“If you have come here to help me you are wasting your time. If you have come because your liberation is bound up with mine, then let us work together.” - Lilla Watson

There was a light

Figma Sites was announced at Config 2025 on May 7, 2025. In tangentially related news, the product announcement came 26 years and 2 days after WCAG 1.0 was announced.

WCAG 1.0 was developed by the W3C in response to the runaway communications revolution caused by a surge in Internet connectivity, and the widespread and rapidly accelerating adoption of these groundbreaking new protocols. As these innovations moved beyond academic, hobbyist, and niche commercial spaces into the everyday lives of millions, WCAG 1.0 emerged as the first international standard aimed at making web content more accessible to people with diverse information access needs. Today, it is primitive and low-tech guidance - but was at the time, crucially, a bright, optimistic and serious statement of intent: the web is the future of information - and this future must include us all.

This first pass can be considered the genesis of many of the ideas, philosophies and practices that ended up informing WCAG 2.2 - the latest formalisation of now long-established guide-rails designed to ensure that Internet products arrive and remain accessible in 2025. It is not only today's gold standard in it's ambition to help web products work for the widest group of mixed-access users possible, but also a type of acknowledgment that the Internet and the web have, in the blink of an eye, become the most dominant piece of techno-social infrastructure of our time.

The impact and significance of the work of accessibility standards today sits well beyond the perimeter of the philosophical or the technically practical. This is not a box-ticking exercise, nor a pat on the back for busywork. It now also informs legal compliance, ethical design, search ranking and most importantly of all: an individual's fundamental right to equal access to digital information and services. In many countries, the law considers accessible web sites and applications as part of the criteria that comprises equal participation in society. People must be able to access vital information and digital services - particularly public sector ones. And the future, in the EU at least, hammers this home for the private sector too.

This very brief scenic route through some of the accessibility history of the Internet is important context for an industry that despite 30 years of standards work, is still having serious problems considering accessibility a first-order concern and not some sort of minimum-effort, performative bolt-on.

A builder of builders

Figma Sites is the new hotness in visual website-builders, and it's designed to be useful for 'non-technical' users (although, I hate that term). It allows users to take the designs that they've already created in Figma, and through some sort of machine transformation, automatically convert them into responsive web layouts - removing the difficult and often expensive portion of the work that involves a web developer actually building it. I suppose people involved in running product strategy for Figma probably referred to this sort of thing as 'closing the loop', or something, which is often what people in companies say before they bolt on new features that are a complete departure from the thing that made them successful in the first place. To my mind, Figma's big announcement is unfortunately, not really a big announcement at all. The fact is that these sorts of visual website builders have been present in one form or another for the last 20 years. Some of the OGs of this space are Netscape Composer, Microsoft FrontPage, and (amazingly, still) Adobe's Dreamweaver - all products that developers of a particular age love the memory of and revile in practice. Despite a number of them being fairly successful at 'democratising' the process of launching websites, even the good ones take heat for being difficult to use and limited in flexibility. I don't think it's much of a reach to say that pretty much all of them are notorious in one way or another for pumping out machine-generated client-side code that does not pass many of the most basic web accessibility standards, let alone those of serious practitioners.

The code generated by the tooling that drives these products is sometimes so poor that it is practically impossible to read, and so immediately, becomes extremely difficult to reliably debug and maintain. Because's it's often not semantic, and contains either spotty, or no, aria work, screen readers have no real idea what to do with it. Styling, markup and interactivity are often schlacked into a godless, mutant code abomination that only even renders because modern browser engines are so extremely tolerant of godless, mutant markup that they sometimes do a pretty sterling job of handling it regardless. These companies have sometimes been staggeringly successful in spite of what eventually comes to light as transparent disregard for the fundamental standards of the platform they have chosen to profit from.

In it's first iteration, at least - Figma Sites unfortunately does not break away from the problem systemic in these products. None of the output generated by the Figma Sites builder today passes basic tests for semantic markup, and here's a really good video explainer of the wider issue, by Kevin Powell, which helps contextualise the rest of this article. There are also interesting BlueSky threads here and here too.

Roughly, the GDP of Namibia

Figma is a single San Francisco technology company with a $12.5 billion dollar valuation, which incidentally is slightly larger than the GDP of the country of Namibia (population ~2,963,000). Although Google's core product also seems to get worse every day, a quick search reveals that Figma appears to have 1,600 employees. The company has unseated much of the existing competition for design prototyping in the space, and today is the effective de facto tool of choice for most web designers who work on creative teams for large companies. Their original success is generally attributed to their collaboration tools. It's a cloud-based service, and they use some pretty bonkers websockets stuff to allow multiplayer, real-time design work to happen - which in the space, and at the time, was a genuine innovation. Particularly at the point that distributed collaboration was crucial when technology companies had to keep their lights on with fully remote teams.

Despite the history of the space, it is still hard for me to imagine that a company of this calibre chose to ship the product while it still contains this scale of bug in 2025. It is hard for me to imagine that in a company of Figma's size and profile, no experienced front-end engineer looked at the output and did not immediately say "ah, that is 29 nested <div> tags". It is worse to imagine the product function knowing about it and it still shipping it regardless - given that this is a website designed with the explicit charge of making other websites. It is really a bad thing that this is the beginning of the life of those websites. With another cursory search, it turns out that Figma has approximately 4 million individual users so the potential blast radius, so to speak, is enormous. It is not a bright prospect for an accessible Internet at a time when we have to be more committed than ever to doing the right thing online as a reaction to a currently unfolding, extremely severe degradation in web product and information quality. I do not believe it to be a coincidence that this degradation has been astonishingly profitable for a small number of highly motivated money types.

I don't think that being an extremely online scold that takes great pleasure in scathingly pulling apart the technical work of teams at big companies for sport has ever really been effective, even thought it sometimes feels required, to remind yourself you are not losing your mind. For that reason, I am not very interested in the condemnation of one specific company, although I think it is both fair and true that Figma are far too big and established a web organisation to have allowed this to happen - because this does not feel like the work of people who deeply understand the world wide web. But all of the best failures are usually multiple separate failures in a trenchcoat, so to speak. I know that a lot of clever people probably worked impossibly hard on this to make sure that the product could be announced at Config 2025. They are not the first company to find themselves here, and they will not be the last. One of the many side-effects of This Sort Of Thing is that it is the transparent revelation of what a product development organisation takes seriously, internally. It also demonstrates what they are willing to compromise on to make hard deadlines work. I think now is as good an opportunity as any to be introspective about whether the platform's accessibility standards are ever an acceptable compromise given the civic and social role of the Internet in 2025.

Regardless of whether or not they were looking for it in first place, companies that end up cornering an entire section of a market inherit a mesh of difficult responsibilities, many of which live in diametric opposition to their revenue-generating agenda. This is to say: what you do at any size matters. What you do at Figma's size really, really matters.

This Sort Of Thing as a Service

I feel intuitively that This Sort Of Thing is a symptom of the well-known and frankly, disaster, of an issue that the commercial priorities of very large, venture-backed Internet companies reward product practices that are the direct counter of what makes an open and accessible web. It shows the competitive space that short-term, broadly synthetic financial pressure means that being first is better than being even adequate. It is also a crystal-clear signal to developers at these businesses that huge success can happily arrive at the expense of an open, semantic, free web of information. I worry about the ramifications of that when these developers eventually go on to found their own products or development houses. In a similar vein, when companies of the size and might of Figma go ahead and do this stuff, it automatically informs the quality of the other design tooling in the space - each under their own kinds of constant pressure.

In my mind, it contributes to something worse - a propagation of a web of the future that does not prioritise diversity of need or access - by the very companies that only exist because of the rigour of the platform. Without a sophisticated way to express what a thing should be (which, is really, just coding) there's no real way to infer meaning from an arbitrary graphic that can be transposed into rich HTML, unless we are to believe that AI tooling can now do this for us. Which posits a different question. If AI tooling can now do this for us - why isn't it? This is declarative work, based on now millions of lines of agreed, factual standards training data. Is this not the perfect candidate for the work of the prediction machine?

It's worth saying that I don't know how much of this new AI/LLM stuff Figma uses at the moment, if any. I'd find it hard to believe that the answer is 'none' - but only because I have never, in my career, seen so many tech-forward companies scrambling to try to find or create problems to solve with an emerging technology. Nobody knows what to build. What they do know is that it should be AI.

However we got here, any outcome that looks like 'this lets us do what we already do poorly, faster' is a frankly crushing indictment of the imagination of the industry's leading technologists and the capability of the technology itself. It is a result of the pervasive philosophy that the most powerful companies in the world get to talk constantly about 'disruption', and 'changing the world' - without listening to the answers of the immediate follow up questions: "Who are you changing this for?" and "In what ways are you changing it for them?". We already deeply understand the ramifications of allowing corporations to be insulated from the effects of closed, radical, for-profit agendas - while they spout progressive marketing positivity at the same time.

The future of the web will continue to include companies that look and behave like Figma. And although the responsibility of agitating for an accessible, progressive, quality web is the job of us all - the institutions that enjoy the rewards of disrupting our existing paradigms must also be measured by the energy they spend honouring the difficult responsibilities that they naturally inherit. I hope that we will see Figma improve this work because this is almost certainly not a technical competence issue. But this isn't really about Figma. It is about the imagination that more than ever, the true and hard work of the future - technologically and politically, is that it must include us all.

Until tomorrow. 🕶 🖤