Know what the rating describes
A product rating and a seller-service rating are not interchangeable. Before comparing two displayed averages, identify the item being reviewed, the source, and the scale. Keep a rating of a whole product family separate from feedback on one variant when that distinction is available.
An average is a summary of the records behind it. It does not explain whether customers are discussing the same product, the same period, or the same kind of experience. This guide offers a reading framework, not a ranking of named stores.
Keep the count beside the average
Schema.org provides separate properties for rating value, rating count, review count, and the highest and lowest rating values. Those distinctions help define what a displayed number means. A rating count need not mean the same thing as a count of written reviews.
Record the source’s labels faithfully. If it does not distinguish ratings from reviews, preserve that uncertainty. Do not combine differently labeled counts into a single total and then present it as a precise measure.
| Field | What to check |
|---|---|
| Item reviewed | Product, variant, store, or service |
| Rating value | The displayed average |
| Scale | Highest and lowest possible values |
| Count | Ratings, written reviews, or an unclear total |
| Observation time | When the summary was recorded |
Do not average averages without their weights
Consider an invented example: one group has an average of 5 across 2 ratings, and another has an average of 3 across 98 ratings. Simply averaging the two averages gives 4. Weighting each by its rating count gives 3.04, assuming the same scale and compatible records.
That arithmetic alone does not make it appropriate to merge sources. You still need a reason to treat their items, methods, and populations as comparable. When those conditions are unclear, report the two summaries separately rather than creating a combined number.
Look for the explanation before the conclusion
Inspect the distribution and recent written feedback where they are available and permitted to use. A single average can hide a mix of strongly positive and strongly negative experiences. A changed product, policy, or collection method may also make old and new feedback difficult to compare.
Present the evidence limits with the result. A visible review collection is not automatically representative of all buyers, and a difference in averages does not identify its cause. Use the observation to choose the next question, keeping the source records available for review.
Sources & further reading
This is a practical framework. Numerical examples are illustrative, not observations of a named retailer. Check current platform documentation before applying product-specific requirements.
Our research and corrections policy ↗