A hardware shop in Nairobi called us last year. They wanted a website. Nice photos, a product list, a contact form. Standard job, standard price.
Ten minutes into the conversation the real story came out. Three branches. Stock counted by hand every Saturday. Prices that differed between branches because nobody could remember what had been agreed. A WhatsApp group where staff sent photos of handwritten orders. The owner driving between branches to check whether an item was actually in stock before promising it to a customer.
They did not need a website. A website would have looked lovely and changed nothing.
This is the most expensive misdiagnosis in Kenyan business technology, and it runs both ways. Businesses that need a simple site get sold expensive software they never open. Businesses drowning in manual work get sold a five-page brochure and wonder why nothing improved.
Here is how to tell the difference before you spend anything.
The distinction in one line
A website tells people about your business. A system runs part of your business.
A website is mostly read. Visitors arrive, look, and leave — hopefully after calling you. The content changes occasionally and one person controls it.
A system is mostly written to. People log in and change things. Stock moves. Orders progress. Invoices are issued. Bookings are made and cancelled. Multiple people, with different permissions, changing the same data at the same time — and the data has to still be correct afterwards.
That last sentence is the entire reason systems cost more. Everything else is detail.
The five-question test
Answer honestly. Count your yeses.
1. Does the same information live in more than one place? Stock in a notebook and in a WhatsApp group. Customers in your phone, in a Google Sheet and in your head. Prices on a printed list and in your memory. If information has to be manually copied from one place to another, you have a systems problem, and it is already costing you.
2. Does someone spend more than an hour a day on repetitive admin? Writing receipts. Reconciling M-Pesa messages against orders. Counting stock. Typing the same customer details into a third form. An hour a day is roughly 22 hours a month. At even a modest salary, that is real money leaving the business every month, forever.
3. Do you regularly not know something you should know? How much stock you have right now. Who owes you money. Which product actually makes you the most margin. Whether last month was better than the month before. If answering these takes more than five minutes, you cannot manage what you cannot see.
4. Do mistakes happen because of the process, not the people? Double bookings. Selling stock you had already sold. Missed follow-ups. Wrong prices quoted. Good staff making the same error repeatedly is not a discipline problem — it is a design problem.
5. Would growing 50% break something? If doubling your orders would require doubling your admin staff, your business does not scale. It just gets heavier.
Reading your score
0–1 yes: you need a website. Your operations are fine. What you lack is visibility. Spend KES 35,000–120,000 on a fast, well-built site with proper SEO and get on with it. Do not let anybody sell you software you will not open.
2–3 yes: you need a website plus one focused tool. Not a full custom platform — one specific thing fixed. Online bookings. A simple order form that feeds a dashboard. Digital invoicing. Solve the loudest problem first and see what changes. This usually lands between KES 90,000 and KES 180,000.
4–5 yes: you need a system. A website alone will make you look better while the underlying problem gets worse. Budget from KES 150,000 and hire someone who builds software rather than pages.
What "a system" actually means in practice
Vague word. Here is what it looks like for real Kenyan businesses.
A retail shop with three branches gets one stock database all branches read from, a POS that deducts stock as it sells, per-branch pricing controlled centrally, low-stock alerts, and a phone screen showing today's sales across all three. The Saturday stock-count disappears.
A school gets student records, fee tracking with automatic M-Pesa reconciliation, balance statements parents can view themselves, and a report that shows outstanding fees by class in two seconds instead of two days.
A salon or clinic gets online booking that respects each staff member's real availability, automatic SMS or WhatsApp reminders that cut no-shows, client history the practitioner can see before the appointment, and a schedule nobody has to phone in to check.
A distributor gets a rep-facing order app, credit limits enforced automatically instead of remembered, delivery tracking, and a route-profitability report that usually surprises the owner.
A property manager gets a tenant portal, rent tracking against M-Pesa payments, maintenance requests with an audit trail, and automatic arrears reminders that do not depend on anyone remembering to send them.
Notice what these have in common. None of them are about looking good online. All of them are about hours, errors and money.
The maths that makes the decision for you
The objection is always the same: "KES 200,000 is a lot for a system."
It is. So compare it to what you are already paying.
Take a business where one person spends three hours a day on manual admin — order entry, receipts, stock counts, chasing payments. At a KES 40,000 monthly salary, three of eight hours is roughly KES 15,000 a month in labour spent on work a system does instantly.
Add the errors. One wrong stock promise a week that costs you a KES 3,000 sale is another KES 12,000 a month. Add the payments you forget to chase. Add the customer who does not come back because you double-booked them.
Call it KES 20,000–30,000 a month, conservatively. A KES 200,000 system pays for itself in seven to ten months, then keeps paying every month after that, for years.
Now do the same sum for a KES 200,000 website. Unless it demonstrably brings in customers, it does not have a payback period. It has a cost.
This is the real argument for systems, and it is why we lead with it rather than with technology. A website is a marketing expense and should be justified as one. A system is an operational investment and should be justified with a payback calculation. If your supplier cannot do that calculation with you, they are selling you software, not solving your problem.
The compliance forcing function
There is a second reason this conversation has become urgent, and it has nothing to do with efficiency.
Since January 2024, KRA has required businesses — including many non-VAT-registered traders and freelancers — to issue invoices through eTIMS. From 1 January 2026, KRA validates income and expense entries in tax returns directly against e-invoicing data, and expenses without a compliant eTIMS invoice are not deductible.
Read that last part again from your customer's point of view. If a business cannot issue you a compliant invoice, your payment to them is not deductible. So businesses are quietly beginning to prefer suppliers who can.
Handwritten receipts were an inconvenience. They are becoming a commercial disadvantage. If you are going to digitise invoicing anyway — and you are — digitise it inside a system that also knows your stock and your customers, rather than bolting a compliance tool onto a paper process.
Doing both at once costs meaningfully less than doing them a year apart.
When you genuinely do not need a system
We turn this work down regularly, so here is the honest counter-case.
If you are under about ten transactions a day, a system is overhead. A notebook and a Till number are genuinely fine. Come back when the volume hurts.
If your process changes every month, you are still figuring out how the business works. Software encodes a process; encoding one that is about to change is expensive. Stabilise first.
If nobody will maintain it, do not build it. A system nobody updates becomes wrong within weeks, and wrong data is worse than no data — people trust it and act on it.
If you can solve it with off-the-shelf software, do. QuickBooks, a standard POS, Google Workspace, an off-the-shelf booking tool. Custom is for when your process is genuinely unusual or the off-the-shelf options do not speak M-Pesa or eTIMS properly. A good supplier will tell you this. It costs them a sale, which is exactly why it is worth hearing.
The right sequence
For most Kenyan SMEs, in order:
- Get findable. A fast, well-built website with a Google Business Profile. KES 35,000–120,000. Do this first, always — a business nobody can find has no volume problem to solve. (What that actually costs in Kenya.)
- Get paid properly. M-Pesa integration, digital receipts, eTIMS-compliant invoicing. Often bundled with the above. Here is how M-Pesa integration actually works.
- Fix the loudest operational problem. One tool, one problem. Bookings, or stock, or orders. KES 90,000–180,000.
- Connect them. Once each piece is proven and in daily use, join them so data flows and reports build themselves. KES 150,000+.
Businesses that skip to step four buy something impressive that nobody uses. Businesses that never leave step one stay busy and stay stuck.
Frequently asked questions
What is the difference between a web application and a website? A website presents information. A web application lets people log in and do work — create records, change data, run processes. Your bank's homepage is a website; your bank's online banking is a web application.
Can I start small and add to it? Yes, and you should. Build the one thing that hurts most, use it for three months, then extend. Anyone insisting you must build everything at once is selling scope, not solutions.
How long does a custom system take? A focused single-purpose tool: four to eight weeks. A multi-module system such as POS plus inventory plus reporting: three to six months. Anyone promising a full business system in two weeks is describing a spreadsheet with a login screen.
What happens if my developer disappears? The question every Kenyan business should ask on day one. Insist on three things in writing: you own the source code, you have access to the hosting and database, and there is written documentation. If a supplier hesitates on any of those, you have your answer.
Is off-the-shelf software cheaper? Usually, and often better. The case for custom is when your process is genuinely unusual, when you need deep M-Pesa or eTIMS integration that packaged tools handle badly, or when per-user licensing across a large team costs more over three years than building once.
How do I know if I can afford it? Do the maths in the section above. Count the hours currently spent on manual admin, multiply by a realistic hourly cost, add the errors. If the monthly figure is above KES 15,000, you are already paying for a system — you are just paying in labour instead of in software, and getting no asset for it.
How Werok Systems handles this
We build both, which means we have no incentive to push you toward the expensive one — and we say no to system projects fairly often, usually because the business is not ready or an off-the-shelf tool would do the job.
Website Design starts at KES 35,000. The SME Growth Pack starts at KES 120,000. Custom Business Systems — CRM, inventory, POS, bespoke workflows — start at KES 150,000. Digitize-My-Business, which combines a custom system with a website and staff training, starts at KES 220,000.
Every build is handed over with the source code, the hosting access and the documentation in your name. You should never be locked into your supplier by anything other than the fact that they are good.
Not sure which side of the line you fall on? Tell us what a normal day in your business looks like — where the time goes and what keeps going wrong — and we will tell you honestly whether you need a website, a tool, or a system. Sometimes the answer is that you need none of them yet.
Get a free quote → · Talk it through on WhatsApp
eTIMS requirements described here reflect KRA guidance current as of August 2026. Tax rules change and this is not tax advice — confirm your specific obligations with KRA or a qualified tax adviser.