How much does it cost to build an MVP in Kenya?

By Bonaventure Ogeto, Founder and CEO, Mctaba Labs
Published

The short answer

An MVP in Kenya costs what its scope costs. Three things set the price: how many jobs the first version does, how many outside systems it connects to, such as M-Pesa, and who builds it. Set your budget first, write the scope on one page, send it to three builders and pay in stages. No single price exists.

You ask three developers what your MVP will cost. One sends a figure by WhatsApp within the hour. One sends a figure many times larger with a long PDF. The third says "let us have a call first". You are left wondering which of them is lying.

Probably none. They were asked to price a word, and each priced a different product. The honest answer to "how much does an MVP cost in Kenya?" is that there is no single price, and anyone who gives you one without seeing a written scope is guessing. What you can do is control the three things that set the price, which are how much the first version does, how many outside systems it connects to and who builds it, and then make every builder price the same written page.

This guide does not quote market prices, because any figure here would be wrong for your product. It gives you the method instead.

What actually sets the price of an MVP?

Three things, in this order of weight.

1. How much the first version does. Every flow a user can complete, every kind of user and every business rule is work to build and work to test. A product where a customer places an order is one flow. Add a vendor who accepts it, a rider who delivers it and an admin who settles everyone, and you have four products that must agree with each other.

2. How many outside systems it talks to. Payments, SMS, WhatsApp, maps, an accounting package, identity checks. Each connection has to be built, tested against the other company's rules, and kept working when that company changes something.

3. Who builds it. A student, a freelancer, a small team and an established firm will price the same sheet differently. You are paying for different amounts of experience, testing, backup and care after launch. Cheaper is not worse and dearer is not safer, but they are not the same purchase.

Two smaller drivers sit under those: which platforms you ask for, and how much custom design you want. The table shows all five and the cheapest sensible choice for a first version.

Driver What pushes it up The MVP choice
Scope More flows, more kinds of user, more rules and exceptions One kind of user, one flow, start to finish
Platforms Separate Android and iPhone apps, app store accounts and reviews A website that works well on a phone
Payments Several payment methods, refunds, payouts to third parties, wallets One route for collecting M-Pesa, with no payouts
Other connections Each extra outside system One at most. Send the rest by hand
Design Custom illustrations, animation, many rounds of changes A clean standard layout, with care on the payment step
Who builds Larger teams, more process, more after-care The smallest team that can show you similar live work

How do I cut the price without breaking the product?

Cut in the order of the table, because the top rows save the most.

Cut scope first. Write the one sentence your MVP must make true: who the customer is, the one job they do, and how they pay. Then move everything that sentence does not need to a list called "later". Admin dashboards, reports, ratings, referral codes and a second kind of user almost always belong there. If you can do something by hand for the first twenty customers, such as approving sign-ups or sending a weekly statement, do it by hand.

Choose one platform. A website that opens from a WhatsApp link needs no download and no store approval. Store apps add their own costs and steps. Apple charges a yearly membership fee and asks an organisation for extra identity checks, both listed on its enrolment page. Google asks new personal developer accounts to run a closed test with real testers before an app can go public, as set out in its Play Console help. Neither is a reason never to build an app. Both are reasons not to begin with one.

Keep payments to one route. For most MVPs, the cheapest build is a payment provider that already supports M-Pesa. Paystack, for example, lists M-Pesa among its channels for businesses in Kenya in its developer documentation. The provider takes a fee on each payment, and in return the builder writes less code. A direct connection to Safaricom's Daraja platform is more work up front. Ask your builder to price whichever they have done before, and leave payouts, refunds inside the product and stored balances for a later version.

Spend design effort in one place. The step where a stranger sends you money must look and feel safe. The rest can be plain.

What you must not cut: testing of the payment flow with real money, ownership of the code and accounts, and a short period of fixes after launch. Savings there come back as larger bills.

How do I get quotes I can compare?

Send every builder the same page. Here is a scope sheet you can copy and fill in.

MVP scope sheet

The one sentence: [Customer] can [do one job] and [get one result], and pays by [method].

Who uses it: [one kind of user, plus you as admin]

What they can do, in order:

  1. [for example: sign up with a phone number]
  2. [choose a service and a time]
  3. [pay a deposit by M-Pesa]
  4. [receive a confirmation by SMS or WhatsApp]

What the admin can do: [see the list of bookings and payments]

Platform: a website that works on a phone

Payments: [one provider, collection only]

Other connections: [SMS for confirmations. Nothing else]

Not in this version: [the "later" list, written out]

Please quote, as separate lines: design, build, payment connection, testing, set-up of hosting and domain in my name, fixes for the first [number] weeks after launch, and monthly running costs you expect.

Please also tell me: what you have assumed, what is not included, the date of the first working demo, and two similar products of yours that are live.

Three rules make the answers useful.

  1. Send the identical sheet to at least three builders. Do not improve it between conversations.
  2. Refuse a single total. A quote with no lines cannot be compared or negotiated.
  3. Read the "not included" answers first. That is where the real price hides.

If a builder replies with good questions about your sheet, that is a mark in their favour. The one who sends a figure without reading it has told you how the project will go.

How should I pay for the build?

In stages, each tied to something you can open on your own phone and test. Never pay against time passed, and never pay everything up front.

Here is an illustration. The shares are round, made-up numbers to show the shape, and yours will differ.

Milestone What you can check yourself Share of the build fee (illustration)
Agreement signed, accounts opened in your name You can log in to hosting, domain and the payment provider 20%
First working demo You can complete the main flow on your phone with test data 30%
Payments live You send a small real amount by M-Pesa and see it arrive and get recorded 30%
Launch, plus an agreed fix period You have run the full flow with five real people you invited, within an agreed number of weeks, and listed bugs are fixed 20%

The principle is that money and risk move together. A deposit to begin is fair, because the builder is setting time aside. A final share held until after launch is fair too, because it keeps attention on your project during the weeks when real users find the problems.

Put the milestones, the dates and what "done" means for each into a short written agreement, along with a line saying the code and accounts are yours once each stage is paid.

How do I work out a budget before I have any quotes?

Work backwards from what you can afford, not forwards from a wish list. Fill this in with your own figures. The example column uses a deliberately round, hypothetical total of KES 100,000 so the arithmetic is easy to follow. It is not a price for anything.

Line Your figure Example split of a hypothetical KES 100,000
A. Money you can spend on this test and lose without harm KES ____ 100,000
B. Held back for changes after real customers use it KES ____ 25,000
C. Running costs for the first six months (hosting, domain, messages, fees) KES ____ 15,000
D. What is left for the build (A minus B minus C) KES ____ 60,000

Line D is the number you take to builders. If every quote for your scope sheet comes in above it, do not stretch the budget. Shrink the sheet, move more to "by hand", or begin with a no-code or manual version and come back when customers are paying.

Line B is the one founders skip, and it matters most. The first weeks of real use always show you something you got wrong. If the whole budget went into version one, you will know exactly what to fix and have nothing left to fix it with.

For the costs that continue after launch, and for products larger than an MVP, the fuller guide to what an app costs to build in Kenya goes through hosting, message charges and upkeep line by line.

Which quote should I choose?

Not the lowest and not the most impressive. Choose the builder who priced every line on your sheet, told you plainly what was left out, showed you a similar product that is live, and offered a working demo within the first weeks.

Then judge them at the first milestone. If the demo arrives when promised and does what the sheet says, carry on. If it does not, you have spent one stage payment, you hold the accounts and the code so far, and you can stop. That is what the whole method is for: a smaller, known amount at risk at every step, so that the cost of your MVP is a figure you chose and not one that happened to you.

Key takeaways

  • No honest builder can price "an MVP". They can price a written list of what it does.
  • The number of flows and kinds of user moves the price more than anything else.
  • Start with a website that works on a phone, one payment route and one kind of user.
  • Send the same scope sheet to three builders and compare line by line, not total against total.
  • Pay against milestones you can test yourself, never against time passed.
  • Hold back part of your budget for the changes real customers will ask for.

Frequently asked questions

Why do quotes for the same MVP differ so much?
Because each builder is pricing a different product in their head. One assumes a web page with one payment button, another assumes Android and iPhone apps with an admin system. Give all of them the same written scope sheet and the gap narrows to differences you can ask about.
Is a fixed price or an hourly rate better for an MVP?
A fixed price per milestone suits an MVP, because the scope is small and written down. Hourly billing suits work that is still being discovered. If a builder insists on hourly, ask for a cap per milestone and a demo every week so you can stop when the money stops making sense.
Should I give a developer equity instead of paying?
Only if you want that person as a long-term partner and would choose them for the role anyway. Equity given to get a cheaper build is expensive later, and a builder who is not paid tends to put paying work first. For a one-off build, pay cash for a smaller scope.
Do I own the code if I paid for it?
Do not rely on it. Ownership of commissioned work depends on the law and on what you signed, so write it down. Put it in the written agreement that the code, the designs and every account are yours once each milestone is paid, and make sure the hosting, domain and payment accounts are opened in your name from the first day.
Can I build the MVP for free myself?
Sometimes. A manual version, with orders taken on your phone and tracked in a simple sheet, costs only your time, and no-code and AI tools can produce a first version for the price of a subscription. You pay in hours and in limits you may meet later, which is a fair trade while you are still testing the idea.

Sources

  1. The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses, Eric Ries (book, 2011). Checked on 19 November 2026.
  2. Payment Channels, Paystack Developer Documentation. Checked on 19 November 2026.
  3. Enrollment, Apple Developer Program. Checked on 19 November 2026.
  4. App testing requirements for new personal developer accounts, Google Play Console Help. Checked on 19 November 2026.