This site argues that most small business websites are brochures and should be built like brochures. That argument is only worth anything if it also tells you when it stops applying. Here is where it stops.

The test is one question: is the platform doing something, or is it only doing something to your invoice? Three situations where the answer is unambiguously the first.

1. Memberships and gated content

The moment some of your content is visible to some people and not others, you need accounts. Accounts need registration, password resets, roles, billing status, and a way to revoke access when someone stops paying. That is a real application with real state, and there is no version of it that is nine lines of HTML.

You could build it yourself. You should not. This is exactly the problem WordPress plus a membership plugin has been solving for fifteen years, with the edge cases already found by somebody else — the expired card, the refund, the shared login, the person who changes their email address. Buy it.

2. A real shop

"Real" is the load-bearing word, so here is what it means: stock levels that change, variants, shipping rules that depend on where the customer is, tax that depends on what they bought, refunds, and an order history someone has to look things up in.

That is inventory management with a storefront attached, and it is worth every dollar of a proper platform — WooCommerce, Shopify, whichever fits. The thing that makes it a shop is not the payment button. It is everything that happens after the payment.

3. Several people publishing, often

One person changing a phone number twice a year can edit a file. Three people who are not technical, publishing several times a week, cannot — and asking them to is how you end up as the bottleneck for your own website.

What they need is the boring part of a CMS, and it is genuinely valuable: a login, a draft state, a preview, an editor that does not require knowing what a tag is, scheduled publishing, and a revision history for when someone deletes a paragraph they should not have.

Notice what this case is actually about. It is not about the website being complicated. It is about who is allowed to change it and how often. That is an organisational fact, not a technical one, and it is the most commonly missed reason to buy a CMS.

The question is not how complex the site is. It is how many people change it, and how often.
The one that gets missed

The honourable mentions

  • Multiple languages with real translators. Managing parallel content in four languages by hand is a filing problem that gets out of sync in month two.
  • Bookings and events with availability. Anything where two people can claim the same slot needs a database and someone thinking about race conditions.
  • A large, growing catalogue of anything. Two hundred case studies that need filtering and search is a structured content problem. Twelve is a folder.
  • Regulated content with an audit trail. If you have to prove who changed what and when, you need software that records it.

Which side are you on

Buy the CMS Write five files
Who edits Several non-technical staff You, or nobody
How often Weekly or more A few times a year
Accounts Customers log in Nobody logs in
Stock Changes daily, has variants Fixed, or none
Page count Grows on its own Five, and it has been five for years
If two or more rows land on the left, stop reading this site and go buy the platform.

If you are buying, buy it well

Choosing a CMS correctly does not mean the five-year bill stops existing. It means you get something for it. Four things to insist on:

  • A short plugin list, justified item by item. Every plugin should map to a feature you can name. Nineteen is not a plugin list, it is a sediment layer.
  • Two-factor on every admin account, and an offboarding step. When someone leaves, their account goes. Write it down as a process now, because nobody remembers in three years.
  • A maintenance agreement that says what is being maintained. "Care plan" is not a specification. Updates applied within how long? Backups tested how often? Restore tested by whom?
  • An export plan, written on day one. Ask how you would get your content out and how long it would take. If nobody can answer, you are not buying a platform, you are buying a landlord.

And if you are not

If you got to the bottom of the table and everything landed on the right, then the honest answer is that you have a document, and a document does not need an application underneath it. Start with what the platform costs over five years, then the contact form, then somewhere free to put it.