Who edits the site, and what that decides

The question "which system should the site run on" is not really about software. It is about permission: who may change a page without calling anyone, and what they may touch. Three answers exist, and the licence, the editor and the templates all fall out of the one you picked, often without anyone choosing it out loud. The answer flips when the site changes twice a year rather than twice a week: then the cheapest correct arrangement is usually the code you already have, plus one person who can be reached.

What are the three models?

Nobody edits it, so the studio does. One person edits text and images inside a fixed page. Anyone edits the structure, assembling pages out of blocks. They are not three products but three amounts of trust, and each needs different preparation before it is safe.

Model	What it assumes	What it costs over a year
The studio edits	The content is stable	Every change waits in a queue
One editor, fixed pages	Someone owns the words	Defining the fields, training that person
Anyone edits structure	A team ships campaigns	Coherence, and a design that drifts

What does it cost to have nobody edit the site?

It is cheap and honest: five pages, no login, no editor to maintain. The cost arrives later, as latency. Each change becomes a request, the request joins a queue shared with other clients, and a number changed on Monday goes live on Thursday.

It stops being fine when the site carries something that changes without warning: prices, hours, stock, a menu. Then every correction is a small project, and people stop asking. A site one season out of date is worse than a small one.

What does one editor with fixed pages cost?

The middle model, and the one most businesses want: the shape of each page is decided once, and the words and pictures inside it are fields. The editor changes a headline, a photograph, a price, and cannot delete the navigation by accident.

The real cost is not the software, it is the ownership. Somebody has to be named. Not "the team", not "whoever is at the desk": a person who knows the site is theirs, who notices the offer ended two months ago. Where that person does not exist, the fields stay as they were on handover, and the site becomes a monument.

The second cost is field discipline: a form with a subtitle field eventually receives a promotion, because the promotion had nowhere else to go. The form holds only what the page can render.

What does anyone-editing-structure cost?

Freedom, and the bill for it. Blocks let a marketing team build a landing page on a Friday without a developer, a real advantage where campaigns ship weekly. The bill is coherence: three people assemble three pages with three spacings, and the site reads as a collection of templates rather than a business.

Blocks also invite accumulation: a section added for one promotion and never removed, because removing it is nobody's job. Two years in, the home page is eleven sections long.

This fits a content-heavy site with a team behind it and someone accountable for the design. It is wrong for a five-page site whose phone number is the most important element on it, where freedom buys nothing.

What has to happen before a price goes live?

The person who knows the price is rarely the person who types it, and that gap is where a wrong number reaches a customer. One step closes most of it: a preview that renders the page as a visitor sees it, opened by someone other than the editor.

That step needs three habits. The approval is a name, not a message in a group chat, because "I thought you had checked" has no author. Every publish is recorded, so "when did this change" has an answer. And the preview must be the same page, because one that renders differently teaches everyone to skip it.

What does a client need before editing safely?

Three things, and no editor works without them. A preview, so a change is seen before it is public. A way back, so a mistake lasts a minute rather than a week. Someone to call, with a name and a promised speed.

Then half a page of writing: what may change without asking, what needs review, what needs a developer. Without it the editor either asks about everything, which defeats the purpose, or changes the navigation on a Friday afternoon, which is the incident.

Is keeping the site in code ever the right answer?

Yes, more often than the industry admits. A CMS is a second system: another login to protect, another set of updates, another way to stop working while nobody was looking. If the site changes twice a year, that is a cost paid for a capability nobody uses.

What makes code acceptable is not the technology, it is the access. You need a way to request a change measured in days, and a repository another developer could pick up. If both hold, a small site in code is fast and cheap to run. If neither does — if the only person who can change a comma has stopped answering — the freedom of a CMS is worth its maintenance.

In what order do you decide this?

1. Count last year's changes, and who asked for each one.
2. Name the person who will edit. If no name comes up, the answer is the first model.
3. List what may change without asking: a price, a phone number, the hours, a photograph.
4. Pick the smallest model that covers that list, not the most capable.
5. Build the preview and the way back before the keys are handed over.
6. Write the half-page rule, and who to call for what.
7. Review after three months, when the real pattern is visible instead of the imagined one.

Where this goes wrong

• An editor installed to avoid phone calls, then nobody trained on it.
• Every field is free text. A headline field receives a paragraph, and the layout breaks invisibly.
• The preview is a different page. It drifts, and nobody uses a preview they distrust.
• The handover has no way back. One bad edit stays live for a week, and the fix becomes a project.

What this does not cover

It names no products and quotes no prices: both rot and neither decides the question. It does not cover the design of the editing screen, nor the internal politics of who owns the website in a company where marketing and management both believe they do. It assumes a defined set of pages; a shop with a blog and a catalogue has three content models.

We build sites where the editing model is chosen before the first page, so that nobody discovers it after the invoice.