ML tool that predicts SME digital payment adoption for Absa Bank Kenya & Airtel Money Kenya
# 🇰🇪 SME Digital Payment Adoption Predictor
> Built specifically for the **Absa Bank Kenya & Airtel Money Kenya**
> strategic partnership on digital payments for SMEs.
## 🎯 Project Overview
Millions of Kenyan SMEs still rely on cash for daily transactions despite
the availability of digital payment infrastructure. This project uses
machine learning to predict which SMEs are most likely to adopt digital
paybill payments — helping financial institutions like Absa and Airtel
target the right businesses with the right interventions at the right time.
**Live Dashboard:** ("
sme-payment-adoption-predic…")
## 💡 Business Context
| Company | Relevance |
|---|---|
| **Absa Bank Kenya** | Investing KES 3 billion annually in digital banking. Paybill 303030 is a key growth channel for SME payments |
| **Airtel Money Kenya** | Recently separated into Airtel Money Kenya Limited. Partners with Absa to allow payments via *334# and My Airtel App with 100% cashback incentives |
Instead of marketing digital payments to every SME blindly — which is
expensive — this model helps both companies prioritize the highest
probability adopters, saving marketing budget and improving conversion rates.
---
📊 Model Performance
| Model | Accuracy |
|---|---|
| Logistic Regression | 86.00% |
| **Random Forest** | **89.00% ✅ Best Model** |
### Detailed Scorecard (Random Forest)
| Metric | Non-Adopters | Adopters |
|---|---|---|
| Precision | 86% | 90% |
| Recall | 74% | 95% |
| F1-Score | 79% | 93% |
### Key Finding
Mobile money transaction frequency is the single strongest predictor
of SME digital payment adoption — stronger than revenue, business age,
or location. When the model predicts an SME will adopt, it is right
90% of the time, and it catches 95% of all actual adopters. This is
a directly actionable insight for Absa and Airtel's marketing teams.
---
## Feature Importance
The top drivers of SME digital payment adoption discover …