Back to Work Case Study · Mobile · Productivity

TaskForge: How a 12-person agency cut task chaos in 8 weeks

Duration: 8 weeks Team: 2 engineers 2025 Lagos, Nigeria

The Challenge

Chidi ran a small design agency with 12 freelancers scattered across Lagos and Accra. They tracked tasks in WhatsApp groups, project deadlines in spreadsheets, and client feedback in email threads. The result: lost billable hours, missed deadlines, and constant "did you see my message?" pings.

Off-the-shelf tools like Asana and Trello didn't fit their workflow — the team needed something lightweight, mobile-first, and reliable on poor Nigerian network connections. Custom builds quoted elsewhere started at ₦8M with a six-month timeline and a team they didn't trust.

My Role

I was the sole engineer on TaskForge. I handled architecture, backend design, mobile app development, and deployment. Chidi handled product decisions and design reviews.

Process Highlights

Key Decisions

1. Firebase over a custom backend. The team needed a working product in 8 weeks, not 16. Firebase gave us real-time sync, auth, and push notifications out of the box. We accepted less query flexibility in exchange for speed — and it paid off.

2. Offline-first architecture. Most team members worked from phones on unreliable connections. We built local persistence with background sync so the app stayed usable with zero network. This was the single most impactful decision.

3. Three-tap task creation. Every feature request that added friction to task creation was rejected. The core loop had to feel instant. This shaped the entire UI.

Results

4,200+Active users in 90 days
4.7★App Store rating
18%D30 retention rate

Within three months of launch, TaskForge had 4,200+ active users. The agency cut task-related meetings by roughly 60%. They've since onboarded two partner agencies onto the platform.

Learnings

If I rebuilt TaskForge today, I'd add analytics from week one. We added it in week six and missed critical early usage patterns that would have shaped the onboarding flow. That's the single biggest change I'd make.

I'd also push harder for a web version early. A surprising number of users asked for desktop access — something we didn't anticipate.

Have a similar project in mind?

I'm currently accepting new work for Q1 2026.