Ratings on retail platforms cluster heavily at the top of the scale, with a smaller cluster at the bottom and very little between. The shape is produced by the collection process, not by unusually good products.
Only two moods reliably produce a review
People who received exactly what they expected have no strong reason to return and write about it. The experience closed cleanly and required nothing further.
Delight and frustration both create a motive to speak, so the two ends of the scale are overrepresented relative to how often those experiences occur.
The middle of the distribution is not missing because it did not happen. It is missing because satisfaction without surprise rarely converts into a rating.
The prompt arrives at a chosen moment
Platforms send review requests on a schedule, typically shortly after delivery, when a purchase is still novel and problems have not yet surfaced.
A product that fails after several months is reviewed only by someone motivated enough to go back and find the order, which is a smaller group.
Early enthusiasm is therefore captured systematically while later disappointment is captured only by exception, and the average reflects that asymmetry.
Sellers treat four stars as a failure
On many platforms a seller's standing depends on maintaining an average close to the top, so anything below five is treated as a problem to be resolved.
Some sellers contact reviewers to offer replacements or refunds in the hope the rating is revised, which pushes the distribution upward over time.
Buyers who understand this begin to read four stars as a real complaint, which further hollows out the middle of the scale.
Reviews attach to listings, not products
A listing can accumulate ratings across variants, editions or even different items if the seller changes what the page sells while retaining its history.
The visible count then reflects a mixture, and a strong average may rest largely on a version no longer being shipped.
Reading the most recent reviews rather than the headline number is the practical response, since recency is the only thing that anchors the score to what is currently sold.
The written text carries the information
Reviewers who describe conditions of use, what they compared against and how long they have owned an item give a reader something a number cannot.
Those details also make a review harder to fabricate, since invented ones tend toward generic praise without specifics.
The useful signal in a ratings section is the concentration of specific, consistent observations across many reviews rather than the average printed above them.