The downfall of the headless

Some things move like a pendulum. Take the way we build websites. For 30 years we’re moving back and forth between website-builders, and hand-coding.

The trend over the last couple years is to go ‘headless’. Meaning to hand-code the ‘front-end’ of a website.

The idea is that by doing this, you prevent yourself from running into the limitations of a website builder. And more importantly, you don’t limit yourself to being only able to use the systems inside the website builder. Instead, you can ‘mix and match’ your own mix of systems, and yourself combine them into an awesome user experience.

This way you take control over your data, your integrations, your user experience. On top of that, you can optimise your website to be super-fast, by cutting out everything you don’t need.

Headless has been all the rage. And as a business case and an idea, it is very appealing.

It is also very abstract.

And abstract ideas tend to be forgotten. Or pushed to the background.

Techies still have a clear picture in their head. But I fear, that more and more, non-technical people are getting frustrated with the downsides of headless.

Headless can be quite complex to build and maintain. For changes, you’re very quickly reliant on a developer or a development agency, with their own hourly invoicing. And while there are many providers who make deployment and ‘edge computing’ easy and accessible, the cloud hosting bills are also stacking up.

I suspect there will be an upward trend of website builders. A new generation, that takes more of the hassle out of it than before. Tools like Squarespace, Framer and Webflow, for example.

The beauty of these tools is that they trigger the imagination. They have a very direct feedback loop. Change the font or image, and you instantly ‘feel’ what that new website would be like to use. AI and automatic responsive design takes much of the hassle out of it. And, if architected right, I think that ‘pluggability’ does not have to be sacrificed. There is an opportunity for a website builder that makes it very easy to integrate with 3rd party services. To loop in data from a REST or GraphQL API. And that makes it easy to create simple custom (JavaScript) apps.

Headless. Headless, the dream of software engineers.

Headless is

A lot of things move like a pendulum, back and forth.

So do websites, and the way we build websites.

Over the last 30 years we’ve pendled back and forth between various website-builders, and hand-coding. Each improving with each iteration.

Over the last 30 decades we’ve had various iterations of ‘website builders’. Each with

Headless wave is coming down

Idea of headless is an abstract one. Hard to stay excited about an abstract utopia when Vercel billing, countless bugfixes and developer/agency lock in are your day to day

Some things move like a pendulum

WYSIWYG website builders have come and reinvented themselves in various shapes and forms

Each time with limitations and at the cost of good code

Given rise to a quite persistent trend of headless

Which in turn fueled trends of squeezing performance out of websites, with lighthouse pagespeed and Vercel edge networks