Image recognition technology, including deep learning, has developed rapidly in recent years.

The development of these methods has been significant and is becoming deeper and more widespread in practice.

Their prevalence is expected to increase rather than decrease in the future, and the impact on marketing operations is expected to be significant.

So, the question is how marketing research should use deep learning approaches.

This study considers what further research developments can be expected by combining the deep learning approach with traditional marketing research methods and concepts.

The ability to leverage the accumulated knowledge of marketing research and the deep learning approach will give us an advantage in practice.

Today, in many consumer goods markets, the physical and chemical differences between competing products are diminishing, largely due to the standardizations of products and production techniques.

In such a situation, the important role of marketers is to develop and market attractive advertising that reinforces their brand positioning.

They must also ensure that their advertising does not reinforce the brand positioning of competing brands.

To date, brand confusion experiments have been used to address this problem.

Brand confusion experiments test for brand confusion using confusion matrices, which show the frequency with which each brand is guessed when a print ad is displayed, obtained by presenting consumers with print ads of competing brands and asking them to guess which brand is being advertised.

This study analyses the visual content of advertisements using a deep learning approach to predict differentiated positioning and brand confusion.

The results will show the differentiation and confusion of each brand based on each company's storytelling and will indicate which communication strategies, such as differentiation and imitation strategies, are being used by each company.