The ceiling of no-code

There is a very specific feeling I have learned to respect. It is the moment a project's requirements list has one honest line on it that the platform cannot do, and I catch myself quietly redesigning the requirement instead of the website. The first time, it feels like pragmatism. The tenth time, it is a message.

Let me say the kind part first, because this is not a breakup letter written in anger. Webflow gave me years of the best trade in the business. Interactions without a slider plugin. A CMS without a metabox jungle. Hosting without update-day roulette. It replaced my entire WordPress plugin stack with features that simply came with the platform, and I have recommended it to anyone who would listen.

But ceilings do not announce themselves. You find them with your head.

Where the ceiling is

Five bruises, in the order I collected them.

• CMS scale. Collections have item caps, field caps, and collection caps, and a real product catalog laughs at all three. When the data outgrows the container, you start flattening rich information into fewer fields, and every flattening throws detail away.
• Logic. Filtering, search, calculators, conditional anything: the honest answer keeps turning out to be custom JavaScript pasted into an embed block. Which means I am writing code anyway, just in the least comfortable place imaginable, with no version control and no good way to test it.
• Data ownership. The content lives in their database, on their servers, under their terms. An export is a pile of CSV files and static markup, not a working system. That is not ownership. That is visitation rights.
• The cost shape. Per site, per seat, per plan tier. The better the business does, the higher the rent, and the rent buys the same product it bought last year.
• Performance. You can tune everything the platform lets you tune, and not one thing more. The heaviest parts of the page are not mine to remove, and past a point the score simply stops moving.

None of these are scandals. They are the natural shape of any platform: it optimizes for the middle of the market, and the middle is exactly where my work is trying not to be anymore.

The same shape as last time

Here is what actually unsettles me. I have taken this walk before, and I recognize the scenery. WordPress was the right choice for years, until the plugin stack stopped being the toolbox and became the liability I spent my weeks managing. The pattern was identical: adopt the platform for what it hands you, grow until you find the edges, then spend more and more effort routing around the tool instead of working through it.

My measurement for this is what I call the workaround ratio. What fraction of the week goes to doing the work, and what fraction goes to convincing the platform to allow the work? On WordPress the ratio crossed the line when plugin conflicts became my main job. On Webflow it is crossing the line now, in embed blocks and flattened collections and support articles that end with "this is not currently possible".

When the ratio tips, no amount of affection saves the tool. I loved Webflow. Present tense, honestly: I love Webflow. And I am leaving anyway.

What the framework says

Years ago I built a decision framework for client projects: template, wireframe on a builder, component library, or fully custom code, chosen by goals and budget rather than by fashion. It has settled every scoping argument since. So it is only fair to run myself through it.

The goals changed. The work I want to do next is data-heavy, logic-heavy, and performance-obsessed, and against those goals the framework does not mumble. It points at the last option on the list, the one I have been circling for years: write the code yourself and own the whole machine.

So that is the destination. Real frameworks, real version control, content that lives in files and databases I hold. Not because no-code failed. Because it succeeded all the way to the edge of its map, and my next projects live past the edge. Webflow was the right chapter, exactly like WordPress was the right chapter. The skill was never loyalty to a tool. It is noticing the moment when the reasons you came in have quietly become the reasons to go.