Design a versatile document system

PayFit is a payroll SaaS where admins can manage their payroll. On the platform we have a dedicated space where admins can add, view and edit all their documents (company documents, payslips, employment contract…).

The document experience was designed many many years ago and needed a redesign because admins & accountants struggled to retrieve employee documents efficiently (CSAT: 2.9)

When I joined the project, the discovery part was already done and directed by the previous Product Manager but it was a project without PM.

We discovered our users were struggling with a document system that buried important files in layers of folders and subfolders organized by month and year.

Finding a single document required multiple clicks through a maze of empty folders, while uploading files was equally frustrating. The result? Users dumping everything into an "Other documents" folder just to avoid the navigation nightmare.

Our metrics revealed this pain point was significant.

Instead of the typical folder hierarchy, we envisioned a unified document list with powerful filtering capabilities. My design partner had already sketched initial wireframes that transformed folders into document category attributes.

After analyzing competing systems and questioning several design choices, I focused on improving the filtering system—the most crucial element of our solution. We hit an unexpected roadblock: PayFit lacked proper filter components. The existing selectors and multi-selectors weren't designed for filtering use cases.

Rather than compromise, we designed a new filtering system from scratch. We even created PayFit's first payroll period selector with preset options. This breakthrough enabled users to quickly find exactly what they needed without endless clicking.

While my partner prepared user tests scripts, I developed interactive prototypes.

After going through design critics, we collaborated with the development team to ensure clear transmission of technical specifications.

Just before the launch, I was transferred to another topic and Charlotte Marmet became the Designer responsible for the first post launch iterations.

Due to the absence of a product manager, we initially launched without the bulk import feature — a critical mistake. Users struggled to navigate the new environment, resulting in negative feedback. We also had very positive feedback, just waiting for the bulk export.

dz

A complete feature rollout from the start would have prevented initial negative feedback. The lack of bulk import at launch weakened the refactored system's usability.

Additionally, the radical shift from a folder-based to a filter-based document organization system might have been too abrupt for users. A limited release to a small group would have allowed us to de-risk the rollout, gather rapid feedback, and iterate before the full launch.