Almost everything peer review knows about logo design was measured on logos that nobody had ever seen, printed in black and white, with the brand name removed. That is not a criticism invented here. It is the third stated limitation in the article that validated the field’s central framework across ten countries, co-signed by the researchers who built it: “our recommendations may not apply to brand/logo combinations.”
Which is a striking sentence, because a brand and logo combination is the only thing a company ever actually ships.
This does not make the research worthless. It makes it narrower than the confident advice built on top of it. Three things are genuinely established, one popular claim is contradicted by the primary evidence, and a surprising amount has simply never been studied. This article sorts them, and names the source and the sample every time.
The framework that replicated
The foundation is Henderson and Cote’s 1998 article in the Journal of Marketing. Its method, from the published abstract: “An empirical analysis of 195 logos, calibrated on 13 design characteristics.”
Those thirteen characteristics collapse into three underlying dimensions, and those three are what survived twenty-five years of subsequent work:
- Elaborateness is not intricacy. It is described as “design richness and the ability to use simple lines to capture the essence of something”, combining complexity, activeness and depth.
- Naturalness is “the degree to which the design depicts commonly experienced objects”, combining representativeness and organic form.
- Harmony is “a congruent pattern or arrangement of parts that combines symmetry and balance”.
In 2009, a team including both original authors tested this across ten countries in Marketing Science: Argentina, Australia, China, Germany, Great Britain, India, the Netherlands, Russia, Singapore and the United States. The three dimensions were present in all ten, and the countries reduced to three clusters, broadly the West, Asia and Russia.
That is a real replication, and it is more than most design advice can claim.
Three objectives, and you cannot have all three
The 1998 article’s real contribution is not the dimensions. It is the trade-off it makes explicit, in a single passage from the abstract:
“High-recognition logos (accurate recognition created by high investment) should be very natural, very harmonious, and moderately elaborate. Low-investment logos (false sense of knowing and positive affect) should be less natural and very harmonious. High image logos (professional look and strong positive image) must be moderately elaborate and natural.”
Read the parenthesis on the second one carefully. A low-investment logo buys a false sense of knowing. People feel they have seen it before when they have not. That is a genuine commercial asset for a company that cannot afford to buy repeated exposure, and it is a different asset from accurate recognition, which requires that exposure.
So the first question a company should answer is not what its logo should look like. It is which of those three outcomes it is buying, given what it can spend on being seen.
A caution before anyone tunes a design to those adjectives: the precise calibration levels differ between the published abstract and how later reviews summarize the original table. The direction is safe. The exact settings are not worth arguing over without the source table in hand.
The mechanism is a mistake your brain makes
Why does any of this work? The answer, established across four studies in the Journal of Consumer Research in 2001, is less flattering than the design profession’s account of it.
Seeing a logo repeatedly makes it easier to process. The brain notices the ease but does not know where the ease came from, so it attributes it to the object: this thing feels good, therefore I must like it. Preference does not come from a direct affective response to repetition. It comes from misattributed processing fluency.
That single finding explains most of what follows. It explains why familiarity reads as quality. It explains why a redesign feels wrong before anyone can say why: the fluency is gone, and the discomfort gets attributed to the design. And it explains why the answer to “simple or complex” turns out to depend entirely on how many times someone has seen the thing.
Simple or complex is the wrong question
The claim that simpler logos perform better is repeated constantly and is not supported by the primary research. Two peer-reviewed studies point the other way, and neither is obscure.
The exposure study. A 2016 article in the Journal of Marketing Communications ran two experiments: 68 participants on six unfamiliar logos, then 164 participants on eight familiar logos, both crossing complexity with number of exposures, with recognition timed in milliseconds. Its conclusion, verbatim:
“increases in exposure led to an increase in brand recognition and to more positive attitudes in particular for complex brand logos, suggesting short-term benefits for simple brand logos, and long-term benefits for complex logos.”
Worth noting how that article opens: “Although good logos are essential for creating brand awareness and brand equity, the effects of logo design features have not been tested empirically.” That was published in a peer-reviewed journal in 2016.
The complexity study. A 2014 article in Psychology and Marketing went further and split complexity in two. It used 505 participants recruited on Amazon Mechanical Turk (56.2 percent women, mean age 33.48) rating 140 black-and-white unfamiliar logos, with visual complexity coded independently by two professional graphic designers and two marketing experts, across three exposure conditions.
The result is a double crossover. Visual complexity helps on first sight and hurts with repetition. Conceptual complexity hurts on first sight and helps with repetition. Two opposing fluency mechanisms, running in opposite directions over time.
So the honest answer to “should our logo be simple” is: it depends on how visible you can afford to be, on whether the complexity is visual or conceptual, and on whether anyone already knows you.
Descriptive logos work, until people know who you are
The strongest recent finding is about whether a logo should say what the company does.
A 2019 article in the Journal of Marketing Research used six studies plus an analysis of 597 brand logos, and found that more descriptive logos “can positively influence brand evaluations, purchase intentions, and brand performance”, because they are easier to process and therefore “elicit stronger impressions of authenticity, which consumers value”.
But the two moderators the authors identify are what make the finding usable. The positive effect is “considerably attenuated for brands that are familiar (vs. unfamiliar) to consumers” and it is “reversed (i.e., negative) for brands that market a type of product/service linked with negatively (vs. positively) valenced associations in consumers’ minds”.
Translated into a decision: if nobody knows you, a logo that shows what you do buys you something real. Once people know you, it buys you very little. And in a category people already dislike, saying what you do out loud makes things worse.
The authors also report that in their sample of 597 logos, practitioners “might not fully take advantage of the potential benefits of logo descriptiveness”. The abstract categories most agencies gravitate toward are, on this evidence, the wrong default for a company nobody has heard of.
A redesign is judged hardest by the people who like you most
The only substantial body of work on changing a logo points in an uncomfortable direction.
A field experiment published in 2010 in the Journal of Product and Brand Management used, in the authors’ own words, “a field experiment with 632 respondents” examining two athletic shoe brands, New Balance and Adidas. The finding: the more strongly a previously angular logo was rounded, the more negatively highly committed consumers evaluated the redesign, and brand attitude followed. Consumers with low commitment reacted positively.
Which inverts the usual assumption. A redesign is not primarily a risk with people who do not know you. It is a risk with the people who already chose you. And they are the ones who notice.
The one longitudinal study located, published in 2021 on a university merger, found the same arc over time: the new logo first met resistance, then became a source of positive associations. That is consistent with the fluency mechanism. The discomfort is the cost of relearning, and it is paid up front by whoever already knew the old mark.
No peer-reviewed study linking a logo change to sales was found. Everything in this literature measures recognition, attitude, perceived authenticity or brand equity ratings. Anyone who tells you a redesign will move revenue is not citing research, because there is none to cite.
What nobody has studied
The 2019 systematic review in the Journal of Marketing Management is the most useful document in the field, because it is honest about the holes. Its method is worth stating: 434 articles found in Web of Science and 843 in Scopus, filtered to the seventy marketing journals listed in the Academic Journal Guide, read in full, reduced to a final sample of 124 articles spanning 1988 to 2018.
Its declared gaps include:
- Nobody has studied how logo elements work together. Typography, shape and color are each studied in isolation. Whether they reinforce or conflict has not been tested.
- Nobody has studied how to expose a logo. Which channel, at what frequency. The review states plainly that “no clear practical advice has been made.”
- Nobody has studied the reverse causality. The literature assumes logo perception drives company perception. Whether a liked company makes people like a mediocre logo has not been tested.
- Nobody has explained the minimalism wave. On the redesigns where components were deleted, the review notes that “little research has explored the underlying motivation of such changes in logo redesign”.
And the structural limitation, in the review’s own words: “the existing findings in the logo literature are mainly based on surveys or controlled lab-experiment data.”
What to actually do with this
Four things follow, and none of them require a research budget.
Decide the objective before the design. High recognition, low investment or high image are three different purchases. Only one of them is available to a company that cannot buy repeated exposure. The mark then has to survive being extended, which is why it is worth drawing the logo as part of a graphic system rather than as a single picture.
If you are unknown, let the mark say what you do. The evidence supports descriptiveness precisely in the situation most companies are in, and stops supporting it once you are known.
Weigh a redesign against your existing customers, not against your taste. The measured penalty falls on committed buyers. If the current mark is not actively failing on a constrained use case, the fluency you already own is an asset you are proposing to spend.
Do not accept “simpler is better” as a reason. It is a practice, documented as a practice, not a finding. The research says it depends on exposure, on familiarity, and on which kind of complexity you mean.
And keep the caveat in view, because it is the honest frame for all of the above. The strongest replication in this field tested unfamiliar logos, stripped of verbal material, judged mostly by students, and told you in its own limitations section that “our recommendations may not apply to brand/logo combinations.” Anyone selling certainty about your logo is selling something the research does not contain.