Building for Impact: Lessons from Scaling EdTech in Rural Egypt

When Vodafone Foundation approached RTG to build Ta3limy — an educational platform for underserved K-12 students across rural Egypt — we were excited and, if we're honest, a little overconfident. We had built consumer apps, government platforms, and enterprise tools. How different could it be? The answer: profoundly different, in ways that taught us more about product design than a decade of commercial work combined.

Designing for Constraints, Not Around Them

The average internet speed in rural Egyptian governorates is between 1–4 Mbps, frequently interrupted, and often shared across multiple devices in a household. Our instinct was to build a lightweight progressive web app and call it done. What we learned is that constraint-driven design is not about stripping features — it's about rethinking what features mean. Video content was compressed to levels that felt aggressive in a conference room but played flawlessly in a Minya classroom. Offline-first architecture meant lessons continued even when connectivity dropped. Audio-supplemented interfaces meant a student who struggled with reading could still engage with content. Every design decision started with 'what does this mean in the hardest conditions?'

The Literacy Interface Problem

Consumer tech assumes a baseline of digital literacy that does not exist across large segments of rural Egypt — or rural anywhere. When we user-tested early prototypes with students and parents in Upper Egypt, we found that icon-based navigation created profound confusion, that multi-step onboarding flows were genuinely impenetrable, and that voice-guided instruction dramatically improved task completion rates. We rebuilt the onboarding from scratch three times. The final version had a single screen, a single input (a mobile phone number), and an audio walkthrough in Egyptian Arabic. Completion rates went from 34% to 89%.

Working with Mission-Driven Stakeholders

Commercial clients optimize for growth metrics. Mission-driven stakeholders optimize for impact — and the two are not always aligned in the short term. Vodafone Foundation rightly pushed back when we proposed features that would improve engagement numbers but not learning outcomes. This discipline made the product better. We learned to design features with dual metrics: engagement (does the student use it?) and efficacy (does the student learn from it?). The lesson generalises: the best products are built by teams that agree on what success actually means.

Scale Changes Everything

Designing for millions of students in diverse geographic and socioeconomic contexts exposed edge cases that careful planning cannot anticipate. We deployed to a limited pilot, instrumented everything, and used real behavior data to iterate before scaling. Features we were proud of were abandoned because students in Aswan used them differently than students in Alexandria. The humility to let real users teach you is not optional when building at social scale — it is the methodology.