By: Timothy Fadipe
In the last five years, the African tech ecosystem has been obsessed with one metric. That metric is user growth. The logic was simple. If you acquire enough users, the revenue will follow. Founders burned millions of dollars on Facebook and Google ads to show investors a chart that moved up and to the right.
But as we settle into the new economic reality of late 2025, that logic has collapsed. We are seeing startups with millions of registered users shut down because they cannot pay their server bills.
The problem is not that these companies failed to grow. The problem is that they grew the wrong way. They optimised their marketing for Cost Per Acquisition or CPA, instead of optimising for Payback Period.
The math of failure
Here is the scenario I see constantly. A Fintech startup raises a seed round. They hire a performance marketer and tell them to get users for under $2. The marketer floods the funnel with cheap leads from TikTok or display networks. On paper, the campaign is a massive success. They acquired 100,000 users for $200,000.
But when a Growth Engineer looks at the SQL database three months later, the reality is terrifying. Those cheap users have a 95% churn rate. They are not transacting. They are ghost accounts.
The company did not spend $200,000 to grow. They spent $200,000 to rent traffic that added zero enterprise value.
Shifting to cohort economics
To survive the current funding winter, African startups must stop looking at aggregate growth and start looking at Cohort Analysis.
This requires a fundamental shift in data infrastructure. You cannot do this with Google Analytics default reports. You need to warehouse your data. You need to join your marketing spend tables with your product usage tables.
When you do this, you can answer the only question that matters. How long does it take for a new user to pay back the cost of acquiring them?
If the answer is more than 12 months, you should not be spending money on ads. You should be fixing your product retention.
Predictive modeling over volume
The most sophisticated companies are now moving towards Predictive Lifetime Value (LTV). Instead of waiting six months to see if a user is valuable, we use machine learning models to analyse their first 24 hours of behavior.
Did they complete the KYC process? Did they connect a bank card? Did they log in more than twice?
By weighing these actions, we can predict the future value of that user with 80% accuracy. This allows us to tell the ad platforms to find more users like the top 1% of our customer base, rather than finding the cheapest users available.
The hard truth
Growth is seductive. Profitability is hard.
For a long time, founders could get away with vanity metrics because capital was cheap. That era is over. The marketing teams that will win in 2026 are not the ones with the best creative ideas. They are the ones with the strongest grip on their unit economics.
It is time to stop celebrating the number of signups on the dashboard and start celebrating the quality of revenue in the bank.
About the Author:
Timothy Oluwapelumi Fadipe is a Growth Engineer and Performance Marketing Strategist. He has architected data driven growth engines for global startups in the AI, EdTech, and Fintech sectors, driving six figure revenue growth through proprietary attribution modeling and unit economics optimisation.
