Product ratings have become one of the most powerful signals in modern commerce. From booking a hotel to selecting enterprise software, consumers increasingly lean on aggregated scores as a proxy for quality. Yet new research suggests that something as simple as how those ratings are displayed – stars versus numbers—can systematically distort perception, with significant implications for businesses.
A study published in the Journal of Marketing Research has found that consumers interpret fractional ratings differently depending on whether they are presented visually, as stars, or numerically. The distinction is subtle but consequential: a score of 3.5 appears materially higher when shown as stars than when displayed as a number.
The research, led by Deepak Sirwani of the University of British Columbia and Manoj Thomas of Cornell University, sheds light on how cognitive processing influences decision-making. It also raises a broader question for companies: are they accurately communicating value, or inadvertently misleading their own customers?
The psychology behind stars and numbers
At the core of the findings is a divergence in how the brain processes imagery versus numerals. When consumers view a typical five-star rating with a half-star included, they instinctively “complete” the visual, turning the partial star into something closer to a full one.
This visual bias has a measurable effect. In six controlled experiments, the researchers consistently found that participants overestimated fractional star ratings, perceiving them as closer to the next whole number. By contrast, when the same ratings were expressed numerically, consumers tended to anchor on the left-most digit. A rating of 3.5, therefore, felt more like a 3 than a 4.
In essence, the same underlying score produces two different perceptions depending solely on format. The result is a systematic distortion: stars inflate, numbers deflate. For marketers and product teams, this is not an academic curiosity. Ratings influence conversion rates, pricing power and brand perception. In digital marketplaces, small differences in perceived quality can translate into disproportionately large differences in sales.
The significance of these findings becomes clearer when set against the growing importance of ratings in consumer decision-making. Online reviews now function alongside price and brand as primary factors shaping purchasing behaviour.
Previous studies have indicated that even marginal shifts in ratings, sometimes as little as 0.2 points, can affect sales outcomes. The Cornell-led research implies that such shifts may not always reflect real product differences, but rather presentation effects.
This introduces a new layer of strategic complexity. A company displaying ratings numerically may inadvertently undersell its product relative to a competitor using stars. Conversely, star-based systems may create elevated expectations that are difficult to meet, increasing the risk of customer dissatisfaction.
Whether intentional or not, format becomes a competitive variable.
Canadian businesses navigating the rating economy
These dynamics are particularly relevant in Canada’s rapidly evolving digital economy, where platform-based business models and AI-driven recommendations are becoming more pervasive.
- E-commerce (e.g., Shopify-powered stores): star icons are the primary rating signal.
- Travel and booking (e.g., Hopper): star-style or visual scores dominate.
- App marketplaces and SaaS tools: star ratings are the standard.
Shopify, which powers a large ecosystem of online merchants globally. Within its app marketplace and merchant storefronts, ratings play a central role in shaping adoption decisions. A subtle format bias, stars versus numbers, could influence how merchants select tools, ultimately affecting which services achieve scale.
Similarly, Canadian travel platform Hopper, which uses predictive analytics to forecast pricing, relies heavily on customer trust and perceived reliability. In a sector where users compare dozens of options, rating presentation can sway booking decisions, especially when differences between competing services are marginal.
Enterprise AI firm Cohere offers another angle. While its customers are businesses rather than consumers, evaluation metrics and benchmarking scores perform a similar function to ratings. How performance is visualised as charts, scores, or qualitative indicators can shape client perception in analogous ways.
Even in healthcare and life sciences, where firms like Deep Genomics operate, communication of model performance and predictive accuracy must balance clarity with interpretability. Numerical precision is essential, but human interpretation remains central.
Across these sectors, the lesson is consistent: presentation influences perception, even when underlying data remains unchanged.
Implications for trust and transparency
The Cornell findings raise a broader issue around transparency in digital markets. If common rating formats are systematically misinterpreted, then a large portion of consumer decision-making is built on cognitive shortcuts rather than objective assessment.
For regulators and platform operators, this presents a challenge. Standardisation of rating systems could reduce variability, but may also limit flexibility for businesses. Alternatively, hybrid formats, such as combining stars with explicit numerical values, might provide a more balanced signal or, alternatively, confuse consumers further.
There is also a role for design. Platforms could experiment with clearer visual cues, such as scaling indicators or contextual benchmarks, to reduce misinterpretation. Small adjustments to user interfaces may yield outsized improvements in decision quality.
From a compliance perspective, companies should consider whether their rating displays align with principles of fair representation. In sectors where consumer protection is critical, misleading impressions—even if unintentional—could attract scrutiny.
Given that visual cues and numerical data appear to trigger different cognitive pathways, then in a marketplace increasingly mediated by algorithms, reviews and dashboards, these differences matter. Businesses that understand the behavioural dimension of data presentation are better positioned to communicate value accurately.