A user score can collapse within hours over something unrelated to the work itself. The mechanism is simple, and the countermeasures platforms have adopted are a direct response to it.
An unweighted average is trivially moved
A plain average treats every submitted rating identically, so a coordinated group submitting the lowest possible score moves the number far more than ordinary opinion does.
Extremes have outsized leverage. A cluster of minimum ratings shifts an average further than the same number of moderate ratings ever could.
Because most people never rate anything, the participating population is small enough that a determined group can constitute a meaningful share of it.
The trigger is usually external to the work
Campaigns typically follow a controversy about casting, a business decision, a statement, or a change to something the audience valued.
The rating becomes a vote on the grievance rather than an assessment. Many participants have not seen or played the thing they are scoring.
Positive campaigns exist as well, organized in defense, which produces a distribution split between the two extremes with almost nothing in the middle.
Timing exposes the coordination
Organic ratings accumulate gradually and track release patterns. A campaign appears as a sudden spike from accounts with little prior activity.
Platforms detect this by examining account age, rating history and the interval between account creation and the rating submitted.
The signature is consistent enough that identifying a campaign is usually straightforward. Deciding what to do about it is the harder question.
Platforms responded with friction
Common measures include delaying user ratings until after release, requiring verified purchase or playtime, and limiting how quickly new accounts can rate.
Some platforms weight ratings by account history rather than removing them, which reduces the impact of a campaign without deleting anyone's opinion.
Others display the distribution alongside the average, since a bimodal shape communicates a disputed reception more honestly than a single number.
The underlying number was always fragile
Even without campaigns, self-selected ratings are not a sample of the audience. People with strong feelings rate; the indifferent majority does not.
Bombing exaggerates a distortion rather than creating one, which is why platforms increasingly present scores with context instead of trying to produce a single clean figure.
Reading the distribution and the dates is a more reliable practice than reading the average, and it costs the reader only a few seconds.