The imperfection effect: why a perfect 5/5 rating can hurt trust on a landing page

Faced with a customer rating to display on a landing page, the instinct is to reach for the most flattering number possible: 5 out of 5, zero complaints, total satisfaction. That's often the least convincing choice. A visitor who lands on "5.0/5 — 340 reviews" asks a question that "4.8/5 — 340 reviews" almost never triggers: are these reviews even real? Statistical perfection, at a real customer volume, is rare — and a brain trained on reading online ratings knows it, even without putting the thought into words. This is sometimes called the imperfection effect: a small, visible flaw makes a trust signal more credible, not less.

What the research on mixed reviews shows

A study by Sun-Jae Doh and Jang-Sun Hwang, published in 2009 in CyberPsychology & Behavior (see on Google Scholar), exposed participants to sets of ten online reviews whose ratio of positive to negative comments varied across groups. The core finding: sets made entirely of positive reviews didn't consistently score highest on credibility — if anything, perceived credibility of the site and the reviews dropped when the set looked too uniformly favorable. A set mixing a few less glowing reviews into a mostly positive whole was rated more trustworthy than an entirely positive one.

A second study, by Nathalia Purnawirawan, Patrick De Pelsmacker and Nathalie Dens, published in 2012 in the Journal of Interactive Marketing (see on Google Scholar), confirms this mechanism from a related angle: the balance between positive and negative reviews within a set directly shapes how useful that set is perceived to be, independent of the resulting average rating. A perfectly homogeneous set — even a homogeneously positive one — is judged less useful for deciding than a set that lets a little nuance show through.

An older principle: the pratfall effect

This mechanism extends a much older finding from social psychology. In 1966, Elliot Aronson, Ben Willerman and Joanne Floyd published a now-famous experiment in Psychonomic Science (see on Google Scholar): participants listened to a recording of a quiz contestant, either highly competent or mediocre, who did or didn't spill a cup of coffee during the interview. The result, now known as the "pratfall effect": the blunder increased the perceived attractiveness of the competent contestant, but decreased that of the mediocre one. The nuance matters for a landing page: displaying a slightly imperfect rating only works if the overall impression is already strong. A small flaw humanizes an already solid case for trust; it doesn't manufacture one where none exists.

Why the brain distrusts a too-round score

A 5/5 score on a small sample (a handful of reviews) is plausible and raises no particular suspicion. The same score across a large volume — several hundred customers — becomes statistically suspect to a visitor used to Google, Trustpilot or Amazon reviews, where a perfect average at scale is, in practice, almost impossible to find. The brain applies reasoning similar to what's described in our article on processing fluency: information that's too smooth to be true triggers a cognitive discomfort that the visitor attributes not to the number itself, but to the reliability of the source displaying it.

How to apply this without sabotaging your social proof

• Show the real average, even at 4.7 or 4.8/5, rather than rounding up or filtering out the less enthusiastic reviews to hit a round number.
• Never fabricate an artificial flaw — the effect only works with an honest rating. Inventing a mixed review to appear credible carries the same risk as the fake testimonials covered in our guide on how many testimonials to display: a visitor who spots it loses trust in far more than that one detail.
• Leave one or two moderately critical reviews visible in an external review widget (Google, Trustpilot) instead of hiding them, especially when they touch on a minor point rather than the offer's core promise.
• Reply publicly to a critical review whenever possible: the reply itself becomes a proof of seriousness, often more persuasive than the total absence of any criticism.
• Reserve a perfect score for small samples where it stays plausible (under ten reviews, a recent launch), knowing it will naturally settle as volume grows.

What this doesn't mean

This principle isn't a license to go hunting for, or highlight, negative reviews about what actually matters to the buyer — the promised outcome, the service's reliability, the support. The imperfection effect plays out on secondary details (a slightly long delivery time, a feature that's still missing, a less-than-perfect interface), not on the page's core promise. It also doesn't replace the underlying work on social proof as a whole: the diversity of formats, the number of reviews and their individual credibility remain the main levers. An imperfect rating is a signal of authenticity that complements that work, not a substitute for it.

Measuring the effect on your own page

Like any copywriting lever built on a cognitive bias, the real effect depends on context, offer and audience. The best way to check it is to test two versions of the reviews block header — a rounded score against the real one, shown alongside the review count — following the method described in our A/B testing guide, rather than applying the principle blindly across the whole page.

The 10 LanderKit templates (€89 each, €229 for the full bundle) come with a ready-to-use reviews block built to display a real rating and a real review count rather than a rounded-up front, whether on the coach and consultant, single-product e-commerce or SaaS waitlist template. You can see the block in context on the coach-consultant demo or the e-commerce demo before adapting it to your own customer reviews.