The best interface is an invisible one

From 2014 to 2019 my company was called Invisible Interfaces. The name was the brief: an interface works best when nobody notices it.

Geert Valère Allegaert · October 2026

The most profound technologies are those that disappear.

Mark Weiser, 1991

It was a registered company in Belgium, and I ran it for five years. The name was not a joke about a studio that delivers nothing you can see. It was a description of the job. For an interface designer, and for a graphic designer, the work is to get out of the way. People come to a page to book a room, pay a bill or read an article. Every second they spend noticing the design is a second taken from what they came to do.

So the brief was short. No fancy stuff. Be simple. Stick to standards. Design for quietness. None of it is original, and that is rather the point. Typography said it first: in 1930 Beatrice Warde asked for printing to be invisible, a story I told in The quiet is the design. For interfaces the idea has its own history, and a weak spot that is worth knowing about.

Where the idea comes from

In 1991 Mark Weiser, at Xerox PARC, wrote The Computer for the 21st Century for Scientific American. He compared computers with writing. Writing is everywhere — street signs, shop signs, sweet wrappers — and nobody has to concentrate to read it. Only when a technology disappears like that, he argued, are we free to use it without thinking and to look past it at what we actually want. The personal computer was the opposite: even the best notebook still pulled all attention towards one box.

Don Norman took the argument to product design in The Invisible Computer (1998): the PC had become too complex, and the answer was devices so easy to use that they would stop being noticed. Golden Krishna pushed it furthest in The Best Interface Is No Interface (2015), written against screens and apps that compete for your eyes. Even he has a chapter on exceptions, and admits the book was written with the form fields and menus it rails against.

Three people, three decades, one complaint: the computer asks for attention it has not earned. Weiser’s is a vision, not a finding, and the other two are arguments. That does not make them wrong. It does mean the idea rests on observation and taste rather than on a measurement.

On the web, invisible means familiar

A website cannot make its screen disappear. What it can make disappear is the effort. In 2000 Jakob Nielsen wrote End of Web Design, with the line that became known as Jakob’s law: “Users spend most of their time on other sites.” They arrive with habits learned elsewhere and expect yours to work the same way. Nielsen asked sites to tone down their own visual design, their labels, their interaction and their structure, and pointed at blue links, the shopping cart and the Back button.

That is what invisibility looks like on a page. A convention is invisible because people learned it somewhere else. Every invention on your site is something they have to learn here, on your time and theirs. Nielsen called it a law, and named it after himself; it is an observation from years of usability work, not an experiment. But it is what you see every time you watch someone use a site that tried to be clever with its menu.

Readers asked him at the time whether this would make the web dull. His own answer was that plenty remains: working out what the people using a site need to do, and writing every piece of content well. I would put it more simply. The structure should be standard. The voice should not. What makes a site its own belongs in the words and the pictures, not in where the menu is.

Invisible is not the same as hidden

In 2003 Matthew Chalmers and Ian MacColl, at the University of Glasgow, pointed out that Weiser’s invisibility had been turned into a demand for “seamless” systems. They disagreed. A seam is where one part of a system meets another, or where it falls short: a weak signal, an inaccurate position, a step that failed. Hide every seam and people can no longer tell what is going on, or work around it. They noted that Weiser himself, in talks in the mid-1990s, warned that making everything seamless makes everything the same, and asked instead for beautiful seams.

On a website the seams are the moments where something happens: a form is sent or it is not, a page is loading, a payment went through, a link leaves the site. Those must be visible. And the worst kind of invisibility is a page that hides what it does — a script from a third party that the visitor never sees, collecting what the visitor never agreed to give. A quiet page hides itself. It never hides what it is doing.

Here is the brief from the company name, as a checklist.

The name outlived the company

Invisible Interfaces closed in 2019. The brief did not. It is still the first question I ask of any page I make or review: what here is asking to be looked at, and has it earned it? Most of the time the honest answer is to take something away.

Read the originals

• Mark Weiser, The Computer for the 21st Century, Scientific American, September 1991. Where the idea starts; the comparison with writing is in the opening paragraphs.
• Jakob Nielsen, End of Web Design, Nielsen Norman Group, July 2000. The first appearance of Jakob’s law, with the list of what should become standard and what remains to design.
• Matthew Chalmers and Ian MacColl, Seamful and Seamless Design in Ubiquitous Computing, 2003. The case against hiding everything, and the source for Weiser’s beautiful seams.
• Donald A. Norman, The Invisible Computer, MIT Press, 1998. The argument applied to products, from the author of The Design of Everyday Things.
• Golden Krishna, The Best Interface Is No Interface, New Riders, 2015. The idea taken to its extreme, and readable as a polemic against screens that want your time.