Résumé en anglais

To limit the risk of failure in marketing initiatives while preserving confidentiality, businesses employ tests carried out in shopper laboratory stores. Compared with real market testing, real or virtual tests in such stores can reduce costs, and facilitate implementation and control of the environment. This study confirms that tested in ‘real’ or ‘virtual’ environments, product and brand image evaluations bear comparison. However, it reveals significant differences in behavioural indicators supporting non-substitutability of data collection methods: virtual shopping behaviour is not real shopping behaviour. Shoppers spend more time in the virtual aisle, attention to price and prompted product recall are strongly modified, and there are major differences between regular buyers (more likely to buy their usual product in a virtual environment) and infrequent buyers.