If you have started asking what custom software costs, something has usually already gone wrong. A spreadsheet has become load-bearing. Jobs are being booked in a WhatsApp group. Someone spends their Friday copying numbers between two systems that were never meant to talk to each other.
The honest answer is that it depends — but not in the evasive way that phrase usually means. It depends on a small number of specific things, and once you know what they are, you can tell a realistic quote from a bad one.
The short answer
Focused first versions start around R15,000. Most business systems land between R50,000 and R150,000.
That is a wide range because it spans very different things. Here is what sits where.
Around R15,000 — a focused first version
One clearly defined job, done properly, for a small number of users. A quote-request system that captures the right details and emails them to the right person. A booking form that checks availability and takes a deposit. An internal tool that replaces one specific spreadsheet.
This is not a cheap version of a big system. It is a small system that is genuinely finished. The point is to solve your most painful thing quickly, get it into daily use, and learn what actually matters before committing more.
R50,000 to R150,000 — a business system
This is where most operational software lands. Several types of user with different permissions — customers, field staff, office admin, an owner who wants to see everything. Real workflows with states and approvals. Payments, notifications, reporting. Often a mobile app alongside a web dashboard.
Concretely: a job management system where customers book, technicians receive and complete work on their phones, and the office sees everything live. A customer portal that shows accounts, invoices and documents. A booking platform that handles scheduling, deposits and reminders without anyone confirming appointments by hand.
Above R150,000 — platforms
Multi-sided marketplaces, systems serving tens of thousands of users, anything with heavy compliance, or software that has to integrate deeply with several existing systems at once. Real, but rarer, and almost never where a business should start.
What actually drives the price
Six things move a quote more than anything else. When you are comparing proposals, these are what to interrogate.
1. How many kinds of user
This is the single biggest driver and the one most people underestimate. Software used by one admin is straightforward. Software used by customers, field staff, managers and accounts is four applications wearing a trench coat — each role needs its own screens, its own permissions, and its own testing.
If you are trying to reduce cost, reducing roles in the first phase does more than cutting features.
2. What it has to talk to
Connecting to your accounting package, a payment gateway or a messaging service adds work. It also usually removes your most expensive manual step, so it frequently pays for itself faster than anything else in the build. Ask whether the systems you use expose an API — if they don't, that changes the plan.
3. Whether it has to work offline
If your technicians work in basements, rural areas or anywhere signal is unreliable, the software has to keep functioning and sync later. That is real engineering — conflict handling, queued actions, local storage — and it costs accordingly. Worth every cent when it applies, pure waste when it doesn't.
4. Security and compliance
Handling personal information properly is not optional, and in South Africa POPIA applies to you whether or not you have thought about it. The level of rigour scales with what you store. Payment data and health information cost more to handle correctly than a list of job addresses.
5. App stores
Publishing to the App Store and Play Store adds developer accounts, review cycles, store assets and an ongoing obligation to keep up with OS updates. If your users are staff rather than the public, ask whether you need store presence at all — often a web app on their phone does the job for less.
6. Hosting and maintenance
Software that runs a business needs somewhere reliable to live and someone to keep it running. This is ongoing, not once-off, and a quote that omits it is not cheaper — it is incomplete. Expect a modest monthly cost for hosting, and a sensible arrangement for updates and support.
Why not just buy a subscription?
Often you should. If an off-the-shelf tool fits your process, buy it — that is genuinely cheaper and faster, and anyone who tells you otherwise is selling something.
The question worth asking is what you are already paying to work around the tool that almost fits. Staff hours spent re-typing between systems. Exports and imports. The customer question nobody can answer without phoning someone. A process only one employee fully understands, which is an operational risk rather than a software problem.
Add up those hours at what those people cost per hour, over a year. That number is what you are already spending. Compare it to a build cost, and the decision usually stops being about price.
Build it in phases
Nothing has to be built at once, and it usually shouldn't be.
The approach that works is a first version solving the single most painful part of the process, in daily use and earning its keep before the next phase starts. Three things come from that. Cost spreads over time instead of landing as one number. Real usage tells you what to build next, which is almost never what you predicted. And if the first phase does not deliver, you find out for R15,000 rather than R150,000.
Discovery versus development
Any developer quoting a precise number for software they have not scoped is guessing, and you will pay for the guess later through change requests.
A proper discovery conversation maps how your business actually runs today, what the manual process is costing you in hours, who touches it, and what the smallest useful first version would be. That is what turns "build me an app" into a number that means something. You should come out of that conversation with a recommendation you could act on — including, sometimes, the recommendation not to build anything.
How to read a quote
Three questions worth asking anyone who quotes you:
- What is not included? Hosting, maintenance, app store accounts and post-launch changes are the usual omissions.
- Who owns the code? The answer should be you, along with the database and the accounts it runs on. If it isn't, you are renting, and switching later will cost more than the build.
- What happens when I want a change? You will want changes. Find out how that works before signing, not after.
So what should you budget?
If you have one painful process and want to test whether software fixes it, budget around R15,000 and scope tightly.
If you are replacing how a real operation runs — multiple people, real workflows, customers touching it — budget in the R50,000 to R150,000 range and expect it to be built in phases.
And if you genuinely do not know which of those you are, that is the normal starting point. Most people know what is painful, not what to build.