How the average star rating is calculated
An overall star rating is a weighted mean. You take each star level (1 through 5), multiply it by how many reviews gave that number of stars, add all five products together, and divide by the total number of reviews:
Average = (1×n1 + 2×n2 + 3×n3 + 4×n4 + 5×n5) ÷ (n1 + n2 + n3 + n4 + n5)
Here n1 is the count of 1-star reviews, n5 the count of 5-star reviews, and so on. This is exactly the formula that Amazon, Google, Yelp, Trustpilot, and the App Store use to turn a pile of individual ratings into the single number shown next to the stars.
A worked example
Suppose a product has 120 five-star, 45 four-star, 18 three-star, 7 two-star, and 10 one-star reviews. The numerator is (5×120) + (4×45) + (3×18) + (2×7) + (1×10) = 600 + 180 + 54 + 14 + 10 = 858. The total number of reviews is 120 + 45 + 18 + 7 + 10 = 200. So the average is 858 ÷ 200 = 4.29 stars. That is why the average is rarely a whole number: it is the balance point of the whole distribution, not any single rating.
Why the rating is needed
A single average condenses hundreds or thousands of opinions into one comparable figure. Shoppers scan it to decide in seconds; search engines and marketplaces use it to rank and filter products; and businesses track it over time as a satisfaction metric. Because it is a mean, it responds smoothly to new reviews — one extra 5-star review nudges a 4.29 up slightly, while a fresh 1-star review pulls it down.
Common reference points
- 4.5–5.0 — Excellent. Most listings that display "top rated" badges sit here.
- 4.0–4.4 — Strong. Widely trusted; research on marketplaces suggests conversions often peak in the 4.0–4.7 band because a perfect 5.0 can look suspicious.
- 3.0–3.9 — Mixed. Enough negative reviews to give buyers pause.
- Below 3.0 — Poor. Signals widespread dissatisfaction.
Note that a 5.0 average requires every single review to be 5 stars; a single 1-star review among 20 otherwise perfect ratings drops the average to (5×19 + 1) ÷ 20 = 96 ÷ 20 = 4.8.