Glossary

What is data normalisation?

In product-data operations, data normalisation means bringing incoming values into a consistent form that can be processed and compared. It can include parsing numbers, converting units, tidying formats and applying agreed value transformations. The term overlaps with standardisation in practice. This use concerns product values and should be distinguished from relational database normalisation or statistical scaling.

Prepare values without losing their meaning

A normalisation process makes recognised source representations usable in the receiving structure. It may separate a measurement from its unit, remove unintended whitespace or interpret a documented date format. Preserve the original value or a traceable reference where a transformation could need review.

An unknown abbreviation should remain an exception until its meaning is established. Changing it to the nearest familiar option can make the data look tidy while altering the underlying specification.

Example: product lengths from two suppliers

One supplier sends "1.2 m" and another sends "1200mm" for equivalent lengths. A catalogue using millimetres can store both as the numeric value 1200 with the defined unit. A value such as "approx. 1.2 m" carries an additional qualification that should not disappear during parsing.

The team also distinguishes product length from package length. Unit conversion is useful only after the attributes have been mapped to the same meaning.

Normalisation and standardisation

Standardisation establishes or applies an agreed convention; normalisation commonly describes the processing that brings values into a usable common form. Tools may use these names differently or combine the activities. Define the actual rules in the workflow. A successful transformation does not verify that the source measurement is accurate.

Related terms

Sources and further reading

Product Data Services

Explore practical support for organising, improving and maintaining product information.