Data Entities Matching Tool

Replacing the back-end Java team process of matching millions of records from various database sources with an easier and more intuitive user flow and interface.


ICX4 is a London-based fintech company that specialises in data remediation, analysis, and regulation compliance for tier one banking clients.  While working as a UX designer there, I collaborated with data and business analysts, product stakeholders, clients, and front and back-end developers to design and test products to streamline clients’ legacy data systems and improve their efficiency.


I was given a brief to design an elegant user interface so that users (employees in tier one banks) could perform the complex task of proscribing rules for matching entity records from various data sources. The details of entities (e.g. names, addresses) are recorded differently across the source data bases, and even within the user’s own internal systems. A single user currently has to trawl through numerous spreadsheets to identify matches. The process is long and the potential for errors going undetected is high.


I sat down with the back-end developer and the data analyst to understand the process and the many spreadsheets of data. We drew an initial sketch of the process, which helped me  understand the user’s journey and goals. A good outcome for the user was to simplify reams of excel spreadsheets into something more manageable to read, as well as to allow the user to save similar match rules as templates.

I sketched a number of user flows to see where I could reduce the cognitive load on the user, discussing my ideas with the data analyst to ensure the flow would meet the user’s goals.


I sketched wireframes in pen and paper based on the user flows I’d drawn. Stakeholders in the product team expressed a desire for the whole process to be made up of drop-down menus. This would have been straight-forward to build, but I was wary of over-using drop-downs and the potential for the process to be more of a nuisance to users that way. 

I sketched the screens as a linear process, with a prominent call-to-action button and a navigational breadcrumb trail to highlight what stage of the process they’re on.


I wanted to test the user interface and whether the user journey was intuitive with more users, however we weren’t able to source users due to the company’s privacy policies. There were a few people with whom I could test who actually knew the function of creating entity match rules. It was a fine balance during testing with unfamiliar users, between explaining it to them and not posing leading questions. 

Each stage of the process had metrics associated with it, such as the percentage of successful matches. We had insufficient information in the design team, which is why the metrics on these wireframes are rather basic and more of a placeholder graphic.


After testing the screens with data analysts and discussions with the back-end developer, we handed over the PAT tool over to the development team to be built. Unfortunately, the company went in a different direction to focus more on internal data analytics so I wasn’t able to test and re-iterate based on more feedback from other users.