Why The Star Ratings Are All Four Point Seven

Every restaurant, driver, plumber and gadget sits somewhere between four and a half and five stars. A scale with five positions is being used as though it had two.

People review from the extremes

Leaving a review costs time and returns nothing to the person writing it, so the motivation has to come from somewhere. Delight and outrage both supply it. Mild satisfaction does not.

The resulting distribution is J-shaped, with a large pile at the top, a small pile at the bottom and almost nothing in the middle where most actual experiences live.

An average calculated across that shape is not describing a typical experience. It is describing the ratio between two groups of unusually motivated people.

The request arrives at the happiest moment

Platforms prompt for a rating immediately after delivery or completion, when the transaction has just succeeded and nothing has yet had time to go wrong.

Ask a week later, after the thing has been used properly, and the scores come out lower. The timing of the request is a design choice with a predictable effect.

Prompts are also often skipped after failed transactions, so a portion of the worst experiences never gets asked for an opinion at all.

Anything below five is treated as a complaint

In many marketplaces the operating thresholds sit remarkably high, and sellers or contractors whose averages slip risk reduced visibility or removal from the platform.

That turns a four-star rating, which in plain language means good, into a meaningful penalty. Everyone on the supply side knows it, and asks customers accordingly.

Buyers learn the convention too, and start rating generously to avoid causing harm over something minor. The scale is compressed from both ends at once.

The scale collapsed to two values

What survives is a binary. Five means fine, and anything else means something went wrong, with the intermediate numbers carrying almost no information.

This is why a four-point-two average feels alarming while a four-point-seven feels normal, despite both sitting comfortably in the upper half of the range.

Readers have calibrated to the compressed scale without being told it exists, which works until they encounter a platform that compresses it differently.

The average hides the shape

Two products with identical averages can have entirely different distributions, one with consistent middling scores and one alternating between perfect and dreadful.

The second is far riskier to buy, and the single headline number treats them as equivalent. The histogram, where it is shown, is more informative than the figure above it.

Counting reviews matters too. An average built from a handful of ratings tells you mainly that few people have bought the thing.