Ask four suppliers to quote the same MVP and you will get R35,000, R250,000, R600,000 and "let's schedule a discovery workshop". Every one of those is a real quote for a real product. The number on its own tells you nothing, because the four suppliers are not selling the same thing.
Here is what each band actually buys.
The four price bands
R35,000 to R120,000: offshore commodity build
Usually an overseas team, often quoted per screen or per feature. You send a specification, you get code back.
This is a legitimate purchase when your spec is genuinely settled and the build is mechanical: an internal tool, a straightforward marketplace, something you have already validated by other means. You are buying execution, and execution is a real thing to buy.
It goes wrong when you are still figuring out what to build. Nobody in that engagement is paid to tell you the third feature is unnecessary or that your core assumption is untested. You asked for eleven screens, you get eleven screens. If the product does not work, the build was still delivered "correctly".
The failure mode is not bad code. It is a working product nobody wants.
R150,000 to R400,000: senior-led fixed scope
One senior product person owns scope, decisions and delivery, with an engineering team behind them. Six to ten weeks, fixed price, agreed before anyone opens an editor.
You are paying for the arguing. Someone whose job is to say "that feature does not earn its place" before it costs you R40,000. In practice, most of the value in this band is scope you did not build.
This is where SILC works, so treat that as a disclosure rather than an impartial view.
R400,000 to R1.5m: agency or pod
A team of three to five: product, design, engineering, sometimes a delivery manager. Cape Town has several very good ones in and above this band.
The team is real and often excellent. What you are also buying is the overhead that makes a team of that size possible: account management, sales, the bench between projects. For a well-funded product with genuine breadth, that is money well spent. For a first MVP, you are frequently paying for capacity you do not need yet.
R1.5m and up: enterprise consultancy
Accenture and equivalents. Procurement, statements of work, a named engagement lead who manages rather than builds.
Correct for a bank. Almost never correct for a founder.
Where the money actually goes
A R250,000 MVP is not 250 hours of typing. Roughly:
- Scoping and discovery, 10 to 15%. Deciding what not to build. The highest-return line item on the list, and the first one people try to cut.
- Design, 15 to 20%. Enough to be usable and credible. Not a design system.
- Engineering, 50 to 60%. The actual build.
- Launch and instrumentation, 10%. Getting it live and knowing whether it worked. Skipped constantly, which is how founders end up with a live product and no idea which part failed.
The interesting number is the first one. Cutting discovery to save R30,000 is what produces the R250,000 build of the wrong thing.
Why the cheap quote is usually the expensive one
Two builds, both real patterns:
Build A. R80,000, eight weeks, offshore. Delivered to spec. Six weeks after launch it is clear users will not do the one thing the whole model depends on. Total cost: R80,000 and fourteen weeks, and you still do not know what to build.
Build B. R60,000 on a three-week proof that tests the same assumption before the build. The proof says no. You pivot to the adjacent thing users actually asked for, and build that for R220,000. Total: R280,000 and seventeen weeks, with a product in market that works.
Build B costs three and a half times more and is enormously cheaper, because Build A's real cost was never the R80,000. It was the fourteen weeks and the seed money spent afterwards on a wrong answer.
This is not an argument for expensive builds. It is an argument for finding out early.
Reading a quote properly
Five questions worth more than the price:
- Is it fixed or a starting point? "From R150,000" is not a quote. Ask what the total was on the last three projects that started at that number.
- Who is accountable when it goes wrong? A name. If the answer is a team or a process, nobody is.
- What is explicitly not included? A supplier who cannot produce a "not now" list has not scoped the work; they have estimated a wishlist.
- Who writes the code, and have they worked together? A newly assembled team learns on your budget.
- What happens at week four if it is going badly? The honest answer involves stopping. Suppliers who cannot describe that scenario have not had it.
A straight answer
For a genuine first MVP in South Africa, scoped properly: R150,000 to R400,000, six to ten weeks. Below R150,000, either the scope is very narrow, which is fine if deliberate, or somebody is going to discover the real number later. Above R400,000, you are buying team capacity, which is worth it once the product is proven and rarely before.
Two things bring that down honestly: cutting scope, and killing bad ideas early. Everything else is moving cost around.
If you want your own number, book a call. Fixed quote within 48 hours after a scoping conversation, and I will tell you if the range does not fit what you are trying to do.