Two outlets can give the same film four stars and mean substantially different things. The number looks universal, but each publication builds its scale on assumptions that are rarely printed alongside the score.
Scales are not calibrated against each other
One publication may treat three stars as a genuine recommendation while another treats it as a polite dismissal, and neither states which convention it uses.
Some outlets use half stars, some do not, and some avoid the bottom of the range entirely because nothing they choose to review merits it.
A reader comparing two scores is therefore comparing two different instruments, and the difference is invisible unless they read a lot of both.
House policy compresses the range
Publications that review selectively rarely publish very low scores, because titles unlikely to be worth covering simply are not assigned.
That selection pushes the effective range upward, so the practical scale runs from the middle to the top rather than across the full span.
Outlets that cover everything, including titles nobody expects to be good, use the whole range and consequently look harsher by comparison.
The score is added after the writing
At many publications the critic writes the review and the number is applied afterward, sometimes with an editor's involvement, to keep the outlet internally consistent.
The text can therefore carry reservations the score does not reflect, or vice versa, since prose accommodates ambivalence and a number cannot.
Readers who take the number and skip the review lose exactly the nuance the critic spent the piece establishing.
Aggregation flattens the differences
Aggregate sites convert varied scales to a common one, which requires deciding what a given outlet's four stars is worth on a hundred-point range.
That conversion is a judgment, and it can move a review several points in either direction before it is averaged with everyone else's.
Reviews published without a score have to be assigned one or excluded, and both choices change the resulting average in ways no reader sees.
Consumer stars measure something else
Ratings left by buyers on retail platforms are not evaluations against a craft standard but reports of whether an individual experience met expectations.
Those distributions tend to pile at the extremes, since people motivated to post are often either delighted or aggrieved.
Placing a critic's score and a buyer's average side by side treats them as comparable measurements when they are answering different questions entirely.