APL eZRx – B2B Marketplace for PT Anugerah Pharmindo Lestari | Case Study

Phase 1 : Understand

APL's team came prepared. They had already mapped out their internal business processes and compiled a list of requirements and known user pain points before the project kicked off. Our job in this phase was to dig into that documentation, ask the right questions, and develop a shared understanding of how orders actually moved, from a customer picking up the phone to stock leaving the warehouse.

That foundation shaped everything that followed.

Phase 2 : Ideate

With the problem space clear, we ran collaborative working sessions with the Business Analyst and Project Manager to sketch out how the platform should work. These sessions weren't just about features. They were about sequencing: what does a customer need to do first, what information do they need at each step, and where does the current process break down most often?

One question we kept coming back to was how to handle the variety of ordering habits across APL's customer base. Some users usually bought items by a regular listing, ordering one-by-one item to the sales representative. But others uploaded Excel files to place large bulk orders, a habit built over years that was clunky by modern standards but deeply theirs.

We made several deliberate decisions about how each flow should work: how much information to surface at each step, where confirmation moments should sit, and how to reduce back-and-forth without removing user control. For the Excel upload specifically, rather than retiring it, we redesigned what happened after by converting the file into a cart-style review screen so customers could check quantities and adjust before anything was submitted. Familiar entry point, better and more trackable exit.

The digitalized sitemap was developed by the Business Analyst, based on initial sketches and insights gathered during the design sprint ideation phase

Phase 3 : Decide

User flows got turned into low-fidelity wireframes, enough structure to test assumptions without committing to visual decisions too early. We reviewed these with the team in rounds, using each pass to catch gaps in the logic and tighten up the interaction patterns before moving forward.

Initial lo-fi design

Phase 4 : Prototyping & Testing

High-fidelity design came next, working across web and mobile in parallel. The two platforms couldn't be treated as mirrors of each other. A web interface built for a desktop browser behaves differently from a mobile app used on the shop floor. Navigation, hierarchy, and layout were each adapted to suit how users would actually interact with the platform on each device, while keeping the visual language consistent enough that switching between them felt seamless.

Phase 5 : Validation, Handover and Finalize

Clickable prototypes in Adobe XD let us put the designs in front of users and see where things held up and where they didn't. Feedback from those sessions fed back into the final designs before anything went to development. Once locked, I worked directly with the front-end team during build to make sure the implementation stayed true to the design intent.