Full disclosure before anything else: most of the websites we run today are built with Elementor. We chose it deliberately, we used it for years, and it was the right call at the time.

This is not a piece about a tool I regret. It is an explanation of why we are building differently now, written mostly for our own clients, who deserve to hear the reasoning from me rather than notice the change and wonder about it.

Why we used it, and why that was correct

Elementor let a small shop ship work that looked like a big shop's.

Before builders like it, a custom layout meant hand-writing every template. That is fine when you have a team of six. When you are one person carrying a client list, a tool that lets you build genuinely custom design without writing a template file for every variation is the difference between taking the work and turning it down. It is enormously capable, the ecosystem around it is deep, and Elementor themselves have put serious engineering into performance over the years. None of that is in question.

For a long stretch, the trade it asked for was clearly worth it.

What changed, and there is a date on it

On 12 March 2024, Google replaced First Input Delay with Interaction to Next Paint as a Core Web Vital.

INP measures how quickly a page responds when somebody actually interacts with it — taps a menu, opens an accordion, hits a button. To score well, at least 75% of interactions need to respond in under 200 milliseconds. And the three things Google names as the causes of a poor score are long-running JavaScript, unnecessary JavaScript, and large rendering updates.

Read that list again with a page builder in mind.

Every visual builder, by its nature, ships a runtime to the browser and produces deeply nested markup, because that is how it turns a drag-and-drop interface into a web page. That is not sloppiness. It is the cost of the abstraction, and every builder pays it — Elementor, Divi, WPBakery, all of them. The tool did not get worse. The measurement moved onto exactly the thing the abstraction costs.

A carefully built and well-tuned Elementor site can still score respectably. We tune them. But you are optimising against the grain of what the tool does, and there is a floor you cannot get under.

The thing that actually decided it

Here is the part that changed my mind, and it has nothing to do with benchmarks.

A page builder's entire promise is that somebody non-technical can lay out their own pages. Every bit of that overhead — the runtime, the nested markup, the editor scaffolding — exists to buy that one feature.

Almost none of our clients have ever opened it.

They do not want to drag boxes around a canvas. They want the price on the services page changed, so they text me the new number and it is live before lunch. For nearly all of them we function as the outside IT and marketing department, not a vendor who handed over a login and wished them luck.

So we were carrying the weight of a design editor almost nobody used. Once I put it that plainly, it stopped being a close call.

What we build now, and what actually changes for you

Static builds, delivered from the edge, with a proper content system behind them. This site is one: the pages are assembled ahead of time and served as files, with no database query and no builder runtime standing between a visitor and your phone number.

You still get a login. I want to be clear about that, because it is the thing people assume they are giving up. You get our CMS — a clean, plain interface for your words, your photos, your prices, your staff list. Change what you need and it goes live.

What changes is what that login edits. A page builder hands you control of the design, and the price of that control is a runtime shipped to every visitor's phone, forever, whether anybody uses the editor or not. A content system hands you control of the content, and costs your visitors nothing, because the design was already built and compiled before anyone arrived. You edit the things that actually change month to month. We handle the parts that should not be dragged around by hand.

In practice most of our clients will keep texting me anyway, and that is fine. The point is that the login is there when you want it, rather than being a chore you were handed at launch.

On our side, the design stops being a negotiation with a tool. Whatever we can draw, we can build.

If your site is on Elementor right now

Almost certainly it is, and here is the honest position.

Nothing is wrong with your website. Those sites are fast, they score well, and they rank — and that is not luck. We put a great deal of continuous work into keeping them there: tuning, caching, image handling, a carefully chosen stack of optimisation plugins, and regular checks that nothing has drifted since the last update. That work is part of what you pay us for and it is not stopping.

So the honest difference is not that your site is slow and a new one would be fast. It is what it costs to hold that line, and how much headroom is left above it. On Elementor we reach the number by working for it, continuously. On the new architecture the number is roughly where the site starts, because there is no runtime to optimise around in the first place.

Which is why nobody is being migrated in a panic, and why there is no separate invoice coming for a platform move. The plan is simpler than that:

When your next rebuild comes around — the one already in your roadmap — it gets built on our new infrastructure, with our own CMS behind it. A redesign you already wanted, on better foundations, at the point you were going to redesign anyway.

If that is three years out, your site stays exactly where it is and we keep it in shape. That is the entire plan, and if anybody tells you otherwise it is not coming from me.

Why I am writing this at all

I published a piece recently arguing that you should pick tools by whether they let the job get done properly, and that being loyal to a tool is a way of putting your own comfort ahead of the client's outcome.

That is easy to write about somebody else's favourite platform. This is what it looks like pointed at your own — the one you know cold, the one your whole portfolio is built in, the one where switching means learning a new way to work and rebuilding your own habits from scratch.

The evidence moved. So we moved. That is supposed to be uncomfortable, and if it never is, you are not actually choosing tools — you are just defending the last decision you made.