Recommend the next best product per user with deep learning
Who:
A leading Spanish electronics retailer
Strategy:
To maximize its product recommendations and drive users further down the funnel to a sale event, this retailer experimented with a state of the art deep learning recommendation algorithm, which it used to automatically predict the next product(s) each user was most likely to engage with.
Show this use case for:
Hypothesis:
The key to unlocking exponential revenue gains from product recommendations largely depends on a brand’s ability to immediately identify shopper intent and dynamically recommend products based on their needs in that moment and over time. Unable to accomplish this with traditional machine learning models, this retailer decided to run a 50/50 split test comparing an advanced deep learning algorithm against its existing recommendation strategy on the homepage. In just 16 days, the deep learning-based recommendations produced a 252% increase in purchases and over €290,000 in incremental revenue.
