Duhqa
Overview
Product: Duhqa is a B2B on-demand logistics platform that connects small retailers with manufacturers and suppliers operating in Nairobi, Kenya. It consists of a customer ordering mobile app, rider mobile app, and warehouse admin web app. The platform is designed to coordinate ordering, fulfillment, and last-mile delivery.
Role: Solo UX/UI designer working with the development team.
Duration: The project took 11 months.
Challenge
Problem: Nairobi is the biggest economic center in East Africa, growing rapidly but facing issues with road infrastructure. At the time of the project, many parts of the city had roads in poor condition or no roads at all. Local grocery shopping was usually done in small local shops that had major challenges in getting supplies. Duhqa wanted to solve this by having its own warehouse in the city and providing a delivery service through an on-demand delivery platform similar to Wolt, but for B2B customers focusing on small retailers. They used an off-the-shelf Warehouse Management System (WMS) that did not fully work for their needs, and they did not yet have a solution for a customer ordering app.
Constraints: With underdeveloped road infrastructure came challenges such as not having a detailed map of the city, unnamed streets, and missing address data. Main issue we had as Europeans is lack of local context. Another constrain was that this was a small startup with limited budget so we decided early on to use an off-the-shelf design system in order to cut development and design time.
Approach
Research: My project manager and I conducted a field research in Nairobi for 7 days. We shadowed warehouse administrators and workers, conducted interviews, spent time in delivery vehicles with drivers, and talked with customers. We also completed a competitive analysis to contextualize our findings.
Key insights:
Field research showed that many assumptions we had did not hold in real conditions. There were many specific local contexts and needs to be covered:
We could not rely on packing lists from producers, as they were often incorrect.
Shop owners used delivery moments for upselling, sometimes buying items directly from the delivery vehicle that were not originally ordered.
Upsell requests often came from non-users, making it difficult for drivers to keep track of the sails. This made impossible to track most of the upsales in the warehouse dashboard.
Failed delivery rate was high, around 30%. We noticed that drivers who frequently called and texted customers to confirm delivery locations and time had significantly lower failed deliveries. Although company tried to incentivize all drivers to do this, it was hard to keep track of driver-customer communication.
Market research indicated this was an untapped niche, and that we were early players in this segment.
Design decisions: We decided to take into account all local context and try to align with known similar models like Uber, Wolt, and DoorDash. Many concepts had to be adapted, and new ones had to be developed.
Solution
Final designs: Interesting solutions I made included allowing users to write descriptive instructions on how to reach their location, since GPS coordinates alone were not enough. We developed a system where drivers always carried extra popular products with them and could add items to orders on the spot for customers, even create profiles for them. Another important feature was workflow where driver needed to contact every delivery before they could mark delivery failed, they could also set new delivery time they agreed on with the customer.
Marketing Website: I created marketing website primary facing retail customers, so that it can be used in marketing campaigns and draw new users. Website was created in Webflow, so that administration can easily edit contents by themselves.
Impact
Results:
Improved delivery success rate by reducing location ambiguity through descriptive address inputs.
Increased order flexibility by enabling drivers to handle on-the-spot product additions.
Reduced operational friction between warehouse, drivers, and customers through clearer workflows.
Improved reliability of deliveries in areas with poor infrastructure and incomplete mapping data.
Established a warehouse management system that supported both warehouse operations and last-mile delivery.
Lessons learned:
Field research was critical in uncovering constraints that would not be visible in desk research or assumptions.
Designing for infrastructure limitations required rethinking standard UX patterns used in Western logistics platforms.
Flexibility in the system (drivers, orders, delivery instructions) was more valuable than strict structure.
In emerging markets, logistics UX is heavily shaped by human coordination rather than purely digital workflows.