What the Screen Shows and How It Looks: Three Routes to Judging FinTech Interfaces

For anyone using a mobile wallet, the interface is where money becomes visible. Balances, payees, fees, confirmations and error messages all appear on a small screen, often in a hurry. How people judge that screen shapes whether they trust the service and keep using it. So providers with limited design budgets face a practical question: which qualities of the user experience should we improve first if we want users to rate the interface more favourably?

UX research offers plenty of candidates: attractiveness, efficiency, clarity, dependability, usefulness, trust, content quality and more. The trouble is that these qualities are strongly correlated. Put them all into one regression and the model explains a lot, but the individual coefficients become unstable, can flip sign or lose significance, and the "order of importance" depends on which predictors happen to be included.

Two competing answers

Two lines of theory point in different directions. Research on aesthetics shows that people who find an interface beautiful also tend to judge it usable and good (Tractinsky et al., 2000; Hassenzahl, 2004), which suggests attractiveness should lead. Research on information systems success and online trust points instead to the quality of information and to cues that reduce risk (DeLone & McLean, 2003; Cyr, 2008). In a financial interface, where every number matters, what the screen shows may weigh more than how it looks.

The study frames this through the components of user experience (CUE) model (Thüring & Mahlke, 2007), which separates instrumental qualities (such as efficiency and controllability) from non-instrumental ones (such as visual aesthetics). The CUE model does not treat the information a system presents as a separate component, even though in many applications that information is exactly what users come for. The study therefore tests three routes to interface judgement:

• Instrumental: efficiency, clarity, controllability and usefulness.
• Aesthetic: attractiveness and personalisation.
• Informational: content quality, and trust and security.

If the CUE model is complete, the informational route should add little once the other two are taken into account.

How it was studied

The data come from an online survey run between 30 March and 2 May 2023 among users of Palestinian FinTech applications. Of 310 respondents, 267 used a FinTech app and rated the one they used most (Jawwal Pay, Reflect, PalPay or Maalchat). Forty respondents (15%) gave identical answers to all 50 items, so the main analyses used the remaining 227 active users, and every model was re-run on the full sample as a check. The screened sample was mostly men (86.3%), mostly aged 26–44 (67.4%), mostly holders of a bachelor's degree (71.8%), and mostly living in the West Bank (64.3%).

Eight UX qualities were measured with four items each, adapted from the User Experience Questionnaire (Laugwitz et al., 2008) and UEQ+ scales (Schrepp & Thomaschewski, 2019). Interface evaluation was measured with 18 items covering visual design, interaction design, information architecture and brand.

Instead of relying on regression coefficients alone, the study used relative-weight analysis (Johnson, 2000) and dominance analysis (Budescu, 1993). Both divide the explained variance among correlated predictors so that the shares add up to the total, and both are recommended alongside regression when predictors overlap (Tonidandel & LeBreton, 2011). A set-level version of dominance analysis (Azen & Budescu, 2003) compared the three routes.

Key findings

1. Together, the qualities explain most of the picture

The eight qualities jointly explained 71.9% of the variance in overall interface evaluation. When the straight-line respondents were included, this figure rose to 79.3%, a reminder that uniform answering inflates apparent associations. User characteristics and the app used explained only 13.6% on their own.

2. Content quality and attractiveness lead

• Content quality: 17.8% of the explained variance.
• Attractiveness: 16.8%.
• Clarity: 14.5%, then efficiency: 12.9%.
• A middle group: trust and security (10.9%), controllability (10.3%) and personalisation (9.8%).
• Usefulness came last, with 6.9%, even though it had the highest average rating.

Content quality and attractiveness could not be statistically separated, but content quality carried more weight than trust and security.

3. The coefficient table would have misled

Looking at regression coefficients alone, usefulness had a negative coefficient (β = −.06) even though its correlation with the interface evaluation was .59, a classic sign of suppression. Controllability had a coefficient of zero although it accounted for about a tenth of the explained variance. Read alone, the table would suggest that useful apps get worse interface ratings, which the data do not support.

4. Different parts of the interface, different leaders

• Visual design: attractiveness led (21.6%).
• Information architecture: attractiveness led (17.8%).
• Interaction design: content quality led (19.5%), the part of the experience where users read confirmations, feedback and error messages.
• Brand: content quality led (16.4%), and personalisation mattered more here (14.5%) than anywhere else.
• Clarity contributed a stable 12.7% to 15.7% to every component.

5. Three routes, equal weight, unequal concentration

Each route contributed roughly a third of the explained variance: informational 34.3%, instrumental 33.5%, aesthetic 32.2%. None of the differences was significant. But the instrumental route needed four qualities to reach its share, while the aesthetic and informational routes did it with two each, so the average weight per quality was about twice as large. The results held when personalisation was moved to the instrumental route, in the full sample, and after controlling for the app, region, gender, age group and education.

What it means for designers and product teams

The results translate into an order of priorities:

1. Invest in the content layer. Accurate and current balances and statuses, transparent fees, and confirmations and error messages that people can actually understand. In a financial app, the content is the transaction.
2. Keep the visual design coherent and appealing. It drives judgements of visual design and information architecture.
3. Keep clarity high on every screen. It contributes steadily everywhere.
4. Treat personalisation as a brand investment, not a usability fix: it had the lowest average rating but mattered most for brand evaluations.
5. Do not expect better functions to fix interface ratings. Users seem to credit usefulness to the service (transfers, bill payments), not to the screen.

Trust deserves a closer reading. It sat in the middle, behind content quality, but that does not make it unimportant. At the interface level, users appear to build trust partly through the information they see: clear fees, clear confirmations and accurate balances are both content and trust cues.

In information-intensive apps, what the screen shows carries as much weight as how the screen works or how it looks.

For theory, this suggests the CUE model is incomplete for such applications: on-screen information behaves as a distinct route to judgement, not a sub-part of usability or usefulness. For research practice, it shows why relative importance analysis should accompany regression whenever UX qualities are correlated.

Limits

All measures came from the same respondents at one point in time, so associations may be inflated by common-method variance and by a general "halo" in evaluations (Nisbett & Wilson, 1977; Podsakoff et al., 2003). A single-factor test suggested the items did not collapse into one general evaluation, but this bias cannot be ruled out. Relative importance describes shared variance, not causal leverage. The convenience sample was mostly male and recruited online, some items were adapted from semantic-differential scales into agreement statements, and apps were unevenly spread across regions. The data were collected before October 2023 and do not describe the current situation in Gaza. Future work should link these profiles to behaviour (task completion, error rates) and test content and visual design experimentally.

About the original study

Qadah, W. What the Screen Shows and How It Looks: Instrumental, Aesthetic and Informational Routes to Judgements of FinTech Interfaces. Unpublished study by the author, 2026. The survey data were collected for the author's doctoral research at Istanbul Arel University.