CMS in 2026: do we still need a publishing system?
It has never been easier to get a website up. Describe it to an AI tool and it is standing there fifteen minutes later. The question is not whether that works. The question is what happens to that website in two years. Here is why we still think a publishing system is the right call for most businesses.
Something really has changed
We are not going to pretend otherwise. Tools like Lovable, v0 and Bolt have made it possible to describe a website in plain language and have it built. The result is often good looking, fast to produce and perfectly usable.
For some purposes that is exactly the right choice. A campaign site meant to live for six weeks. A landing page for a single event. A prototype you need to show an idea internally. In those cases we would also say a full publishing system is overkill.
So the simplest part of the market has moved, and that is fine. It is the other part we want to talk about.
The problem is not that it works
Imagine a capable colleague builds your website in an afternoon. It looks good. Everyone is happy. Case closed.
Then eighteen months pass.
The colleague has changed jobs, or simply moved on to other things. The tool that built the site has changed, introduced new pricing, or altered the way things work. Nobody else in the company has opened the project since it was made.
And now you need to change an address in the footer.
That should take thirty seconds. Instead somebody has to work out where the project lives, get access, understand how it was built, make the change in the code, and publish it again without breaking anything else. Suddenly a trivial edit has become a small task that needs planning.
Two dependencies that are easy to miss
Websites built this way tend to get stuck on two things.
- The person. The knowledge of how the site was made sits with whoever made it. Often it exists nowhere else.
- The tool. The site is built in one particular service, with one particular way of doing things. If that service changes, or disappears, you depend on somebody being able to move the project out.
This is not an argument against AI. The same thing happened to websites built by a single freelancer in a homemade setup fifteen years ago. The difference is that it is now much faster to end up there, and that it happens to sites which look professional from day one.
Technical debt does not come from bad work. It comes from work nobody maintains.
What a publishing system actually gives you
A CMS is not primarily technology. It is a division of labour.
People without a technical background can change the content
The marketing manager changes the address. The communications adviser publishes a news item. Nobody has to open code, and nobody has to ask a developer for permission first.
Content is separated from appearance
Text, images and prices live in a database, not inside code files. If the site gets a new design in three years, the content comes along. In the long run this may be the most important difference of all.
Several people can work at once, in different roles
Some can publish, others can only suggest. There is version history, and you can see who changed what. For an organisation with more than one person, this is not a luxury.
It is predictable to hand over
A website in a well known publishing system can be taken over by any agency that knows that system. That is real insurance. We see the consequence of the opposite every time we take over a project where nobody knows how things fit together any more.
Security has become more important, not less
This is the part we care most about right now.
Websites are scanned automatically, at scale, by tools looking for known weaknesses. It is nothing personal, and no site is too small to be found. AI has made both the looking and the attacking cheaper to carry out, and it shows.
A maintained publishing system has machinery around it. Security updates arrive, warnings go out, and there is something to update to. A website that was generated once and never touched again has none of that. It just sits there.
That is why we think choosing a system in 2026 has to be about more than how quickly you can get started. It has to be just as much about who looks after the site afterwards.
What we use, and why
We use Craft CMS on most of our projects, and have done for many years. We are an official Craft CMS partner.
The reasons are simple. It is tidy for the people editing the content, flexible for those of us building, and it is kept well maintained. Craft 5 has support through 2031, and Craft 6 is on its way. So we know what we are getting for the next few years, and that is part of the point.
At the same time, Craft is not the only answer. Other publishing systems solve the job well, and the right choice depends on what you are doing, who will work in it, and what you already have. The important thing is not which system you land on. The important thing is that there is a system, that it is maintained, and that somebody other than one person can take it over.
So, do you need a CMS?
If the site is meant to live for a few months, perhaps not. The new tools are often the better choice there, and we are happy to say so out loud.
If it is meant to live for years, be edited by several people, and still work long after whoever built it has moved on? Then we still think the answer is yes.
Get in touch if you want an honest assessment of what suits your situation.
