Marketing practice has long treated color and typeface as matters of taste. The empirical literature tells a more disciplined story: visual surface features shift brand personality attributions and change what a shopper believes a discount is worth. That same literature has also produced failed replications and outright reversals of its most-cited effects. This paper argues that the findings cohere around one mechanism, processing fluency, and that this mechanism explains both why surface features matter and why simple prescriptions about them fail.
The foundational claim is that hue carries semantic content that transfers to whatever it is applied to. Labrecque and Milne (2012) mapped hues onto brand personality dimensions across four studies, finding red conveyed excitement and blue competence, with saturation and value amplifying those traits independently of hue. The mechanism is learned rather than innate. Elliot's (2015) review describes color-in-context theory, in which a hue's effect depends on its setting: red signals failure in an achievement context and attraction in an affiliative one. That context-dependence is both the theory's strength and its practical liability.
Cross-cultural evidence complicates this without dissolving it. Jonauskaite and colleagues (2020) surveyed 4,598 participants across 30 nations and 22 languages and found color-emotion associations that were substantially universal, averaging a pattern-similarity coefficient of r = .88, with residual variation predictable from linguistic and geographic proximity.
Typeface operates on the same principle. Henderson, Giese, and Cote (2004) reduced typeface design to six dimensions and showed they predict strategically valuable impressions such as reassuring and engaging, while documenting the trade-offs: a face optimized for one impression forfeits another. Piovesan, Sinico, and Perondi (2023) later consolidated 34 studies covering 229 semantic qualities across 635 typefaces into a single open dataset, motivated by the observation that practitioners were choosing type from personal experience because the evidenceColor and existed only in scattered form. Whether these effects extend beyond Latin letterforms remains untested.

Processing fluency, the subjective ease with which information is handled, supplies the common mechanism. Song and Schwarz (2008) showed its reach: participants who read exercise instructions in a difficult typeface estimated the routine would take longer and were less willing to attempt it. The typeface changed nothing about the task, only the felt difficulty of imagining it, and that feeling was misattributed to the task.
Fluency also explains why color effects reverse. Kim and Jang (2022) tested the retail convention that red signals savings. With a deep discount, red raised perceived value relative to white, blue, or green; with a shallow discount of three to five percent, red significantly lowered it below the neutral controls, an effect fully mediated by processing fluency. The color did not fail.
It made a promise the offer could not keep, and the incongruity registered as disfluency.
Caution is warranted about headline results. Mehta and Zhu (2009) reported in Science that red enhanced detail-oriented performance while blue enhanced creative performance. The
finding entered marketing curricula quickly, but Steele (2014) ran a high-powered replication and failed to reproduce the interaction.
Color and typography are semantic channels, not ornament, and they are read before the copy is. Their effects are relational rather than absolute: congruity between cue and claim
predicts outcomes better than any property of the cue alone, which is why Cyr, Head, and Larios (2010) found color appeal to be a significant determinant of website trust while resisting a single optimal palette across cultures. The practical rule is negative, since a cue that overstates the offer will be punished.
Elliot (2015) closes his review by calling the literature nascent. Effect sizes are modest, replication records uneven, and most studies rely on undergraduate samples. The defensibleColor and claim is not that a particular blue sells a product, but that a message's surface and its substance are read together, and a mismatch between them is costly