Definition
A product attribute is one named characteristic of a product, held as a defined field with three parts. Those parts are a name, a data type that constrains what may go in the field, and a unit where the characteristic is a measurement. A product then records a value in that field — the attribute is the field itself, not the value. That distinction is what makes attributes comparable across products rather than merely descriptive of one, and it is the unit of meaning that structured product data is assembled from.
Key points
- The attribute is the field, the value is the content: Color is the attribute, graphite is one product's value for it.
- The data type constrains the value — free text, a number, a boolean, a date, or a selection from a controlled list.
- A measurement attribute carries its unit separately from its number, so 12 and inches are two fields, not one string.
- Every attribute is either required or optional for its category, and only the required ones make completeness measurable.
What does a single product attribute look like in practice?
Take a monitor. Screen size is an attribute: type number, unit inches, required for the category — one product's value for it might be 27. Panel type is a second attribute: type single-select, no unit, allowed values IPS, VA, and TN, required. VESA mount pattern is a third: type single-select, allowed values drawn from a fixed list, optional. Each is one characteristic, stored once, typed, and separately addressable. A shopper filtering for a 27-inch IPS monitor is filtering on two of those fields independently. If either one held its number and unit together as text, or held two characteristics in one string, the filter would not resolve.
Common pitfalls
- Storing a measurement as free text with the unit baked in, so values cannot be compared, sorted, or converted.
- Overloading one field with several characteristics — a single Details attribute holding color, material, and care instructions at once.
- Adding attributes without ever marking them required or optional, so no product is ever flagged as incomplete.
FAQ
What is the difference between a product attribute and a product attribute schema?
An attribute is one field. A product attribute schema is the definition layer above it — the document that declares which attributes exist for a category, their types and units, and which are required. The schema defines; the attribute is one of the things it defines.
Is an attribute the same as its value?
No. The attribute is a stable field defined once for a whole category; the value is what one product records in it. Two products share the attribute Material while holding different values. Conflating the two is why some catalogs end up with a separate field per value.
Source
The schema.org Product vocabulary publishes an expected value type alongside every property it defines, and models measurements the way described above: width, height, depth and weight each expect a typed Distance or Mass, or else a QuantitativeValue — which, as the vocabulary's own size entry puts it, is a value "with a unitCode," held separately from the number. It does not, however, govern how a catalog stores its own fields. The required-versus-optional distinction is instead shown concretely in Google's Merchant Center product data specification, which marks each field "Required," "Optional," or "It depends."