A film can be praised by critics and rejected by audiences, or the reverse, and the divergence is treated as a scandal by both sides. It is mostly a product of who is doing the rating and under what conditions.
The two groups watch different amounts
A critic sees a very large number of films each year, which makes familiarity itself a factor. A structure that feels fresh to most viewers may be the fourth version of it that month.
That volume rewards novelty and formal ambition, and it penalizes competent work that does nothing unexpected. The judgment is comparative against an unusually large sample.
A general viewer choosing one film for a Friday evening is not making that comparison, and a familiar structure executed well is precisely what they came for.
Audience scores are self-selected
People who rate a film have usually chosen to see it, which means the sample is drawn from those already inclined toward its genre, cast or subject.
Critics are assigned across everything released, including films they would never choose, so their sample includes viewers a title was never meant to reach.
A genre film with a devoted audience and cool reviews is often not a disagreement about quality at all. It is two groups answering different questions.
Viewing conditions differ substantially
Press screenings happen in the morning, in a quiet room, among people taking notes, sometimes several films in a day. That is not how anyone else watches.
An audience arrives in the evening, socially, having paid, which raises both anticipation and the cost of being disappointed.
Comedy is particularly sensitive to this, since laughter is contagious and a sparse professional screening suppresses the response the film was built to produce.
Aggregate scores hide the shape of opinion
A single number cannot distinguish a film everyone found acceptable from one half the audience loved and half rejected, though those are very different works.
Divisive films are exactly where the two aggregates diverge most, because the groups split along different lines and the averages land in different places.
Reading the distribution rather than the average resolves much of the apparent contradiction, and most platforms make that distribution available. The shape of opinion is more informative than its midpoint.
Coordinated rating adds noise
Audience scores can be pushed by organized campaigns before a film is widely seen, particularly where a title has become a subject of unrelated argument.
Platforms respond with verification requirements and delayed opening of ratings, which reduces the effect without removing it.
The practical consequence is that early audience numbers on a contested title measure attention rather than reception. They tend to settle toward something more meaningful once ordinary viewers arrive.