Academia.edu
Researcher Experience
Redesigning the new-user onboarding and feed experience for Academia.edu - turning an empty, impersonal first impression into a personalized, topic-driven research discovery system.
Client Project
Product Design
Lead Product Designer
Web
Jun 2022
A platform for the world's research - with a first impression problem
Academia.edu hosts over 34 million papers with a mission to make research freely accessible. But new researchers signing up were landing in an experience that felt empty, unstructured, and hard to navigate - undermining the platform's promise from day one.
34M+
Papers on Academia.edu
4
Connected surfaces designed
2
Core user audiences
1
Systemic problem to solve
Problem
The empty state that was costing researchers
New researchers were making their first decision about Academia.edu based on an experience that was working against them. Understanding why meant looking beyond a single screen to the entire onboarding and discovery flow.
Root Cause
No Personalization Infrastructure
The empty feed wasn't a visual problem — it was an architectural one. Without knowing what a new user cares about, the platform had no basis for surfacing relevant content. The fix had to happen earlier in the flow, at sign-up.
Why This Matters
First Impression = Retention Signal
For a research platform, the moment a new user fails to find relevant content is the moment they decide Academia isn't for them. A strong, personalized first experience isn't a nice-to-have - it's the primary driver of whether researchers return.
I deliberately began the experience at the Logged-Out Single Work Page (LOSWP) - the primary entry point where researchers first discover Academia.edu via search or a shared link. This grounded the design in the real user journey, not an assumed one that starts at a homepage.
Research & Strategy
Understanding the researcher's mental model
Before designing, I needed to understand how researchers actually think about discovery - and what makes a research platform feel credible and worth returning to.
Audience Analysis
Who Are Researchers on Academia?
Researchers visit Academia with a specific intent: to find papers, understand a field, and discover connections within their area of study. Unlike social platforms, credibility signals matter more than social signals. A researcher needs to trust the platform before they'll invest time in it.
Competitive Analysis
How Research Discovery Works Elsewhere
Looked at how Google Scholar, PubMed, ResearchGate, and Semantic Scholar handle new-user onboarding and content discovery. Key pattern: platforms that invest in explicit topic selection at sign-up dramatically outperform those that rely on behavioral inference alone for early personalization.
Content Audit
What Academia Already Has - and What's Being Wasted
Academia has a rich taxonomy of research topics, 34M+ papers, author profiles, supplementary visuals, and videos. The problem wasn't a lack of content - it was an architecture that failed to surface it meaningfully. This shaped the design principle: organize what already exists, don't just add more.
Strategic Direction
Topics as the Foundation of Personalization
I landed on research topics as the organizing principle of the entire experience. Topics are explicit (unlike behavior), durable (they don't change session-to-session), and immediately actionable - they give the system everything it needs to populate a meaningful feed from day one.
Personalization can't be retrofitted onto an empty experience. It has to be built into the foundation - at the moment a user tells you who they are.
Topic selection as the gateway
The redesigned sign-up flow introduces a mandatory step between account creation and the news feed: users must select at least three research topics. This single addition transforms the entire downstream experience - the feed is never empty, and personalization is immediate.
Entry point begins at the LOSWP - where most researchers first encounter Academia - and flows naturally into account creation before the topic selection step.
Structure over volume
The redesigned feed organizes content by topic section, with a fixed hierarchy within each section. Users' chosen topics appear prominently in a hero bar at the top as direct navigation anchors - giving researchers an immediate path to what matters most.
A central hub for each research field
Each topic has a dedicated landing page that extends the feed's content hierarchy. When a researcher navigates from their feed to a topic page, the same organizing logic - papers first, then supporting content, connections, and recommendations - makes the experience immediately familiar.
Topic pages also serve as a discovery entry point for breadth - a researcher deep in Neuroscience might discover a Linguistics connection they wouldn't have found through search alone.
Personalization that follows researchers off-platform
The weekly digest email extends the personalization system beyond the site. Curated around a user's chosen research topics, it surfaces only the most relevant papers from the past week - mirroring the feed's visual structure and organizational logic.
This creates a low-effort re-engagement loop: researchers who haven't visited in a week receive a prompt that feels personal and useful, not generic. The email isn't a broadcast - it's a continuation of the experience they already know.
A Connected Multi-Surface System
Designed a coherent experience across four surfaces - sign-up flow, news feed, topic landing pages, and email digest - all anchored to the same personalization logic. Each surface reinforces the others rather than operating in isolation.
Information Architecture Framework
Established a reusable content hierarchy (papers → supporting content → connections → recommendations) that can be applied consistently across topic sections, feed cards, and landing pages. Structure becomes the design system.
Root-Cause Thinking Over Surface Fixes
Rather than redesigning the homepage, identified that the problem lived upstream - in the absence of a personalization signal at sign-up. The one mandatory topic-selection step solved the empty-state problem without requiring behavioral inference or heavy ML infrastructure.
Presented company-wide
Presented company-wide - to the founder, directors, and cross-functional teams. The concept demonstrated that strategic, systems-level design thinking can emerge from a focused brief and resonate beyond the design org.
01
Fix the infrastructure, not the interface
The empty feed was a symptom. Adding more homepage modules or better visual design wouldn't have solved it. The real fix was one mandatory step earlier in the flow - a reminder that the most elegant design solutions often happen upstream of the visible problem.
02
Hierarchy is the product
For a content-heavy platform like Academia, the information hierarchy isn't a layout decision - it's a statement about what the platform values. Putting papers first in every section wasn't arbitrary; it was a deliberate signal about what kind of platform Academia wants to be.
03
Cross-surface thinking separates good design from great design
Extending the design to include an email digest wasn't scope creep - it was what made the concept feel complete. A personalization system that only lives on-site misses half the opportunity. The email was what closed the retention loop.