USSD vs WhatsApp vs Apps: The Channel Economics

The Connectivity Divide

Before comparing channels, understand the baseline. Sub-Saharan Africa has:

• 25-28% mobile internet penetration — three quarters of the population cannot access apps or WhatsApp
• 960 million people in the "usage gap" — they live within mobile broadband coverage but don't use mobile internet (cost, device, literacy barriers)
• 40%+ on 2G networks — the radio infrastructure only supports voice and signaling, not data
• Device cost = 26% of monthly GDP per capita — the cheapest smartphone costs a quarter of an average monthly income

Any channel strategy that starts with "download our app" has already excluded the majority of the addressable market.

User Data Cost: Who Pays What

The user's data cost determines adoption friction. In markets where 1GB costs 5-10% of monthly income, every megabyte matters.

USSD is the only channel with zero user-side data cost. WhatsApp benefits from carrier social bundles (e.g., Safaricom's "WhatsApp Bila Bundle") but rich media conversations still consume data. Apps impose the highest burden — the download alone can cost more than a day's data budget for rural users.

Enterprise Cost Models

Each channel has a fundamentally different cost structure. The right choice depends on your volume, your users, and your margin.

USSD: Low CapEx, Linear OpEx

USSD infrastructure is cheap to start but scales linearly. Every session costs money. There's no marginal cost reduction at volume.

• South Africa benchmark: R850/month + R0.08-0.10 per session
• Nigeria NCC regulated rate: NGN 6.98 for a 120-second banking session
• Tanzania: Up to 75% of OpEx is session fees on time-sliced carriers
• Kenya (Africa's Talking): KES 1.00 per session on shared shortcodes

At 100,000 sessions/month in Kenya, USSD costs ~KES 100,000 ($770). At 1 million sessions, it's KES 1,000,000 ($7,700). Linear. Predictable. No economies of scale.

WhatsApp Business API: Per-Message, Category-Priced

Meta charges per conversation, categorised by type. Pricing varies by country and conversation category.

Native Apps: High CapEx, Near-Zero Marginal Cost

Apps are expensive to build and expensive to acquire users for, but once installed, each additional interaction costs almost nothing.

• Development: $50,000-500,000+ depending on complexity
• User acquisition: Cost Per Install (CPI) of $1-5 in African markets
• Ongoing: Server infrastructure, App Store fees (15-30% on transactions), version maintenance
• Marginal cost: Near zero per additional user interaction

At scale, apps are the cheapest channel per interaction. The problem is reaching scale. An app with 10,000 installs in Kenya cost $10,000-50,000 in acquisition marketing alone. And 70% of those users may never open it twice.

Deployment Speed

USSD's killer advantage is deployment speed. A new betting market can be live in days. A new feature ships with a server-side code change. No app review. No template approval. No version fragmentation. When Kenya's BCLB changes compliance requirements overnight, USSD operators can update the journey in minutes. App operators need a release cycle.

Offline Capability

In rural Sub-Saharan Africa where data connectivity is intermittent, USSD is the only channel that works 100% of the time. A farmer in rural Tanzania with a Tecno T301 feature phone and no data plan can dial *149*10# and place a bet. That same farmer cannot use WhatsApp or an app.

WhatsApp in Low-Data Environments

WhatsApp's behaviour in African markets creates unexpected failure modes:

• Data toggling: Users turn mobile data on only when needed, then off immediately to preserve their bundle. WhatsApp messages sent while data is off queue silently and may arrive hours or days late
• TTL vulnerability: Authentication messages have a 10-minute TTL. Utility messages expire after 30 days. If the user's data is off when the message is sent, it expires before delivery
• Algorithmic throttling: Meta imposes a 30-second delay per message during bulk sending. A "flash sale" notification to 50,000 users takes 17+ hours to deliver. By then, the event is over
• Tier limits: New WhatsApp Business API accounts start at 1,000 messages/day and must "level up" through engagement quality metrics. You can't launch at scale

The Hybrid Playbook

Winning operators in Africa don't pick one channel. They deploy all three with a clear role for each:

• USSD: Acquisition channel and universal fallback. Every user can reach it. Handles registration, first deposit, first bet. Falls back to USSD when app connectivity fails
• Native App: Retention and deep engagement. Live odds, in-play betting, cash-out, loyalty programmes. For users who have smartphones and data
• WhatsApp: Push notifications and customer support. Bet settlement alerts, promotional offers, complaint resolution. Not a transaction channel

The USSD layer is the foundation. It's the only channel that reaches 100% of the mobile subscriber base. The app and WhatsApp layers sit on top for users who can access them.

How This Plays Out in Practice

M-Pesa: USSD to Super App

Safaricom launched M-Pesa on USSD via STK in 2007. By 2024, the M-Pesa Super App has 30+ million downloads. But the STK/USSD channel still processes the majority of transactions. The app didn't replace USSD — it layered on top. M-Pesa's "Offline Mode" queues transactions locally and syncs when connectivity returns, acknowledging that even smartphone users lose connectivity.

MTN MoMo: True Omnichannel

MTN MoMo operates across USSD (*165#), the MoMo app, WhatsApp integration, and web. The USSD channel handles the highest transaction volume because it reaches users the app can't.

Equity Bank: EVA WhatsApp Bot

Equity Bank Kenya deployed EVA, a WhatsApp-based banking assistant. It handles balance inquiries and mini-statements. But actual transactions (transfers, payments) still route through USSD or the Equity app. WhatsApp is the engagement layer, not the transaction layer.