Stardog Documentation System
Restructured Stardog’s documentation around how users actually work replacing product-first organization with clear, task-based pathways.
Tools
Figma
Slack
Google Suite
Notion
Team
1 PM & Research Lead
1 Engineer
4 Designers
Contribution
• Conducted user interviews to uncover pain points with the existing documentation system
Co-designed the research survey used to gather quantitative user feedback
Performed the IA audit in partnership with a teammate, evaluating content structure, navigation, and labelling across the site
Contributed to the competitive analysis and SWOT analysis to benchmark the documentation against industry standards
Helped facilitate affinity mapping sessions to synthesise research findings into themes
Conducted usability tests on the current documentation system to observe real user behaviour and validate audit findings
Led a card sorting study to understand how users mentally organise content, directly informing the left navigation restructure
Contributed to prototypes for IA restructuring and improved search functionality
00 OVERVIEW
About
As part of a UX consultancy engagement through UMD iConsultancy, our team partnered with Stardog an enterprise knowledge graph platform to analyze and improve their documentation experience. Our goal was to identify usability gaps and develop a content strategy to make the docs more accessible for both technical and non-technical users.
The Problem
Stardog's documentation was built for engineers but its users aren't all engineers. Internal employees, business stakeholders, and external clients were all struggling to navigate the same docs, leading to increased reliance on internal support, slower onboarding, and user frustration.
Research Approach
• Competitive Analysis across 5 platforms (Neo4j, Amazon Neptune, Ontotext, Graphwise, Palantir) to benchmark documentation patterns and identify opportunities
Heuristic Evaluation to assess usability against established UX principles
Information Architecture & Content Audit to map the existing structure and surface inconsistencies
9 User Interviews with a mix of internal employees and external clients ranging from engineers to sales reps
2 Usability Tests with internal Stardog participants
Survey of 59 internal users across engineering, support, product, consulting, and sales
We ran a multi-method research sprint to fully understand the problem space before jumping to solutions:
01 RESEARCH
User Interviews - What are our Users Saying
To better understand how Stardog’s documentation supports real users, we conducted interviews with 9 participants. 5 internal employees and 4 external clients. Our goal was to uncover friction points across onboarding, navigation, search, and content clarity, and to identify where the experience breaks down between technical and non-technical users.
We aimed to understand how people actually use the docs not how they’re intended to be used.
The documentation architecture favors those who already understand the system. It lacks clear pathways, strong cross-linking, and task-oriented routes.
The documentation reflects how engineers think about the product, not how different personas experience it.
There is no strong editorial system. Documentation feels decentralized and engineering-owned rather than user-centered and curated.
The documentation is not the fastest path to clarity so users bypass it with LLM's. Using it for intent interpreters, summarizers, jargon translator, search capabilities
Competitive Analysis - What are others doing that Stardog isn't?
We analyzed 5 competing platforms: Neo4j, Amazon Neptune, Ontotext, Graphwise, and Palantir. We were looking for documentation patterns, navigation structures, and features that could inform our redesign direction.
What we noticed across competitors:
Most used both a left global nav panel and a right local nav panel alongside breadcrumbs giving users multiple ways to orient themselves
Platforms like Neo4j used storytelling techniques and use-case-driven structure to engage non-technical audiences
Several competitors offered role-based entry points and dynamic layouts with infographics
Neo4j’s search functionality incorporated natural language processing to find conceptually relevant results, even if they don't contain your exact keywords resulting in a better search experience.
This told us that Stardog's documentation structure is solid for developers but lacks the layering and personalization that helps non-technical users find their footing. There was a clear opportunity to introduce storytelling, role-based paths, and richer visuals.
Auditing the Information Architecture
After conducting a full IA audit of Stardog’s documentation ecosystem, we uncovered systemic structural issues that were not isolated to individual pages. They reflected deeper inconsistencies in hierarchy, labeling, and mental model alignment.
Our discovery wasn’t about broken content. It was about broken relationships between content.
There was no single, scalable taxonomy governing how products were categorized.
Label inconsistency disrupted scanning behavior. Users could find pages but only if they were highly specific in search queries.
The IA lacked macro-grouping. It was comprehensive but not synthesized.
The structure reflected how Stardog internally organizes its products not how users think about learning and using them.
Users typically think in flows like:
Install → Configure → Query → Develop → Deploy → Troubleshoot
But the documentation followed product segmentation first, task flow second.
There was no unified onboarding framework across the documentation system.This disproportionately impacts new users, business stakeholders cross-functional teams evaluating the product.
Stardog’s documentation is rich in content but difficult to navigate as a cohesive system. The structure mirrors how the organization thinks about its offerings rather than how users progress through real-world workflows, making it harder to move seamlessly from learning to execution.
Card Sorting - Understanding Mental Models
Our IA audit found that the Stardog documentation is comprehensive, it has everything it needs. The problem is not missing content. The problem is that content was added incrementally over time, without a consistent structure or order. The result is a doc site that does not reflect how users actually navigate or think about the product. So we conducted a card sorting usability test to better understand our users mental models.
Open card sort · 4 participants · 28 documentation sections · 3 Tasks
Key Findings
Users sort by mental model, not doc structure
Participants grouped by flow, role, or action not Stardog's existing hierarchy.
Three zones of strong consensus
All 4 participants separated onboarding, core usage, and platform administration.
Gray-zone content causes confusion
Items like ML, Inference Engine, and Tutorials shifted categories depending on interpretation.
Abstract labels get renamed
Participants rewrote bare nouns ("External Compute") into verb-forward phrases ("Configuring External Compute").
LLMs help draft, not decide
LLMs surfaced consensus quickly but over-clumped items; boundary decisions required team judgment.
Auditing the Content
The goal of the audit was to identify structural inconsistencies, usability issues, and gaps in visual communication that may hinder the user experience
The content audit revealed systemic inconsistencies in visual structure, component usage, and accessibility across Stardog’s documentation. While the information itself is comprehensive, the lack of standardized design patterns, clear hierarchy, and consistent interaction cues increases cognitive load and reduces scannability.
Surveys
We surveyed 62 cross-functional employees across R&D (34%), Professional Services (21%), Sales (16%), and other departments to understand how internal teams use and perceive Stardog’s documentation
1. Documentation Functions as a Reference Library , not a Workflow System
2. Findability Is a Structural Problem, Not a Content Problem
3. Applied Examples Are the Biggest Opportunity
4. Troubleshooting Is the Weakest Link
5. Mental Model Misalignment Is Systemic
Who Are We Designing for
We consolidated all our research and identified our core suer group
02 IDEATION
What should be build from here
Intent-aware search
Search results that help users identify the right page before clicking without opening multiple tabs.
Plain-language page summaries per result
Previews of key sections with clickable links
Role badges (Business, Engineer, Data) on each result
Intent-matched keyword highlighting
Filter toggles and consistent result structure
Search
Documentation landing page
A welcoming homepage that routes users to the right content based on their role from the moment they arrive.
Role selector (Business, Engineer, Data, New user)
Quick links to core topics
Card-based visual hierarchy for easy scanning
Consistent navigation cues showing where to go next
Navigation
Role-based content display
The site adapts its content and navigation dynamically based on the role a user selects reducing noise for everyone.
Role prompt on first visit; persistent dropdown to switch
Reduced side navigation showing only relevant sections
Homepage and top nav adjust to match selected role
Navigation
Glossary awareness
A dedicated, visible glossary that all users can find and rely on — with terms organised by role so content stays relevant.
Standalone entry in the side navigation
Role-based tabs (Business, Engineer, Data)
Supports terminology alignment across teams
Content
AI semantic search
A search experience driven by intent and meaning, not just keyword matching — letting users ask questions in their own words.
Natural language query support
Context-aware suggestions as users type
Follow-up suggestions to guide next steps
Search
In Progress
My team and I are currently in the process of conducting user-ability tests with our mid fidelity prototypes, stay tuned to see the rest of our designs!
Check out more of my work
2026 • Hackathon • Concept project
My Proudest Project
Attune
We designed a recovery companion app and wearable that translates invisible biosignals into a garden, helping people with anorexia nervosa perceive their hunger and fullness better.