SENSE: Turning Trial Data into Decisions

SENSE is a data visualization and analytics platform built for pharmaceutical clinical trial operations. It surfaces risk predictions, process interdependencies, and delay signals across active studies — giving trial directors, portfolio leaders, and forecasting managers the visibility they need to make proactive decisions instead of reactive ones.

Clinical trials run between 5 and 8 years on average. They are highly sensitive to external disruption: regulatory delays in one country, site recruitment gaps, supply chain bottlenecks, or a single missed checkpoint can cascade into months of trial extension. Every study contains multiple interdependent processes, and when one slips, the ripple effect across related studies is rarely visible until it is already too late.

I worked as a solo designer from my studio, owning both Interaction Design and Visual Design end-to-end. Our team was a Product Owner, a Development Lead, and a Design Manager from the client side at Accenture.

The project opened with a participatory design-thinking workshop I helped facilitate at the client's headquarters. We invited stakeholder groups spanning operators, management, and leadership — three groups that rarely sat together — and spent two days mapping the problem space from each vantage point.

After synthesizing the output, four latent needs cut across every role. Users needed to predict risk and delays from data before they materialized. They needed to trace individual threads and analyze how processes depended on each other. They needed notifications across levels of operations so the right person was always in the loop. And they needed to communicate proactively with cross-functional partners to actually pivot — not just know something was wrong.

SENSE serves three distinct user groups, each with a different scope of responsibility and a different relationship to the data.

SENSE organizes data across two independent hierarchies. The content hierarchy runs from a single clinical study, to a project (a collection of 8–10 studies), to a portfolio (200+ studies in the same disease area). The geographical hierarchy runs from global, to regional, to country, to individual site — the hospital or location conducting the study.

At any point in the platform, the user chooses their position in both hierarchies simultaneously. The primary workflow flows from the widest view inward: start at portfolio level to get a macro read on risk across studies, deep-dive into a specific process that shows strain, then zoom geographically from global down to a single site. The entire platform was designed to make that traversal feel continuous rather than fragmented.

As the engagement matured, SENSE was deployed as a six-screen control center at the client's headquarters. The deployment is called Bridge.

Bridge places each major data point on its own screen. Users can pin widgets from any screen and compare data points across portfolios, studies, or any level of the geographical hierarchy simultaneously. Where the web platform was built for individual analyst work, Bridge was designed for the briefing room — a shared workspace where portfolio directors could orient teams, align on risk, and make decisions in front of live data.

The same design system, visual language, and interaction patterns carry over from the individual platform to the six-screen layout without adaptation friction. Bridge is SENSE at a different scale, not a different product.

SENSE launched to strong reception. The engagement was extended twice based on platform performance — an outcome that reflected the depth of adoption, not just delivery. Upon launch, the platform received media coverage from Bloomberg and was featured on Accenture's website as an example of excellence in client delivery.

The feedback that landed hardest as a designer: a user emailed the CEO of the pharmaceutical client, unprompted, to thank them for pushing the organization to build an app like SENSE. That kind of signal — a user taking time to write to leadership — does not come from a tool that merely works. It comes from something that changed how people work.

The concentric-circle study widget was the riskiest call in the project. It packs a lot of state into a small space, and while power users read it quickly once they learned it, I would now validate that learning curve with new users earlier rather than trusting that density and elegance would carry it.

I would also pressure-test the prediction language sooner. Solid for actuals and dashed for predictions felt obvious to the team, but the whole value of SENSE rests on a user trusting a dashed red bar enough to act on it before the date arrives. That trust deserves its own round of testing, not an assumption.