De-identified data is information from which personal details have been removed or altered so that a person is no longer reasonably identifiable. Under Australian privacy law, information that is properly de-identified is generally no longer treated as personal information.
De-identification can involve removing names and contact details, grouping data into broad categories such as age bands, truncating or replacing identifiers, and aggregating records into totals. The aim is that no one could work out who a record relates to, even when it is combined with other available data.
In people counting, the simplest approach is to collect only aggregated counts by hour or zone and never store images or identifiers. Retailers can then share store-level traffic with head office, landlords or marketing teams with limited privacy concern.
Pitfalls include assuming a removed name is enough. Small groups, rare combinations of attributes or a unique device ID can still point to an individual, a risk called re-identification. Hashing an identifier is not always sufficient. Review the method, limit who has access and reassess when new data is added. Keep a written record of how the data was de-identified, so the decision can be explained if questioned.
Using de-identified or aggregated data lets a business learn from customer behaviour while keeping risk low, and it supports good privacy practice.