Home About Services The Problem Insights Contact
Data Transformation · East Africa

Your business data
exists. Your

Datakafe Intelligence builds the transformation infrastructure that turns your scattered M-Pesa records, ERP exports, and spreadsheets into decisions your leadership team can trust — in hours, not weeks.

Specialist in East African data
FintechSACCOAgriRetail
Reality for most Kenyan businesses
0
Average time to answer a basic question like "which product is most profitable?"
0
Of analyst time spent cleaning data instead of generating insight for decisions
0
Disconnected systems on average — QuickBooks, M-Pesa, spreadsheets, CRM, all unconnected
Scroll
We work with
Fintechs & Lenders SACCOs & MFIs Retail & E-Commerce Agri-Platforms Health & Logistics
The Problem

Six signs your data infrastructure is broken

These are not separate problems. They are six faces of one root cause — the absence of a transformation layer between your raw data and your decisions. Most Kenyan medium enterprises have all six right now.

If this sounds familiar — you have the problem
Finance and sales report different revenue every month
Your analyst spends most of their week cleaning data
Reports can't be trusted unless one person verified them
Leadership questions take days or weeks to answer
A spreadsheet is your most important data infrastructure
01

Metrics That Disagree

Finance and Sales report different revenue numbers. Nobody knows which is right. Trust collapses. Decisions revert to gut feel.

02

The Human ETL Trap

Your best analyst spends 80% of their week cleaning data. They're a bottleneck. When they leave, everything breaks.

03

Silent Data Failures

Pipelines break quietly. Dashboards update with wrong numbers. You discover it in a board meeting, not from an alert.

04

Knowledge Locked Away

One person knows every edge case in your pipeline. When they leave, months of data intelligence leaves with them.

05

The Load-Bearing Sheet

A Google Sheet that started as a quick fix is now critical infrastructure — undocumented, untested, one mistake from disaster.

06

Slow Answers, Fast Market

The CEO asks Monday. The answer arrives Thursday. Competitors who answer in hours are already acting.

All six problems have the same root cause: no transformation layer between your raw data and your decisions.
How We Work

From scattered data
to decisions you trust

01

Data Audit & Discovery

We map every source — ERP, M-Pesa, CRM, spreadsheets — identify what's broken and missing. The blueprint for everything.

Weeks 1–2
02

Transformation Layer Design

We design version-controlled data models encoding your business logic — one agreed definition of every core metric.

Weeks 2–4
03

Build, Test & Document

Every model ships with quality tests baked in. Every pipeline has alerts. Nothing is a black box. Bad data never reaches a report.

Weeks 3–6
04

Handover & Enablement

We train your team. Knowledge lives in documented code — not our heads. You own everything we build.

Weeks 6–8
What we build for you
📥
Raw InputsM-Pesa · ERP · CRM · Excel
⚙️
Transformation LayerCleaned · Tested · Documented
Quality GatesAutomated tests · Alerts · Lineage
📊
Trusted OutputsReports · Dashboards · Decisions
Our Services

High-impact engagements built for East Africa

⚙️
01

Data Foundation Sprint

6–8 weeks. Takes you from raw warehouse to a clean, tested, documented transformation layer. The core of what we do.

Learn more →
🔍
02

Data Health Audit

Deep diagnostic of your current infrastructure. We surface transformation debt and give you a clear, prioritised roadmap.

Learn more →
📱
03

M-Pesa Data Modeling

Specialists in turning M-Pesa transaction data into structured models that reveal behaviour, patterns, and risk signals.

Learn more →
🏗️
04

Ongoing Data Engineering

A senior data engineering partner without the full-time hire. We extend your transformation layer as your business grows.

Learn more →
Who We Serve

Built for growing East African businesses

Fintechs & Lenders
Turn M-Pesa flows, loan records, and repayment data into risk models and product insights that actually inform decisions.
SACCOs & MFIs
Regulated, auditable pipelines with tested data lineage. Reports that hold up to scrutiny. Infrastructure that replaces spreadsheets.
Retail & E-Commerce
Unified models across SKUs, channels, and fulfilment systems. See what's selling, what's not, and why — in one place.
Agri-Platforms
Complex supply chains, field agent data, and payment corridors modeled into performance visibility at every level.
Health & Logistics
Operational data from facilities, routes, supply chains — transformed into branch, route, and unit-level performance models.
The Datakafe Promise

"We don't just connect your data. We build the foundation that makes every business question answerable — in hours, not weeks — and that stays working long after we leave."

John Waweru · Founder, Datakafe Intelligence · Nairobi

Let's Talk

Ready to trust your data?

Tell us what your data situation looks like. We'll tell you exactly what's holding you back — and what it would take to fix it.

  • No complicated sales process
  • Response within 24 hours
  • Free initial data health assessment
Send us a message
✓ Message received. We'll be in touch within 24 hours.
About Datakafe

We exist because East Africa's data problem is real — and largely invisible.

Most businesses in Kenya are generating more data than ever. Almost none of them can actually use it.

6–8wk
Time to build your data foundation
100%
Ownership transferred to your team
East Africa
Focused exclusively on this market
0
Black boxes delivered. Everything documented.
Our Mission

Make East African business data answerable.

We believe every growing business in this region deserves data infrastructure that works — tested, documented, and actually usable for decisions.

The modern data stack has transformed how companies in the US and Europe operate. We bring that same infrastructure — adapted for East African realities, M-Pesa data patterns, and local tech ecosystems.

What we stand for
No Black BoxesEvery model is documented, readable, and owned by your team.
Local Context FirstM-Pesa structures, EAT timezone, Kenyan business realities — built in from day one.
Testing as StandardWe never deliver untested pipelines. Quality gates are baked in, not bolted on.
Ownership TransferWe succeed when your team runs the infrastructure without us.
Our Values

How we work and why it matters

🔍

Radical Transparency

We explain every decision. You should never wonder what's happening inside your own data infrastructure.

🏗️

Build to Last

Infrastructure your team can operate and extend is worth more than something only we understand.

🎯

East Africa First

We adapt data engineering patterns for the data sources and business realities of this region.

Speed of Insight

Every decision we make prioritises one outcome: how fast your team can answer questions after we leave.

Test Everything

If it hasn't been tested, it hasn't been finished. We treat untested pipelines like unreviewed code.

🤝

Ownership Transfer

Our job is to make ourselves unnecessary. We succeed when your team runs everything without us.

The Founder

Built by someone who has seen this problem up close.

JW
John Waweru
Founder & Lead Data Engineer · Datakafe Intelligence

John founded Datakafe Intelligence after spending years watching growing East African companies make critical decisions on untrustworthy data — not because they lacked talent or intention, but because nobody had ever built the transformation layer between their raw systems and their reports. Datakafe is his answer to that gap: a specialist firm that builds what most companies skip, in the way it should have been built from the start.

Our Services

Four ways we fix your data infrastructure

Focused, high-impact engagements. Clear scope, clear outcomes, clear ownership.

01
Core Service

Data Foundation Sprint

⏱ 6–8 Weeks

Takes your business from raw, siloed data to a clean, tested, version-controlled transformation layer that the whole organisation can trust. By the end, your team has one agreed definition of every core metric. Every report traces back to documented, tested code.

  • Full data source audit
  • dbt transformation models
  • Automated quality tests
  • Agreed metric definitions
  • Pipeline alerting
  • Team handover session
02
Diagnostic

Data Health Audit

⏱ 1–2 Weeks

Before spending on infrastructure, understand exactly what is broken. A deep diagnostic of your current data setup — pipelines, models, documentation, and trust levels across your team. You receive a clear, prioritised report of your transformation debt.

  • Audit all existing pipelines
  • Documentation assessment
  • Metric consistency check
  • Single-person risk register
  • Prioritised fix roadmap
  • Executive summary
03
Specialist

M-Pesa Data Modeling

⏱ 2–4 Weeks

M-Pesa generates the richest transactional data in East Africa. Most companies that process it have no idea what to do with it beyond counting deposits. We take your raw Paybill, Till, B2C, and C2B data and build clean, queryable models that reveal customer behaviour, payment patterns, and risk indicators.

  • Multi-channel consolidation
  • Payment behaviour models
  • Cohort & retention analysis
  • Anomaly detection
  • EAT timezone normalisation
  • Warehouse integration
04
Ongoing

Ongoing Data Engineering

⏱ Monthly Retainer

A senior data engineering partner without the cost or complexity of a full-time hire. New data sources get modeled. New business questions get answered. Pipeline health gets monitored every month.

  • Monthly model additions
  • Pipeline monitoring
  • Incident response
  • Documentation updates
  • Monthly health review
  • Team Q&A support
The Problem We Solve

Most Kenyan businesses are data-rich and insight-poor.

The data exists. It's generated every day. But without a transformation layer, it stays raw, scattered, and unanswerable.

Growing businesses in East Africa are making critical decisions without reliable, timely, or trusted data — not because they lack data, but because nobody has built the infrastructure to connect, clean, and transform it into something answerable.

The result is a company where finance and sales report different revenue numbers, the best analyst spends most of their week cleaning spreadsheets, and answering a basic business question takes days or weeks. This is not a people problem. It is a data transformation problem — and it has a clear, buildable solution.

The Six Symptoms

Not six problems. One problem with six faces.

Every symptom below traces back to the same root cause: the absence of a clean, tested, documented data transformation layer.

01

Metrics That Disagree Across Teams

When different departments pull data from different sources using different logic, they get different answers. Finance uses confirmed receipts. Sales uses closed deals. Neither is wrong — but because nobody has agreed on a single, version-controlled transformation of raw data into business metrics, the numbers will never match. Leadership stops trusting the data. Decisions revert to gut feel.

Impact: Loss of data trust across the organisation
02

Analysts Doing Machine Work

In most Kenyan medium enterprises, the data analyst spends the majority of their time pulling exports, cleaning columns, joining tables by hand, and rebuilding the same base datasets from scratch for every new question. The analyst — hired to find insights — is trapped in a role that requires no analytical thinking at all. You're paying insight-level salaries for data-entry-level work.

Impact: Wasted talent, slow answers, single points of failure
03

Silent Failures in Reporting Pipelines

Most data pipelines were built to move data, not to verify it. When an upstream system changes — an API column name shifts, a timezone updates, a join key duplicates — the pipeline continues running silently with wrong data. The dashboard looks fine. The numbers are wrong. The first sign of the problem is a discrepancy in a leadership meeting, not an automated alert.

Impact: Bad decisions made on data that looked valid but wasn't
04

Institutional Knowledge Locked in One Person

The logic that powers your most critical reports often exists only in the mind of the person who built it. The edge cases, the filters, the workarounds — undocumented, unshared, unversioned. When that person leaves, the company loses months of accumulated data intelligence. The replacement spends weeks reverse-engineering work that should have been written down from day one.

Impact: Knowledge debt that costs months to rebuild
05

The Load-Bearing Spreadsheet

Almost every medium enterprise in Kenya has at least one Google Sheet or Excel file that started as a quick fix and became critical infrastructure. It has no version control, no testing, no documentation. Its logic is invisible unless you already understand it. Yet week after week, decisions worth millions of shillings are made based on what it outputs — and nobody has ever audited whether it's right.

Impact: Fragile, unauditable infrastructure hiding in plain sight
06

Slow Answers in a Fast Market

East Africa's business environment is moving quickly. Companies that can answer strategic questions in hours have a meaningful competitive advantage over those still waiting two weeks for a manually assembled report. Every month without clean data infrastructure is a month of slower decisions, missed signals, and compounding competitive disadvantage.

Impact: Strategic decisions made too late, or not at all
The Real Cost

How long does it actually take?

For a typical Kenyan medium enterprise, here is the honest gap between how fast business questions should be answered and how long they actually take.

Business Question Should Take Actually Takes
What was our revenue last month? Minutes 1–3 days
Which product line is most profitable? Hours 1–2 weeks
What is our customer retention rate? Hours 2–4 weeks (if ever)
Which sales rep performs best, by region? Hours 1 week
What is our invoice collection period? Minutes 3–7 days
Where are we losing customers in the funnel? Hours Never answered properly
What does cash flow look like in 60 days? Hours Done manually, monthly only
Which branch is actually profitable after costs? Hours 2–3 weeks
The Fix

The answer is a data transformation layer — built right.

A clean, tested, documented, version-controlled set of transformation models that encode your business logic once — and make every question answerable from that point forward.

  • One agreed definition of every core business metric
  • Every report traces to tested, documented code
  • Pipelines alert when something breaks before it reaches a dashboard
  • New team members onboard to the data in days, not months
  • Analysts spend their time analysing, not cleaning
  • Answers that take hours, not weeks
Before vs After Datakafe
📊
Metric ConsistencyOne definition. Every team. No arguments.
Answer SpeedHours instead of weeks. Every time.
🛡️
Data QualityAutomated tests catch problems before reports do.
📖
DocumentationEvery model readable by anyone on the team.
🔄
ResilienceNo single person holds the keys to your data.
Insights & Ideas

The data transformation problem — named, explained, solved.

Practical writing on data infrastructure for East African businesses. No jargon. No vendor pitches.

01
Featured Post
Your Finance Team and Your Sales Team Are Looking at Different Revenue Numbers
Data Transformation · 8 min read

And both think they're right. One is presenting to the board next week.

This is not a people problem. It is not a communication problem. It is a data transformation problem — and it is silently happening inside your company right now.

Read Post →
Analyst Productivity
Your Best Analyst Is Not Analysing. They Are Cleaning Data.
You hired a smart analyst. But 80% of their week is work a computer should do.
6 min read
Data Quality
Your Dashboard Is Lying to You. Politely. Every Day.
The numbers look fine. But there's a real chance they've been wrong for weeks and you have no way to know.
7 min read
Knowledge Management
Your Data Pipeline Is a Black Box. Only One Person Knows What's Inside.
What happens when that person leaves? Critical data logic lives in one person's head in most Kenyan businesses.
5 min read
Speed of Insight
Why It Takes Two Weeks to Answer a Question Asked on Monday
The CEO asks a question. The analyst says "give me two weeks." That pause costs more than anyone realises.
6 min read
Infrastructure Risk
The Spreadsheet That Everyone Edits and Nobody Owns
There is a Google Sheet. It started as a quick fix. Now it's the source of truth for your entire operations review.
7 min read
M-Pesa Data
The M-Pesa Data Problem Nobody Is Talking About
M-Pesa generates the richest transactional data in East Africa. Most companies have no idea what to do with it.
8 min read
Cloud Strategy
You Have a Data Warehouse. You Do Not Have a Data Strategy.
Moving to BigQuery was right. But a warehouse without a transformation layer is a library with no shelves.
6 min read
Tools & Strategy
Why Your New BI Tool Will Not Fix Your Data Problem
A better dashboard on top of bad data is a more beautiful lie. The tool was never the problem.
5 min read
Leadership
The Data Team Every Kenyan CEO Actually Needs
Being data-driven is not about hiring analysts or buying tools. It's about the infrastructure that makes it possible.
9 min read
Stay Informed

Get new insights in your inbox.

One email when we publish. No noise, no selling. Just practical writing on data infrastructure for East African businesses.

Get In Touch

Let's talk about your data situation.

No long sales process. No pitch deck. Just an honest conversation about what's broken and what it would take to fix it.

What to Expect

We keep things simple.

Tell us what's going on with your data. We'll respond within 24 hours with either a direct answer, a clarifying question, or a proposal for a free data health conversation.

📍

Location

Nairobi, Kenya — serving businesses across East Africa

🌐

Website

datakafe.co.ke

⏱️

Response Time

Within 24 hours on business days

🎯

Best Fit

Growing businesses with 20–500 staff and real data challenges across their systems

What happens after you reach out
  • We read every message personally — no auto-responses
  • If it's a fit, we'll schedule a 30-min discovery call
  • We'll give you an honest assessment of your situation
  • You'll leave with a clear picture of what's broken and what it takes to fix it
  • No obligation to proceed further
Send us a message
Tell us about your data situation in as much or as little detail as you like.
✓ Message received. We'll be in touch within 24 hours.
Common Questions

Frequently Asked

How do I know if I need the Audit or the Foundation Sprint first?
If you have a vague sense that your data is broken but can't articulate exactly what's wrong, start with the Audit. It gives you a clear picture and a prioritised roadmap. If you already know the problems and want to fix them, the Foundation Sprint is the right engagement.
Do we need to already be on a cloud data warehouse?
Ideally yes — BigQuery, Snowflake, or Redshift. If you're not yet, we can advise on the right path. Some of our work involves helping companies migrate from on-premise setups to cloud warehouses before building the transformation layer.
What happens to our data work after the engagement ends?
Everything we build is yours. All code is version-controlled and transferred to your team's repository. All documentation is written for your team to maintain. We do a full handover session so your team understands everything we built.
We have M-Pesa data but it's a mess. Can you work with that?
Yes — this is one of our specialities. M-Pesa data arrives in multiple formats from multiple channels (Paybill, Till, B2C, C2B) and has significant normalisation work required. We've modeled this data specifically for East African business contexts.
How long does the Foundation Sprint actually take?
6–8 weeks is the standard range. The variation depends on the number of data sources, the complexity of your business logic, and the availability of your team for reviews. We'll give you a more precise estimate after the discovery conversation.
What tools do you use?
We primarily work with dbt for transformation, on top of BigQuery, Snowflake, or Redshift. For orchestration we use Airflow or Prefect. We'll always recommend the right tool for your specific situation — not just what we're most comfortable with.