Beautiful Digital Experiences Research Journal — Cycle 1

Cycle frame

• Cycle: 1
• Largest uncertainty: whether the project can define and measure aesthetic quality without collapsing it into liking, usability, fluency, or fashion.
• Scope: operational definitions, construct boundaries, an initial scientific evidence pass, initialization of the 15 required hypotheses, and a falsification plan.
• Out of scope: claiming universal laws, ranking live products, trend forecasting, cultural generalization, and validating a complete evaluation rubric.
• Repository state reviewed: Composition Science constitution and governance, Project Atlas visual-information foundations and typography work, the familiarity research report, and the Visual Engineering measurement REP.

2026-07-26 — Search decision 001

Decision: begin with construct and mechanism evidence rather than contemporary galleries or awards.

Reason: awards can reveal professional taste but cannot establish that beauty, usability, trust, or performance caused the judgment. The first cycle requires a measurement language before corpus observations can be interpreted.

Queries and source classes pursued:

• aesthetic–usability effect and the separation of perceived from measured usability;
• processing fluency, prediction, interest, boredom, and aesthetic pleasure;
• website prototypicality, complexity, and exposure duration;
• multidimensional measures of perceived website aesthetics;
• repeated exposure and longitudinal use;
• credibility judgments and visual design;
• accessibility standards as constraints, not beauty evidence.

2026-07-26 — Observations 001–010

The following are observations from sources, kept separate from the interpretations below.

OBS-BDE-001

Kurosu and Kashimura reported that apparent usability judgments of ATM layouts were more strongly related to aesthetic aspects than to their experimental inherent-usability measure. The source is a short CHI companion paper and does not establish a universal causal law.

OBS-BDE-002

Tractinsky, Katz, and Ikar found strong relations between perceived aesthetics and perceived usability before and after use. Their experiment also manipulated actual usability, so its results do not warrant rewriting “beautiful is perceived as usable” as “beautiful performs better.”

OBS-BDE-003

Sonderegger and Sauer compared visually appealing and unappealing but functionally identical mobile-phone simulations. Appealing presentation raised perceived usability; objective usability quality did not differ. The participants were 60 adolescents.

OBS-BDE-004

In a two-week field experiment, the positive effect of aesthetic appeal on perceived usability weakened with exposure. Aesthetics and inherent usability were independently manipulated.

OBS-BDE-005

Reber, Schwarz, and Winkielman proposed that processing fluency contributes positive affect and aesthetic pleasure. Their account links preference effects from contrast, repetition, symmetry, and prototypicality through the perceiver's processing experience.

OBS-BDE-006

The Pleasure–Interest model distinguishes immediate fluent pleasure from elaborated interest and explicitly accommodates boredom and confusion. Difficult-to-process work can become liked when it affords rewarding elaboration; fluent work can become boring.

OBS-BDE-007

A 2024 review integrates fluency with predictive processing and epistemic motivation. It treats fluency as expectation-dependent and notes empirical and conceptual challenges to a simple “easier is always more beautiful” account.

OBS-BDE-008

Tuch and colleagues found that visual complexity and website prototypicality influenced aesthetic ratings after very brief screenshot exposures. Low complexity and high prototypicality were favored in that stimulus set. The work concerns first impressions of screenshots, not full experience quality or repeated use.

OBS-BDE-009

Lavie and Tractinsky recovered two distinguishable dimensions in perceived website aesthetics: classical aesthetics, emphasizing clarity and order, and expressive aesthetics, emphasizing creativity and convention-breaking.

OBS-BDE-010

Moshagen and Thielsch's VisAWI work recovered four interrelated facets—simplicity, diversity, colorfulness, and craftsmanship—across a multi-study measurement program. This is evidence that perceived visual aesthetics is not well represented by one undifferentiated adjective.

2026-07-26 — Interpretations 001–008

These are working interpretations, not direct source findings.

INT-BDE-001

“Beauty” should be modeled as an attributed quality of an experience, not an intrinsic scalar property of a screen. The attribution depends on stimulus relations, perceiver state and history, task, category expectations, culture, and time.

INT-BDE-002

The aesthetic–usability effect is most defensible as a halo on perceived usability and related expectations. Effects on task performance are conditional and can be null, positive, or negative.

INT-BDE-003

Fluency is one pathway to pleasure, confidence, and familiarity. It cannot explain interest, expressive difficulty, culturally learned meaning, or rewarding surprise by itself.

INT-BDE-004

Prototypicality is a plausible mechanism for category legibility and predictive fit, but high prototypicality cannot be treated as a universal optimum. It may reduce identity, interest, and memorability, and the principal website evidence reviewed here uses static screenshots.

INT-BDE-005

At least two partly independent aesthetic routes are needed in the initial theory: coherence/fluency and expression/interest. “Appropriateness” gates both by category, audience, task, culture, and brand.

INT-BDE-006

Repeated use is not merely a stronger version of first-impression measurement. It can reverse or attenuate halo effects and expose unresolved states, friction, or superficial novelty.

INT-BDE-007

Accessibility conformance defines necessary constraints and behaviors for inclusive use; it is not evidence that a design is beautiful. Conversely, no reviewed evidence supports the claim that inaccessible low contrast or hidden controls are necessary for beauty.

INT-BDE-008

The proposed eight-layer model is better treated as a dependency graph than a strict ladder. Functional failure can dominate evaluation, but expression and coherence can influence first impressions before usability is tested; contextual mismatch can also damage trust despite high craft.

Failed assumptions and corrections

FA-BDE-001 — “Fluency is the common denominator of beauty”

• Status: rejected in strong form.
• Why: the pleasure–interest literature explains how difficulty can support interest and how fluency can become boring. Fluency remains a mechanism, not the construct.

FA-BDE-002 — “The aesthetic–usability effect improves usability”

• Status: narrowed.
• Why: the strongest recurring evidence concerns perceived usability. Objective performance effects are conditional and sometimes absent or slower.

FA-BDE-003 — “First impressions are noisy previews of durable judgments”

• Status: unsupported.
• Why: longitudinal evidence indicates attenuation, while screenshot studies omit interaction, real content, errors, mobile states, and repeated exposure.

FA-BDE-004 — “Accessibility is another aesthetic facet”

• Status: rejected.
• Why: accessibility and aesthetic evaluation overlap through perceptual and interaction mechanisms but answer different questions and require separate measures.

Counterexamples sought

• A highly fluent, prototypical website that feels generic or sterile.
• A demanding or initially disfluent experience that becomes admired through mastery.
• An attractive interface that produces slower completion or more consequential errors.
• A visually unconventional interface that remains legible because semantic and perceptual relationships are preserved.
• An accessible interface whose visible constraints increase expressive quality.
• A meticulously polished product that remains untrustworthy because of content, business model, or category mismatch.

No named product counterexample is promoted in this cycle; candidate products require dated, viewport-specific case records and evidence beyond memory or reputation.

Confidence update

• Multidimensionality of perceived website aesthetics: medium.
• Aesthetic influence on perceived usability: medium.
• Reliable improvement of measured task performance: low / conditional.
• Fluency as one contributor to aesthetic pleasure: medium.
• Prototypicality as a contributor to rapid website preference: low-to-medium, bounded to the studied screenshot conditions.
• Repeated-use attenuation of aesthetic halo: low-to-medium.
• Universal or cross-cultural transfer of any current claim: very low.

Challenge questions

• Are we confusing fashion with beauty? Not in this cycle; trend sources were deliberately deferred. The risk returns when the case corpus begins.
• Are we confusing professional taste with public preference? No professional-jury evidence was used as population evidence.
• Are we confusing visual simplicity with quality? VisAWI and fluency research make simplicity one facet or route, not the whole construct.
• Are we evaluating screenshots instead of experiences? Some foundational website evidence does. This is recorded as a major boundary and motivates within-product, repeated-use studies.
• Are we overgeneralizing from Apple or luxury brands? Neither was used as a foundation in this cycle.
• Does design work because of visual system or content, photography, reputation, and budget? Unresolved; a factorial art-direction experiment is needed.
• Would the same design work in another category? Unknown; category transfer is a planned falsification.
• Does it remain strong with real content, mobile, and accessibility constraints? Not inferable from the reviewed screenshot studies.
• Are participants reporting social prestige rather than preference? Existing measures do not rule this out.
• Are we measuring perceived instead of actual usability? Frequently; all records now name the construct explicitly.
• Are we treating Western preference as universal? The evidence base is too narrow for cultural generalization.
• What would prove the current conclusion wrong? Stable objective-performance improvements caused by aesthetic manipulation across tasks, populations, devices, and repeated use would overturn the current narrow account of HY-BEAUTY-001.
• What important counterexample is unexplained? Rewardingly complex interfaces and expressive expert tools are not yet represented.
• What loved design violates the theory? Not yet tested with a defensible corpus.
• What theory-conforming design still feels unattractive? A coherent, prototypical, fluent but anonymous interface is the priority candidate class.

Highest-value next investigation

Build a stratified case-study protocol and pilot corpus focused on category expectation and complete-experience evaluation. Sample at least finance, healthcare, public sector, developer tools, editorial, luxury, and one non-Western product family. Observe each at mobile and desktop viewports, with realistic content and at least one task and one failure/empty state. Use this to test whether category fit predicts appropriateness and trust independently of beauty, and whether controlled novelty predicts distinction without reducing clarity.

Revision history