Back to Work OpenMetadata
Overview Problem Process Design Impact

OpenMetadata · Process

The full process

The research, personas and flows behind OpenMetadata. The case study covers the decisions; this is the work underneath them.

02 — User Research

User Persona & Goals

Three enterprise data personas with distinct responsibilities, workflows, and pain points — each requiring the platform to speak a different language while staying within the same interface.

👤
Shreya Bhat
Data Engineer, 31
Goals
  • Discover datasets quickly without tribal knowledge
  • Track data lineage across upstream and downstream
  • Set up automated data quality checks
Pain Points
  • No central catalog — discovery depends on who you know
  • Hours spent searching for the right table
🧑
Kiran Nair
Data Scientist, 38
Goals
  • Understand data context before modelling
  • Find trusted, certified datasets reliably
  • Track experiment metadata across projects
Pain Points
  • Stale documentation makes context untrustworthy
  • Unclear data ownership leads to repeated work
👩
Priya Iyer
Data Governance Lead, 44
Goals
  • Enforce data policies across the organisation
  • Track compliance with regulatory requirements
  • Understand how data is being used enterprise-wide
Pain Points
  • Manual governance spreadsheets don't scale
  • No audit trail for data access or usage

03 — Business Challenges

Core Challenges

CHALLENGE 01
🔍
Data Discovery at Scale

With thousands of tables, pipelines, and dashboards, finding the right dataset without a catalog meant relying on tribal knowledge — slow, inconsistent, and impossible to onboard against.

CHALLENGE 02
🛡️
Trust and Data Quality Signals

Data teams needed visible signals of data freshness, ownership, and certification — without those signals, every dataset required manual verification before it could be trusted in analysis.

CHALLENGE 03
⚖️
Governance Without Bureaucracy

Traditional governance tools added friction that teams resisted. OpenMetadata needed to embed governance naturally into the discovery workflow — making compliance the path of least resistance.

CHALLENGE 04
🔄
Metadata Freshness

Stale documentation and outdated metadata was often worse than no documentation — it created false confidence. Automated, always-current metadata was a technical and UX requirement.


04 — Secondary Research

Market Insights

FINDING 01
40%
Data Engineer Time Spent Searching

Data engineers spend 40% of their working time searching for and understanding data — time that should be spent on analysis, modelling, and building pipelines.

FINDING 02
63%
Analytics Projects Delayed by Discovery

63% of analytics and data science projects experience delays caused by data discovery bottlenecks — a systemic problem that better tooling directly addresses.

FINDING 03
3×
Faster Onboarding with Data Catalogs

Organisations with active data catalogs onboard new data professionals 3× faster — making the catalog a strategic talent and productivity investment, not just a governance tool.


07 — User Flow

The Journey

01
Connect Data Source
Connect warehouses, BI tools, and pipelines via 500+ native integrations with zero-code setup
02
Auto-discover Assets
AI automatically scans and indexes all data assets, building the catalog without manual entry
03
Enrich Metadata
Teams add descriptions, tags, ownership, and quality rules to enrich auto-discovered assets
04
Search & Explore
Users search the catalog with semantic queries, filters, and type-ahead to find trusted data in seconds
05
Track Lineage
Visual lineage graph shows upstream sources and downstream consumers for every asset
06
Govern Policies
Governance leads define, apply, and audit data policies across the entire estate from a single dashboard

08 — Toolkits

Tools & Workflow

Tools and methods used throughout the design process — from enterprise user research through information architecture, interaction design, and final delivery.

🎨FigmaUI Design
🗂️FigJamWorkshops
📋NotionDocumentation
🗺️MiroJourney Mapping
🧪MazeUsability Testing

Design Process

From chaos to catalog

A six-phase process that started with enterprise user research and ended with a cohesive design system deployed across both the product and the marketing website.

01
Enterprise User Research
Interviews with data engineers, analysts, and CDOs across 15+ enterprises. Journey mapping, pain-point taxonomy, persona definition.
02
Information Architecture
Designed the IA for 200+ entity types — tables, pipelines, dashboards, ML models, topics, containers. Hierarchical taxonomy and relationship mapping.
03
Navigation System
Rebuilt the global navigation to support role-based contexts. Data engineers, analysts, and governance officers each needed a different primary path.
04
Search & Discovery UX
Semantic search with faceted filtering, type-ahead, relevance signals, and intelligent ranking. Reduced discovery time from 2.5 hrs to under 20 minutes.
05
Lineage Visualization
Interactive graph-based lineage UI — nodes, edges, collapse/expand, upstream/downstream isolation, impact analysis overlays.
06
Marketing Site Design
Designed open-metadata.org to convert enterprise buyers — clear value props, integration directory, pricing clarity, documentation entry points.
Back Return to the case study