Introduction

Yara’s digital agri-marketplace, Bodega, helps farmers purchase high-quality agricultural inputs directly from verified dealers. But while Bodega’s web platform was gaining traction, a large segment of the market remained unreachable: smallholder farmers in rural Kenya who didn’t own smartphones or had unreliable internet access.

The solution? Bring Bodega to USSD.

USSD (Unstructured Supplementary Service Data) is an older mobile communication protocol, still widely used in countries like Kenya. It allows users to access simple, menu-based services from even the most basic phones. We leveraged this technology to give every farmer access to the digital marketplace, regardless of their device or location.

Context: Designing for Real Access, Not Just Interfaces

In emerging markets, digital inclusion isn’t just about apps or mobile-first design—it’s about infrastructure. In Kenya, a significant portion of farmers use feature phones. While mobile penetration is high, internet access isn’t.

This wasn’t about creating a lite version of an app. It was about designing an experience that could:

• Work on the most basic phones

• Load instantly, without data or Wi-Fi

• Be understood by users with limited formal education

• Reflect the trust and usability of the digital Bodega experience

We had to rethink everything from navigation to confirmation flows, adapting complex marketplace functionality into simple, time-limited USSD menus.

The Challenge: Making USSD Work for Farmers and Business

Designing for USSD is fundamentally different from designing for web or mobile apps. You’re working with:

• No visuals

• No persistent memory or history

• Character limits (usually 182 characters per screen)

• Strict session timeouts (often under 3 minutes)

• Linear interactions only (no backtracking, no multi-column logic)

All of this while preserving the business goals of Bodega: improve conversion, simplify the ordering process, and build trust across dealer networks.

How might we help farmers place an order without needing internet or a smartphone?

How might we design a flow that works for users with low literacy and tech exposure?

How might we make a 6–7 step purchase flow feel simple and secure?

How might we preserve business rules like pricing, inventory, and dealer routing inside a stripped-down interface?

Goals and Strategy

Enable non-smartphone users to access the Bodega marketplace

This meant making the experience easy to use through just a keypad, without assuming prior digital knowledge.

Simplify product ordering for low-literacy users

Use clear language, minimal steps, and logical menu structures to avoid cognitive overload.

Support Yara’s mission of agricultural accessibility

By expanding the platform to non-digital-first users, we helped Yara scale inclusively and reach more of its target market.

Process: Cross-Continental Collaboration, Local Understanding

As the UX/UI Designer on the project, I worked across teams from Kenya, India, Singapore, Tanzania, Indonesia, and Thailand. This included close collaboration with product owners, developers, QA, commercialization, and localization teams.

We began by reviewing:

• Feature phone usage patterns in Kenya

• How USSD services (like M-PESA) trained user expectations

• Regional language and literacy considerations

• Limitations of telco USSD platforms

Using this research, we mapped out key journeys:

• Product ordering

• Location selection

• Dealer assignment

• Order confirmation and payment prompts

Wireframes took the form of linear scripts—each screen written out with character counts and timed flows. Every iteration was reviewed for tone, clarity, and speed.

Key Features and Design Decisions

Menu structure optimized for decision speed

We used short menus (no more than 4–5 options per screen) and placed the most common actions first. For example:

1. Buy fertilizer

2. Check prices

3. Find dealer

4. Exit

We avoided deep menus and kept the full purchase flow to under 6 steps where possible.

Context-aware ordering

Based on the user’s location input, we dynamically routed them to the nearest Bodega-approved dealer. Inventory and pricing could shift based on this, but the user only saw simple confirmations- never raw database logic.

Built-in failsafes for session timeouts

To reduce friction, we designed checkpoints that saved inputs mid-session and allowed users to complete an order even if the connection dropped halfway through.

Language and literacy consideration

We used plain, localized Swahili and English with no jargon. Messages were structured like spoken instructions, not abstract UI labels.

Results and Impact

The USSD product was piloted across regions of Kenya and quickly gained traction among rural farmers. Key outcomes included:

• Successful integration with Bodega’s existing logistics and dealer network

• Significantly expanded reach to non-smartphone users

• Improved order volume in previously underserved regions

• Positive feedback from field agents and users who found the flow “easy to follow” and “faster than walking to a store”

The project also strengthened Yara’s position as a leader in inclusive agri-tech—showing that access isn’t just about apps but about infrastructure and intention.

What I Learned

Designing for USSD is one of the most constraint-heavy UX challenges—and one of the most rewarding.

Lessons that stuck with me:

• Simplicity is a skill, not a default—especially under technical constraints

• Designing for low literacy requires more empathy, fewer assumptions, and tested copy

• Success often depends on alignment with local norms and existing mental models (like how M-PESA trains users)

• Collaborating with cross-continental teams can be a strength when aligned around the user

Most importantly, this project reminded me that inclusive design isn’t about targeting edge cases - it’s about designing for the majority in underrepresented contexts.

Conclusion

YaraBodega USSD brought agricultural e-commerce to farmers who were otherwise excluded by the digital divide. It wasn’t flashy, but it was powerful—proof that great UX can exist even in 182 characters at a time.

Curious how this applies to your project?

Let’s chat.