The Forefront of Marketing Thought and Practice

The accelerating pace of change in marketing means it is vital for marketers to ensure they have the latest tools and capabilities needed to face the challenges ahead. A core component of this is the increased use of data and analytics to drive decision making and customer experiences. The Marketing Science Institute (MSI) has existed since 1961 to bring the best of science to the complex world of marketing. It funds academic research and promotes its practical application with the aim of advancing marketing knowledge. I have the honour of serving on the board of the MSI and working with the top academics and business leaders in the marketing science space. What spurred me to write this post though, is the enormous practical value that MSI research offers marketers, helping them to be at the forefront of marketing thought and practice.

I store up the latest reports for the occasional quiet Sunday morning when I am up before my family. The highlights from this morning's batch were retail focused examples:

• A study by Vilma Todri, Anindya Ghose and Param Vir Singh on the trade-offs between effective and annoying display advertising. Using a Hidden Markov Model on data from an online retailer, they examine the structural dynamics of advertising effects by allowing the effects to be contingent on the latent state of the purchase funnel in which each consumer resides. They find that consumers at different stages of the purchase funnel exhibit widely different tolerance for annoyance. For example, the threshold of annoyance in display-advertising exposures is about two times higher for consumers who reside in the Interest state compared to those in the Awareness state of the funnel path. As brands strive for hyper-personalised customer engagement, it's important to understand the triggers of annoyance and to cap frequency and tailor ads accordingly. In my experience though, most brands are unable to do this routinely today.
• A fascinating report by Daria Dzyabura, Siham El Kihal and Marat Ibragimov which explores the use of deep learning-based image analytics to improve the prediction of return rates. They used a dataset of over 1.8 million online and offline transactions for 10,000 fashion products over a four year period. With the additional image-based attributes extracted, the team were able to increase the accuracy of returns prediction by 37% over non-image based attribute predictions (such as description, price, category etc.). Given the enormous cost of returns in the online channel for fashion retailers, this additional accuracy could be worth tens or hundreds of millions of dollars to larger brands. Surely then an area where fashion retailers should be experimenting and leveraging similar techniques in order to create their own tailored machine learning to reduce returns at scale.

Benjamin Franklin said "An investment in knowledge pays the best interest". In marketing, I can think of few better investments than the knowledge, insights, networking and learning opportunity afforded by the MSI.

Links

• MSI
• Advertising effectiveness and annoyance dynamics across the purchase funnel
• Leveraging the power of images in predicting product return rate

These opinions are my own and not the views of my employer.