Last year, a Nairobi-based logistics founder paid KSh 2.4 million for a custom system. Eighteen months later, he was still using spreadsheets. The developer had moved to a “bigger project.” The code was a black box. The boss had no idea what his business actually owned.
It’s the kind of story we hear every month at Savannah Software Solutions — and it’s exactly why we’ve spent the last several years obsessing over how software gets built for Kenyan businesses, not just what gets built.
This is a behind-the-curtain look at the principles, the fights we have internally, and the seven things the smartest companies in Nairobi demand before they sign a single cheque. If you’re a Kenyan founder, operations lead, or finance manager about to commission software — read this first.
The Quiet Disaster Costing Kenyan Businesses Millions
Let’s name the pain. It’s not glamorous, but it’s costing Kenyan SMEs more than cybercrime does.
It’s the gap between what a business actually runs on and what the software was built to do.
Picture this. A mid-sized distributor in Mombasa wants to track inventory across three branches. They hire a developer. Three months in, the system can’t talk to M-Pesa for supplier payments. Another two months, the Kenya Revenue Authority eTIMS integration is half-done. By month eight, the original developer is gone, the new one is “rebuilding” what already exists, and the director is still approving purchase orders by WhatsApp screenshot.
Sound familiar?
We’ve met a Mombasa hotel chain that paid for a booking engine that couldn’t accept reservations from Kenyan mobile numbers. We’ve met a Nakuru agribusiness whose KSh 8 million ERP crashed every harvest season because nobody had asked about offline mode. We’ve met a Nairobi fintech founder who only discovered — after launch — that his app didn’t meet the Communications Authority data localisation rules.
None of these failures were about bad code. They were about not asking the right questions before the first line was written.
That’s the crisis. And it’s the reason we’ve built our entire delivery process around a single rule: the most expensive software in Kenya is the software that gets built wrong.
Inside Savannah: The 7 Things We Refuse to Compromise On
When a Kenyan business walks through our (virtual) door, we don’t start with timelines. We start with arguments. Internal arguments. The kind that take an extra week up front and save a client KSh 4 million twelve months later.
These are the seven non-negotiables.
1. We Start With the Ugly Truth, Not the Pretty Mockup
Most dev shops in Nairobi lead with screens. Look how nice this dashboard looks. Look at the colours.
We lead with process maps, not pixels.
Before we draw a single wireframe, we sit with the operations manager, the finance team, the warehouse guys — anyone who actually touches the work. We draw out how stock actually moves through a Nairobi warehouse at 6 a.m. We trace how an invoice really gets approved (hint: not the way the org chart says). We follow the boda boda dispatch sheet, not the system of record.
Why? Because the software has to fit the business, not the other way around. If we build for the org chart, we will lose the people doing the work by week two.
- We reject the “let me just clone a SaaS template” approach.
- We reject “we’ll figure out the workflow later.”
- We reject any project where the client can’t name their three biggest operational headaches in plain English.
It’s slower. It feels uncomfortable. It is the single biggest reason our software actually gets used.
2. We Design for the KENYAN Reality — Not the Silicon Valley Fantasy
Here’s a confession: most software written for African businesses has been written by people who’ve never queued at a Safaricom shop during a token crisis.
Our clients live in a world of:
- Power cuts in Industrial Area that last four hours.
- 3G fallbacks on the outskirts of Kisumu.
- USSD workflows for the 40% of customers who don’t have smartphones.
- M-Pesa prompts that decide whether a sale happens or dies.
- eTIMS compliance that the KRA can audit tomorrow morning.
So when we design, we ask: does this work when the lights go off? Does it degrade gracefully? Can a field agent use it on a Tecno Spark with 1GB RAM and a flaky signal?
For one of our Nairobi retail clients, that meant building a POS that queues transactions locally and syncs to the cloud once a connection returns. For a Kilifi fisheries client, it meant designing a fish-weight capture flow that works in bright sun, with wet hands, on a phone screen that’s seen better days.
This is not a “nice to have.” This is the difference between software that runs in Kenya and software that survives in Kenya.
3. We Treat KSh as a First-Class Citizen — Not an Afterthought
Try this. Go to a globally popular accounting system. Try to set up Kenyan VAT, NHIF, NSSF, PAYE, and the dozens of statutory deductions the KRA expects.
Now try to do it without losing your mind.
Most platforms weren’t built with Kenya in mind. They were built, and then somebody added “international” as a tab.
At Savannah, we do the opposite. Every financial module we design is born Kenyan:
- Multi-tier VAT logic (16% standard, 8% petroleum, 0% exports, exempt categories).
- KRA eTIMS integration for real-time invoice submission.
- NHIF and NSSF calculations that match the latest statutory tables.
- PAYE computations that handle reliefs properly — including the personal relief, insurance relief, and mortgage interest.
- Multi-currency support where KSh is the default, not an option buried in settings.
If your software cannot produce a KRA-compliant invoice in two clicks, your software is not built for Kenya. It’s a template with a Kenyan flag sticker on it.
4. We Build for the People Who Will Actually Use It — Not the People Who Sign the Cheque
In almost every Kenyan boardroom, the person who approves the software budget is not the person who will use the software at 7 a.m. on a Monday.
That gap kills more rollouts than bad code ever will.
So we design for the receptionist, the cashier, the driver, the inventory clerk. We design for the person whose bonus doesn’t depend on the system working, but whose day becomes impossible if it doesn’t.
Concretely, that means:
- Training videos in plain English (and Swahili, when the team asks).
- UI flows that match how Kenyans already work — not how a Silicon Valley designer imagines they should.
- Error messages that say what to do next, not what went wrong technically.
- A “panic button” that any user can hit to call our support team directly.
One of our clients — a fast-growing SACCO in Thika — told us last quarter that their loan officer rejection rate dropped 31% in the first month after rollout. Not because the algorithm changed. Because the interface finally made sense to the people using it.
5. We Refuse to Lock You In — Ever
This is the hill we will die on.
You should own your code, your data, and your ability to leave.
Far too many software vendors in Kenya (and abroad) operate on a quiet hostage model. The contract is signed. The data is locked in a proprietary database. The APIs are undocumented. The exit fee is a polite note saying “migration not included.”
We think that’s a rotten way to do business.
What we offer every Savannah client:
- Full source code escrow from day one (or full transfer, if you prefer).
- Open documentation for every API and database schema.
- Standard data export in open formats (CSV, JSON, SQL dump) — always.
- A 30-day handover period where your internal team can shadow ours.
If you ever decide to leave us, you walk out with everything you need to keep running. We’d rather earn your business every quarter than trap you into a decade.
6. We Treat Security and Compliance as a Day-One Conversation, Not a Phase
Kenya’s Data Protection Act is not a suggestion. The Computer Misuse and Cybercrimes Act is not optional. CBK and CMA regulations are not paper tigers.
And yet, security is still the line item most often cut when budgets get tight.
Not here.
From the architecture phase, we build in:
- Role-based access control mapped to your actual org chart.
- Encryption at rest and in transit, using current best-practice algorithms.
- Full audit trails for every financial transaction (the KRA will ask).
- Data residency options for sectors where local hosting is mandatory.
- Penetration testing before launch — not after a breach.
We also have uncomfortable conversations early. Do you really need to store customer ID numbers in plaintext? Should that admin account be accessible from outside the office? What happens to the data if a staff member leaves under a cloud?
These questions are cheaper to answer on day 14 than on day 414.
7. We Measure Success the Way You Do — Revenue, Time, Compliance
Software vendors love to talk in vanity metrics. Users logged in. Buttons clicked. Dashboards rendered.
Our clients don’t get paid in clicks.
So we measure success the way a Kenyan CFO does.
When we ship a system, we agree upfront on what “winning” looks like. It might be:
- Reducing invoice processing time from 9 days to 2.
- Cutting stock-outs at a specific branch by 50% within one quarter.
- Achieving 100% eTIMS submission compliance before the next KRA deadline.
- Freeing up 18 staff hours per week currently lost to manual reconciliation.
- Increasing M-Pesa collection rate by 22% through a smoother checkout flow.
If we hit those numbers, we celebrate together. If we don’t, we don’t send an invoice for the final milestone. That’s not a marketing line. It’s written into our contracts.
What Smart Nairobi Companies Are Already Doing Differently
The forward-thinking businesses we work with share a few traits. They’re not the biggest names. They’re not the loudest LinkedIn posters. But they are quietly outpacing their competitors in 2025.
Here is what they’re doing:
They treat software as a business decision, not an IT decision. The MD, the CFO, and the head of operations sit in the kickoff meeting. Not just the IT manager.
They demand Kenyan context from day one. They walk away from vendors who can’t talk fluently about eTIMS, M-Pesa Daraja, NSSF, or the Data Protection Act.
They invest in change management. They budget 15% of the project cost for training, internal champions, and feedback loops. They know that the most elegant system fails if the team rebels.
They ask for the exit clause. They want to know, on day one, what happens if this relationship doesn’t work in 18 months. And they’re surprised — pleasantly — when a vendor answers that question with enthusiasm.
They measure outcomes, not outputs. They don’t care that we built 47 API endpoints. They care that the KRA audit took 2 days instead of 3 weeks.
These are the companies pulling away from their sectors in 2025. The rest are still arguing about colours.
What You Should Walk Away With Today
If you take nothing else from this behind-the-scenes look, take this:
Great Kenyan software is not about the latest framework. It’s about respect for the business, the people, and the country it has to run in.
The right partner will:
- Ask hard questions before quoting a price.
- Refuse to start until the workflow is mapped.
- Build for KSh, M-Pesa, KRA, and the matatu route your field agent takes.
- Treat your data as yours, always.
- Tie their fee to a result you can measure in your own books.
If your current vendor can’t tick those boxes, you don’t have a software problem. You have a partnership problem — and that’s the more expensive one to ignore.
Ready to Build Something That Actually Fits?
If you’re a Kenyan business owner, founder, or operations lead who is tired of software that doesn’t match how your team really works — let’s have a conversation before you sign anything.
The team at Savannah Software Solutions has helped dozens of Kenyan businesses — from logistics firms in Mombasa to SACCOs in Thika, from agribusinesses in Eldoret to fintechs in Westlands — design and build systems that their teams actually use, and that pass the KRA audit on the first try.
Visit savannahsoftwaresolutions.co.ke to book a free 30-minute discovery call. No pitch deck. No pressure. Just an honest conversation about what your business actually needs, and what it will take to build it right.
