Safaricom Baze — Creator & Admin Platform
Designed the creator tools and admin backend powering Safaricom's video and music streaming platform — from content upload and monetization to catalogue management and billing across millions of users.
Baze
Platform
2
Systems Designed
6 mo.
Duration
80+
Screens Designed
Overview
Baze is Safaricom's mobile-first video and music streaming platform. Behind the consumer experience sit two operational systems: the creator tools (how content gets onto the platform) and the admin backend (how Safaricom manages it).
I designed both — working directly with engineering and PMs as the sole designer on this workstream.
Note
Production designs remain confidential. Full case study available via BBR.
The Problem
Creator Side
- Upload requirements weren't clear upfront, leading to rejected submissions
- Creators had no visibility into earnings or content performance
- No scheduling — no way to control when content went live
- No actionable feedback when content was rejected; just 'not approved'
Admin Side
- Content review was a bottleneck with no structured queues or escalation paths
- Catalogue tagging was inconsistent — downstream search, filtering, and recommendations all suffered
- Reporting was scattered across multiple disconnected systems
- Creator management, earnings, and compliance status lived in separate places
Creator Platform
Staged Onboarding
Progressive registration that collects what's needed in phases — not one overwhelming form. Creators reach their first upload faster.
Upload with Validation
Client-side checks on format and quality before submission, with metadata capture and draft saving. Rejection rates drop because problems are caught before they reach review.
Content Pipeline Tracker
Creators see exactly where their content is at every stage: uploaded → in review → approved → scheduled → live. No more guessing.
Scheduling & Publishing
Creators control when content goes live, with a calendar view across all their content. Strategic publishing, not just upload and hope.
Earnings & Analytics
Revenue dashboard with the 60/40 split made transparent, payout tracking, and per-content performance data. First time creators could see exactly what they were earning and why.
Actionable Moderation
Rejections come with specific reasons and a clear resubmission path. Not just a status — a next step.
Admin Backend
Catalogue Management
Advanced filtering, batch operations, and standardised metadata templates that enforce consistency at scale. Fix the tagging problem upstream — everything downstream gets better.
Review Workflows
Structured queues with priority sorting, review checklists, and escalation paths. Review stops being a bottleneck when it has a shape.
User Management
Unified creator profiles connecting content, earnings, compliance status, and subscriber impact in one view. The team stops switching between tools to understand a creator.
Analytics
Platform-wide dashboard with drill-down from overview to genre to creator to individual content. Every stakeholder gets the data at the right level of detail.
Billing
Subscription management, creator payout processing, and reconciliation views. Operational finance visible in the same system as content and creators.
The Impact
- Designed the full operational infrastructure behind a platform serving Safaricom's subscriber base
- Replaced scattered tools with a unified backend for content, creators, subscribers, and billing
- Gave creators earnings transparency for the first time — a differentiator against platforms where payouts are opaque
- Structured review workflows that scale from hundreds to hundreds of thousands of content submissions
What I Learned
- Two-sided platform design means every decision has two audiences — a moderation rejection is a workflow for the admin and a moment of frustration for the creator
- Designing at Safaricom's scale means thinking in systems, not screens — every pattern needs to survive 100x growth
- The highest-impact design decision was standardised metadata templates — consistent metadata made search, filtering, recommendations, and analytics all work better downstream
Project Gallery