Transforming enterprise data discovery into a self-service experience that increased adoption by 350%.
Product Design
Retail Banking
Data Management
Overview
1
Enterprise data was fragmented across teams, difficult to discover and expensive to consume. Creating new data products required technical expertise, access was managed inconsistently and knowledge of available datasets often relied on word of mouth.
I led the end-to-end redesign of the platform, transforming it into a self-service data marketplace with guided creation, enterprise search and role-based governance. The result lowered the barrier to data creation while making trusted organisational data significantly easier to discover and consume.
Outcomes
350% increase in adoption, helping shift the organisation toward self-service data consumption.
Reduced creation time from over 15 minutes to under 3 minutes, lowering the barrier for new producers.
Unified enterprise data into a central catalogue, improving discoverability, ownership and governance.
I led the end-to-end product design, partnering with Product, Engineering and Data SMEs from discovery through developer handover. My responsibilities included research, workshop facilitation, workflow architecture, interaction design, prototyping and usability testing.
The Challenge
2
Fragmented Data Discovery
High Barrier to Creation
Creating a data product required significant technical knowledge, making publishing inaccessible for many domain experts.
Weak Governance
Ownership, permissions and access requests varied between teams, creating inconsistent governance and reducing confidence in shared data.
Research Insights
3
I interviewed data producers and consumers across multiple teams to understand how data was created, discovered and consumed. With limited access to end users, research was complemented by weekly workshops with engineering and domain experts to continuously validate assumptions and refine the product direction.
Key Product Decisions
4
I created a searchable catalogue that brought enterprise data into one discoverable location. Familiar search, filtering and categorisation patterns helped users quickly identify relevant datasets without relying on organisational knowledge.
02 Guided Creation
Creating a data product was transformed from a technical API workflow into a guided experience that reduced configuration time while making publishing accessible to a much broader audience.
Breaking creation into logical stages reduced cognitive load while allowing users to save progress and collaborate asynchronously. This transformed a technical workflow into something approachable for non-developers.
Alternatives Considered
We explored exposing all configuration upfront to mirror existing developer workflows. This proved overwhelming for less technical users, so we progressively revealed configuration throughout a guided creation journey.
03 Flexible Consumption (Checkout)
Different consumers required different versions, output formats and delivery methods. Rather than exposing every option immediately, I introduced a configurable checkout that allowed users to tailor consumption without increasing cognitive load.
Review before provisioning
Configure only when needed
Multiple output formats
Version selection
Prevent configuration errors
Familiar checkout experience
Alternatives Considered
Early concepts surfaced every configuration option before requesting access. User feedback showed this increased complexity, so configuration was deferred until users had selected the products they intended to consume.
04 Enterprise Governance (RBAC)
Rather than introducing another permission model, I integrated governance directly with the organisation's existing Active Directory groups. This reduced administration while aligning with established security and compliance processes.
Reflection & Hindsight
5
This project extended beyond redesigning workflows—it required changing how people thought about enterprise data. The challenge wasn't simply making data easier to find, but lowering the technical barriers that prevented teams from discovering, publishing and consuming data independently while maintaining governance.
Borrowing familiar interaction patterns from retail experiences proved to be one of the most valuable decisions. Catalogues, baskets and guided journeys gave users confidence to navigate an unfamiliar technical domain without extensive training. It reinforced that effective enterprise design isn't always about creating new patterns, but about applying existing mental models in the right context.
In hindsight, I'd spend more time measuring behaviour after launch to better understand how teams discovered and reused data over time. While adoption increased significantly, deeper behavioural insights could have informed future improvements to recommendation, discoverability and collaboration across the platform.