/academic purposes but generally do not rely solely on the website's content.
-
Referenced sources, including related articles, and more transparency on Medium's curation process were said to help increase confidence.
-
Average engagement, external fact-checking, and researching the author's record are ways they determine trust.
-
They don't generally engage with the articles.
Based on the research and data gathered, I developed a persona and their journey map to guide me throughout the design process and better identify opportunities.
FEATURE RESEARCH
What else is out there?
After understanding the business and its users, and validating the initial hypothesis, it was time to piece it all together and brainstorm potential solutions. I conducted feature research on other text-based products that included, Twitter, Reddit, Kindle, Wikipedia, and Perusall and came up with ideas on how to make Medium not only more trustworthy but also more collaborative and engaging.
IDEATE
Picking the best solution
After the research phase was completed, I brainstormed several ways to solve this problem.
My initial idea was to present a multi-feature solution that would come together through a confidence-level meter, where writers would earn points and badges by properly referencing articles, having their article curated by another Medium writer, cross-referencing other Medium articles, and having good engagement stats. ​
I later realized, however, that I was way over my head and that this idea would not be realistic due to time constraints.
The final idea
I decided to synthesize my initial solution into one useful feature: a referencing system.
Seen as most users already fact-check elsewhere when reading a Medium article, why not share it with other readers?
Creating this collaborative experience of sharing relevant reads and media to complement an article's content inevitably increases reader confidence and creates a more welcoming environment for engagement (thus increasing Medium's low engagement rate).
Okay, but how would it work?
1
Expand the existing highlight feature
Give users a new section that shows all highlights. Also, let users know how many users and who highlighted each passage.
2
Create a new feature to add a reference
To each highlight, the author and readers can add references to further expand the discussion and validate the information provided in the article.
3
User-generated suggested reads
Change the AI-generated suggestions in the article's sidebar for user-generated suggestions.
INTERACTION DESIGN
Designing the flow for the new feature.
The most important aspect after defining what a new feature will be is to design something that is coherent with the existing product. This is not only how it is visually presented, but also how users will interact with it.
To design this feature I made sure to learn how Medium presents each of its features to the user. For example, when a reader decides to comment on an article, what is this experience like?
After I gathered every necessary aspect for this I was able to design the flow and an initial layout for the UI.
UI KIT
Matching new elements to an existing product.
From the existing features in Medium, I had the challenge of incorporating their UI style into the new feature, integrating it seamlessly into the existing platform.
Existing features in Medium
UI Kit | Referencing System
USABILITY TEST
Test & iterate
Testing the new feature's usability was the next step once the UI was ready. Going in, I was aware that testing this product would be challenging since working with a prototype requires familiar Medium users (at different levels) to consciously perform given tasks rather than casually interact as usual.
I conducted 10 usability tests via Maze, structuring it into 5 missions, each followed by an opinion scale ranging from 1 - very hard to 5 - very easy.
The missions that had the worst performance and ranked lowest on the opinion scale were those that required discovering the new features.
I had previously asked the testers about their familiarity with Medium, which allowed me to conclude that the occasional user had a more difficult time discovering the new icons, either bouncing or rating low on the opinion scale. The more familiar users, however, ranked high on the opinion scale.
Medium's standard for launching a new feature is non-intrusive. The Medium Staff page introduces the feature in an article and more writers, organically (or not), reproduce the content and start using the feature. With time, less frequent users also adapt and adhere to it.
Since this is a feature for an existing product, I believe the best way to approach it is to follow their usual process.
Medium article introducing the new feature
KEY LEARNINGS
What can I take to the future from this project?
Keep it simple!
Possibly my main lesson from this project is that even the simplest developments take a lot of work. I initially intended to solve the problem with a multifeatured approach, and thankfully I was advised to focus on developing one of the features very well rather than several poorly.
Dedicate more time and energy to recruiting
Had I known earlier how difficult it would be to find Medium users who would be available and willing to participate, I would've dedicated more to recruitment. Although I managed to interview a variety of users, some more familiar with the platform than others, I missed interviewing and testing the product with professional Medium writers. This could have given me a more in-depth perspective.
More constraints than expected
This was by far my most challenging project for DesignLab. Even though I wasn't working with a real client, I was working with a real product. This is especially complicated since they have practices that have been around for a decade and we can't know how they evaluate their effectiveness. Following their methods and being coherent with the brand was much harder than I expected.