Ask three agencies to quote on the same brief and you will get three numbers that have almost nothing to do with each other. That is not corruption — it is because "a website" describes an outcome, not a unit of work. Here is what actually moves the number.
What you are really paying for
A website is four separate purchases stacked together.
- Strategy and content. Deciding who the site is for, what it must say, and what a visitor should do next. This is the part most quotes quietly skip, and the part that determines whether the site earns anything.
- Design. A bespoke visual system versus a purchased template someone adjusts. Both are legitimate; only one is cheap.
- Engineering. A handful of pages, or a content model, search, filtering, integrations with your accounting or CRM, multi-language, and a blog your team will actually update.
- Ongoing care. Hosting, backups, security updates, monitoring. Ignoring this is how sites get hacked two years later.
Realistic ranges
For a South African small business, the honest bands look roughly like this.
- R5 000 – R15 000: template, one page or a few. Fine for a business that needs a credible presence and nothing more. Expect to write the content yourself, and expect the site to be slow to change.
- R15 000 – R45 000: a proper small business site. Custom design, five to fifteen pages, a blog or portfolio the team can update, structured data, analytics, correct performance and accessibility. This is the band where most real businesses should land.
- R45 000 – R150 000: sites with actual software in them. Quote calculators, customer portals, booking logic, complex integrations. You are no longer buying pages; you are buying an application.
- Above R150 000: platform work. Multi-sided systems, e-commerce at scale, custom administration. At this point the website is a product, and should be scoped and staffed like one.
Anything under about R5 000 is either a template install or it is being subsidised by something you will pay for later — in hosting lock-in, in a site you cannot edit, or in maintenance you cannot get.
The questions that reveal a good quote
Price tells you almost nothing on its own. Ask instead:
- What is explicitly excluded? A quote without exclusions is not a quote, it is an opening position.
- Who owns the code, the content and the accounts? If the answer is not "you", walk away. You should own the domain, the hosting and the repository from day one.
- Can I edit the content myself? If every text change is a ticket, the site will quietly go stale.
- What happens after launch? Ask for a written handover: documentation, credentials and a support arrangement.
- How will we know it worked? If there is no measurement plan, the project has no definition of success.
The cheapest site is often the expensive one
A R4 000 site that loads in nine seconds on mobile, has no analytics, cannot be edited, and brings in no enquiries has a real cost: the traffic it did not convert, every month, for years. The comparison is not between two prices. It is between admitting the site was a waste and pretending it is working.
Decide what the site is supposed to produce. Then buy the smallest thing that credibly produces it.