RISK LEDGER
Reducing duplicate supplier profiles through better organisation search
A two-week B2B SaaS search redesign focused on improving supplier discovery, reducing duplicate organisation creation, and creating a clearer search-and-connect pattern across client and supplier workflows.
My contribution
Lead Product Designer
The team
1 × product manager
1 × product designer
3 × engineers
Year
2026
Summary
Risk Ledger relies on accurate organisation data. When clients cannot find an existing supplier, they may create a duplicate profile, fragmenting supplier data and weakening the value of the network.
I led the design of an improved organisation search experience to help users find the right supplier faster, understand whether an organisation already exists, and avoid creating unnecessary duplicates. The work combined search result UX, clearer metadata, creation fallbacks, role-specific actions, instrumentation, and a more consistent mental model across client and supplier search.
The project was deliberately constrained: one squad, two weeks, and no major algorithm rewrite without better data. The design focus was therefore on productionising a hackday improvement, tightening the UX, reducing ambiguity, and creating the measurement foundations needed for future search improvements.
Business objective
Improve supplier discovery so clients can find the correct existing organisation before creating a new one.
The project aimed to:
-
Reduce duplicate supplier profiles.
-
Increase successful connections between claimed organisations.
-
Improve confidence when choosing between similar suppliers.
-
Make search behaviour more consistent across client, supplier, and fourth-party workflows.
-
Add observability so future search improvements could be based on data instead of guesswork.
Success measures
Because search had limited existing observability, we defined a combination of primary, secondary, and guardrail metrics.
Primary success was measured by:
-
Increasing connections made between claimed organisations by 5%.
-
Reducing duplicate organisations created per month by 2%.
Secondary success was measured by:
-
Keeping the percentage of searches ending in new organisation creation below 5%.
The guardrail metric was:
-
Weekly users of the search endpoint should not drop by more than 20%.
Part of the project was also about improving measurement. We needed better visibility into how users searched, when they created new organisations, and whether the new search experience reduced avoidable duplicates.
My role
I led the product design work from problem framing through design, implementation support, and QA.
My role included:
-
Framing the problem around duplicate prevention, data quality, and network trust.
-
Auditing existing supplier discovery surfaces.
-
Defining the desired search behaviour across supplier, client, and fourth-party search.
-
Designing the modal states for popular results, recently viewed results, active search, no results, creation fallback, connection states, and error handling.
-
Working with engineering to keep the scope shippable within two weeks.
-
Avoiding unnecessary algorithm changes until the team had better observability.
-
Using Cursor to inspect existing frontend behaviour, understand edge cases, and improve design-to-engineering handover.
-
Supporting implementation QA across different role and relationship states.
Discovery
Search was not one clean experience. It was split across several overlapping surfaces.
Users could search through:
-
The search modal.
-
The Browse all suppliers page.
-
The All suppliers navigation item.
-
Supplier-side client search.
-
Network visualisation tables.
-
Fourth-party supplier search.
Each surface had slightly different data, actions, and expectations.
The biggest issue was that users were not trying to “browse all suppliers”. They were trying to find a specific organisation and understand what they could do next.
The existing Browse all suppliers page mostly existed because the search modal could only
show a limited number of results. That created a broken mental model:
-
Search in the modal.
-
Hit a result cap.
-
Get pushed to a different page.
-
See a different layout.
-
Lose confidence about where to search or what action to take.
There was also naming confusion. “All suppliers” meant suppliers connected to the client, while “Browse all suppliers” meant suppliers across the platform. From a user’s perspective, that distinction was not obvious.
This shifted the design challenge from “make the modal nicer” to “create a clearer organisation discovery pattern”.
Design principles
Help users find before they create
Creating a new supplier should be available, but it should not feel like the default path. The experience needed to make existing suppliers easier to find first.
Show enough detail to choose confidently
Supplier name alone was not enough. Users needed extra signals such as domain, country, and connection status to understand which organisation was the right one.
Keep the pattern consistent across roles
Client search, supplier search, and fourth-party search had different actions, but they should not feel like completely separate interaction models.
Avoid cleverness without data
The team deliberately avoided deeper search algorithm work until better observability was in place. The focus was to productionise a known improvement, tighten the experience, and measure what happened next.
Outcome
Through countless iterations, real-world testing, and navigating technical and environmental constraints, we transformed a fragile inspection concept into a tool that shaved 80 minutes off each inspection workflow. The project revealed the messy reality behind polished UX—highlighting the importance of designing for real-life conditions, aligning cross-functional teams, and adapting to constantly shifting assumptions. In the end, we built not just a usable product, but one that felt intuitive under pressure, proving that great design is forged in the chaos, not in spite of it.