Aesthetics vs Functionality in Digital Service Design: What Research and EU Law Say

Aesthetics vs functionality is one of the oldest arguments in design, and in digital service design it is mostly a false choice. How an interface looks changes how easy people believe it is to use. How it works decides whether they finish the task. A good service needs the two to agree.

The research behind that claim goes back three decades, to a 1995 study of ATM screens that gave the aesthetic-usability effect its name. Later studies complicated the picture in useful ways, and since June 2025 part of the argument has been settled by law: the European Accessibility Act turned contrast, focus states, target sizes, and error handling into compliance requirements for most businesses selling to EU consumers.

The stakes are highest in services people do not choose to use. Nobody browses a tax portal for pleasure or bookmarks a toll payment page. When a service is mandatory, poor design does not cost a conversion. It costs the user time, money, or a fine, and it leaves the operator with a support burden that never goes away. That is why public sector platforms are the clearest testing ground for service design principles, and why commercial designers should study them.

This guide covers what the research says about beauty and usability (including where it disagrees), what EU accessibility law and WCAG 2.2 now require of the visual layer, what three European public services get right, and how to apply all of it to commercial work.

Key takeaways

• The aesthetic-usability effect is real but conditional. Attractive interfaces are judged easier to use, especially on first impression, yet one seven-week experiment found no effect at all.
• Visual quality buys tolerance for minor friction. It never buys forgiveness for a broken flow.
• Under the European Accessibility Act, color contrast, visible focus, target size, and error handling are compliance requirements. EN 301 549 v4.1.1, published in September 2026, moves the benchmark to WCAG 2.2.
• If independent websites exist to explain how your service works, your service has an explanation gap.

The Aesthetic-Usability Effect: Why Beautiful Interfaces Feel Easier

The aesthetic-usability effect is the tendency to judge attractive interfaces as easier to use than plain ones, even when both work exactly the same. It is one of the best-documented findings in UX research, and one of the most often oversimplified.

The research on this predates what most designers assume. In 1995, Masaaki Kurosu and Kaori Kashimura at the Hitachi Design Center showed 252 participants 26 variations of an ATM interface and asked them to rate each on ease of use and on visual appeal. All 26 had identical functionality.

The correlation between perceived beauty and perceived ease of use was stronger than the correlation between perceived beauty and actual ease of use. People were reading aesthetics as a signal of usability, even when they were asked to judge function. Nielsen Norman Group’s write-up of the aesthetic-usability effect covers the study and its practical limits in more depth, and it is the reference worth citing rather than the secondhand versions circulating on design blogs.

Two consequences matter for anyone building a service.

Aesthetics buy you patience, not forgiveness. A well-designed interface earns tolerance for minor friction. Users are more willing to persist, less likely to blame themselves, and more likely to report the experience positively afterward. That tolerance is real and it is valuable.

The effect has limits, and severity is the clearest one. It is strongest on first impressions and small problems. A visually polished service with a genuinely broken flow does not get a pass. It gets a worse reaction, because the visual promise makes the failure feel like a betrayal rather than an expectation met. How long the goodwill lasts is less settled than most design blogs suggest, as the next section shows. This is where “make it pretty” as a strategy falls apart.

The honest reading is that visual quality is a multiplier on functional quality, not a substitute for it. Multiply zero and you still have zero.

Is the Aesthetic-Usability Effect Real? What Later Research Found

Yes, but with conditions. The 1995 ATM study is quoted on nearly every UX blog, but it was one study, run in Japan, measuring opinions rather than behavior. The research that followed is more useful to designers than the original, because it shows where the effect holds and where it disappears.

It held up across cultures. In 1997, Noam Tractinsky repeated the ATM experiment in Israel, expecting the effect to be weaker there. The correlations between perceived beauty and perceived ease of use were very high in both countries, and not weaker in Israel.

It survived actual use, at least in the short term. In a 2000 study titled “What is beautiful is usable,” Tractinsky, Adi Katz, and Dror Ikar measured perceptions before and after participants used a system. The link between perceived beauty and perceived usability stayed intact after use. More striking, how attractive the system was shaped how usable people judged it afterward, while how usable it actually was did not.

It can also vanish. In a 2022 experiment, Jürgen Sauer and Andreas Sonderegger followed 110 people operating a fully automated coffee machine over seven weeks, with both visual aesthetics and usability deliberately varied. They found no effect of visual aesthetics on user experience, including perceived usability, which put them at odds with much of the earlier literature.

Our reading of the evidence: the effect is real but conditional. It is most reliable for screen interfaces judged quickly on first impression, which is exactly how most people meet a digital service they use once a year. It is far less reliable as a strategy for products people operate often, where performance becomes obvious. That is the practical answer to aesthetics vs functionality: visual quality shapes the judgment, functional quality shapes the outcome, and only the outcome keeps users coming back.

It also explains a pattern visible across the web. Polished visual presets make a site look credible on first view, and that low-risk payoff is part of why so many websites now look the same.

How to Tell Whether Good Looks Are Hiding Usability Problems

The real risk is not that users like a beautiful interface. It is that the team stops seeing the interface’s problems, because the same bias that flatters the design also colors usability testing. Four checks separate the two:

1. Score behavior before opinion. Record task success, time on task, and error counts before you read satisfaction ratings. High ratings with low completion mean the visual layer is covering for the flow.
2. Test a stripped-down build. Run the same tasks on a grayscale, low-fidelity version. If completion holds and ratings drop, aesthetics are carrying perception. If completion drops too, the visual hierarchy is doing functional work worth protecting.
3. Test the second visit. Bring participants back a week later with a new task. First-impression goodwill is spent by then, so the friction that remains is what regular users will feel.
4. Trust drop-off data over compliments. Analytics does not see how a page looks. A step that wins praise in design reviews but loses most users is a functional problem. In e-commerce that step is usually checkout, which is where online stores that get traffic but no sales tend to lose people.

Digital Public Services and the New Compliance Floor

Governments and private companies alike have spent a decade moving services online. What changed recently is that the quality of those services stopped being a matter of taste.

The European Accessibility Act became enforceable on June 28, 2025. Service providers are exempt only if they are microenterprises, with fewer than 10 employees and no more than two million euros in annual turnover or balance sheet total. Everyone else offering covered products or services to EU consumers is in scope, which sweeps in e-commerce, banking, transport ticketing, and telecoms.

The technical yardstick is EN 301 549. The version currently cited for legal purposes, v3.2.1, references WCAG 2.1 Level AA. A new version, v4.1.1, was published in September 2026 and adopts WCAG 2.2; it becomes the reference once the European Commission cites it in the Official Journal. Enforcement is national, and for new products and services it started on day one. Transition rules mainly protect service contracts agreed before June 28, 2025 and products already in service, both until June 2030 at the latest, and member states may let self-service terminals such as ticket machines stay in use for up to 20 years. In France, disability associations took four major grocery retailers to court in late 2025 over inaccessible online shopping. The first ruling, in Auchan’s case in May 2026, went to the retailer because the court applied an older French law with a higher revenue threshold, even though it described the site’s accessibility as fairly low.

For designers this is not a legal footnote, it is a set of constraints on the visual layer. Color contrast ratios, visible focus states, text resizing without loss of content, target sizes, and error identification are all design decisions, not developer decisions. A brand palette that fails contrast at 4.5:1 is now a compliance problem, not a debate with the marketing team.

The useful reframe is that accessibility requirements and good design point the same direction. Clear typography, generous target sizes, obvious focus states, and plain language make services better for everyone, not just for users with disabilities. If contrast is where your palette is likely to break, tools like ColorShark let you check schemes against accessibility requirements before they get committed to a design system rather than after a legal notice arrives.

A WCAG 2.2 Checklist for Designers: What the Visual Layer Must Do

Most accessibility checklists are written for developers. The requirements below are decided in design files, before a line of code exists. The legal baseline under the European Accessibility Act is still WCAG 2.1 Level AA, through EN 301 549 v3.2.1. The updated EN 301 549 v4.1.1, published in September 2026, adopts WCAG 2.2 and becomes the reference once the European Commission cites it in the Official Journal. Designing to WCAG 2.2 now means not redesigning the same components twice. The W3C summary of what’s new in WCAG 2.2 lists nine added success criteria. Six of them sit at Level A or AA, and most land squarely on the designer’s desk.

Requirement	WCAG criterion	What it means in the design file
Text color contrast	1.4.3, AA	4.5:1 for body text, 3:1 for large text (about 24px regular or 18.66px bold)
Non-text contrast	1.4.11, AA	3:1 for input borders, icons, and focus indicators against adjacent colors
Color is not the only signal	1.4.1, A	Errors, required fields, and status need an icon, label, or text, not just red or green
Reflow and resize	1.4.10 and 1.4.4, AA	Layouts work at 320 CSS px wide without horizontal scrolling, and text scales to 200 percent without breaking
Text spacing	1.4.12, AA	Nothing clips or overlaps when users raise line height to 1.5 and paragraph spacing to 2 times the font size
Focus not obscured	2.4.11, AA (new in 2.2)	Sticky headers, chat widgets, and cookie banners never fully cover the element with keyboard focus
Target size	2.5.8, AA (new in 2.2)	Tap and click targets are at least 24 by 24 CSS px, or spaced so a 24px circle around each one touches no other target
Dragging alternatives	2.5.7, AA (new in 2.2)	Sliders, sortable lists, and map pins also work with a single tap or click
Consistent help	3.2.6, A (new in 2.2)	Help links, chat, or contact details sit in the same place on every page that offers them
Redundant entry	3.3.7, A (new in 2.2)	Information already given in the same process is prefilled or selectable, never typed twice
Accessible authentication	3.3.8, AA (new in 2.2)	Login does not depend on memorizing or transcribing, and paste and password managers work
Error prevention for payments and legal commitments	3.3.4, AA	Submissions are reversible, checked for errors, or confirmed before they become final

Two of these rows connect directly to the public services later in this article. Redundant entry is Estonia’s once-only principle applied to a single form. Error prevention for financial and legal transactions is the exact problem the Czech vignette portal has to solve, because a paid vignette is tied to a license plate and some purchases cannot be changed after payment.

Type is where most brand palettes and component libraries fail first. Light gray body copy, thin weights at small sizes, and tight line height are the usual offenders, and our guide to web typography best practices for readability covers how to fix them without flattening the brand. Forms follow the same logic: an error message that appears inline, explains the fix, and keeps what the user typed does more for completion than any visual polish. The UX blueprint for a multi-step quote calculator shows that pattern working in a real booking flow.

Three European Public Services and What Each One Gets Right

Rather than one example, three, because the lessons are different in each case.

GOV.UK and the discipline of subtraction. The UK’s government service design work is the most studied in the field, and its defining move is removal. Pages carry one task. UK government design guidance recommends writing for a maximum reading age of 9, even for specialist audiences. The design system is public, versioned, and used across hundreds of services, which means a citizen who learns one service has already learned the next. The transferable idea is that consistency is a usability feature, not a branding preference.

Estonia and the once-only principle. Estonia’s digital state is built on the rule that a citizen should never have to supply the same information twice. That is an architectural decision with a design consequence: forms get shorter because the system already knows the answer. Most commercial products ask users to retype information they already hold. The design fix is often a data fix.

Czechia’s electronic vignette and the mandatory transaction. Czechia replaced paper motorway stickers in 2021 with an electronic system where drivers register a license plate and pay online. The official e-shop is edalnice.cz (also reachable at edalnice.gov.cz), operated by the State Fund for Transport Infrastructure. The interesting design problem here is not aesthetics at all. The service is compulsory, many users are foreign drivers passing through, and a mistyped plate can leave a driver without a valid vignette. The correction rules depend on how and when you buy: at a point of sale, mistakes can be fixed within 15 minutes; online with a later start date, the plate can be changed once; online with immediate validity, the purchase cannot be changed or cancelled at all. Driving without a valid vignette can mean an on-the-spot fine of up to CZK 5,000. That combination puts enormous weight on error prevention, confirmation states, and multilingual clarity, which are unglamorous parts of design that rarely win awards.

What the three share is that the visual layer does a job. None of it is decoration.

When Third-Party Explainers Appear, the Official Service Has a Design Problem

Here is a diagnostic worth adopting. Search for any public service you are studying and count how many independent sites exist purely to explain it. If there are several, the official service is functionally complete but not comprehensible.

The Czech vignette illustrates this well. When we searched “czech vignette” in late September 2026, most of the top results were independent guides or resellers, and the official e-shop was not among them. That ecosystem splits into two very different kinds of site. Information-only guides such as dalnicniznamka.cz explain prices, validity periods, vehicle categories, and exemptions in plain language across several languages, state that they sell nothing and charge no fees, and send readers to edalnice.cz to buy. Resellers do the opposite: they sell the same vignette with a handling fee on top of the regulated price. The problem is large enough that the Czech Ministry of Transport has published a list of dozens of unofficial sites selling overpriced vignettes, and Czech embassies have warned drivers to buy only through edalnice.cz.

Both kinds of site point at the same design finding. When a service is mandatory, its rules are non-obvious, and its users include people who do not speak the language, an explanation gap opens up. Something will fill that gap. If the official service does not, third parties will, and some of them will monetize the confusion.

The lesson for commercial designers is direct. Any time you find a Reddit thread, a YouTube walkthrough, or a comparison site explaining how your product works, that is unpaid user research telling you exactly which part of your interface failed to explain itself. The content those explainers produce is the content your product should have contained.

Service Design Principles That Hold Up Under Real Constraints

The patterns that work in high-stakes services are not exotic. They are the ordinary ones, applied without exception.

Principle	What it looks like in practice	What breaks without it
Reduce cognitive load	One task per screen, no competing calls to action	Users abandon mid-flow and restart
Progressive disclosure	Show the common path, hide edge cases behind clear links	The interface reads as more complex than the task
Plain language	Write at the reading level of your least confident user	Support volume rises permanently
Error prevention	Format hints, input masks, confirmation before commit	Wrong data enters the system and cannot be undone cheaply
Visible system status	Confirmation screens, receipts, status lookups	Users repeat transactions because they cannot tell if the first worked
Consistent components	One button, one form field, one error pattern across the service	Every screen becomes a fresh learning problem
Mobile responsiveness	Test on a mid-range phone on a poor connection, not a design laptop	The majority of real sessions are the worst sessions

The row people skip is visible system status. In any transaction where the user cannot see the result, they will repeat it. Duplicate submissions are almost always a feedback failure rather than a user error.

Multilingual support belongs on this list too, and it is more than translation. Currency formats, date orders, name field assumptions, and address structures all encode assumptions that break at borders. The design principles behind this apply just as much to app interfaces, and our guide to mobile application design principles covers how the same constraints play out on small screens.

What Commercial Designers Should Take From Public Sector Work

Public services optimize for a metric commercial products rarely track: task completion by the least capable user, on the worst device, under the most pressure. That constraint produces patterns worth stealing.

Design for the person having a bad day. Commercial design tends to optimize for the engaged, motivated user. Public service design cannot. Building for the distracted, rushed, unfamiliar user makes the product better for everyone else by default.

Treat the confirmation screen as a real design surface. It is usually the last thing designed and the first thing users screenshot. Give it a reference number, a summary of what was submitted, and a clear statement of what happens next.

Write the error messages first. If you can articulate every way the flow can fail and what the user should do about each one, the happy path largely designs itself.

None of this conflicts with visual ambition. A portfolio site, a booking platform, or a product landing page can be visually distinctive and still respect these constraints. The relationship between visual identity and commercial outcome is covered in more depth in our piece on where web design meets digital marketing, and for product teams specifically, the new SaaS homepage brief looks at how design decisions now have to survive being cropped, screenshotted, and redistributed.

What Designers Ask About Beauty, Usability, and the European Accessibility Act

Design That Works Is Design That Looks Like It Works

The form-versus-function debate is a false one. Thirty years of research, starting with the 1995 ATM study, point the same way on the part that matters: people often read visual quality as a promise about how a thing will behave. Good design keeps that promise. Decoration breaks it.

What has changed recently is that in Europe, part of this is no longer discretionary. Contrast, focus states, target sizes, and clear error handling are now compliance requirements for a large share of commercial digital services, not preferences a designer argues for. That is not a constraint on good design. It is a floor that most good design was already above.

The services worth studying, public or private, are the ones where you cannot tell where the aesthetics end and the function begins.

Sources

Research, standards, and official guidance cited in this article, checked on September 30, 2026:

• Kurosu, M. and Kashimura, K. (1995). Apparent usability vs. inherent usability: experimental analysis on the determinants of the apparent usability. CHI ’95 Conference Companion, ACM.
• Tractinsky, N. (1997). Aesthetics and apparent usability: empirically assessing cultural and methodological issues. Proceedings of CHI ’97, ACM.
• Tractinsky, N., Katz, A. and Ikar, D. (2000). What is beautiful is usable. Interacting with Computers, 13(2).
• Sauer, J. and Sonderegger, A. (2022). Visual aesthetics and user experience: A multiple-session experiment. International Journal of Human-Computer Studies, 165.
• Directive (EU) 2019/882 on the accessibility requirements for products and services (the European Accessibility Act), including Article 32 on transitional measures.
• ETSI EN 301 549 v4.1.1 (2026), as summarized in the European Commission’s AccessibleEU announcement of the updated standard.
• W3C, Web Content Accessibility Guidelines (WCAG) 2.2, and the W3C summary of what changed from WCAG 2.1.
• Czech Ministry of Transport, guidance on electronic motorway vignettes and its warning about unofficial resellers.
• UK Home Office Design Manual, readability guidance for written content.

Editorial note: Sites Gallery has no commercial relationship with dalnicniznamka.cz or with any other site linked in this article. It is included as an example of an information-only guide that sells nothing and sends readers to the official e-shop.