12-02, 16:30–17:00 (Europe/Amsterdam), Ernst-Curie
An ever increasing number of people are discovering mobile grocery shopping as an alternative to brick-and-mortar supermarkets. This talk will cover how we can use machine learning to make these customers' grocery shopping as smooth and frictionless as possible. We do this by applying ML models that rank products in agreement with the customer’s intent: e.g., by detecting personal shopping habits, and by striking a balance between query relevance and margin.
In online grocery, the wide range of available choices can easily overwhelm a customer. Moreover, failure to find the desired products may lead to customers not converting at all. It’s therefore crucial to optimise ranking, in accordance with the customer’s intent; and to construct sensible algorithms that capture this intended behaviour.
In this talk, I will provide a holistic view of how we approach ranking in the online grocery context. Depending on an app page’s intended functionality, we might aim to make rebuying as frictionless as possible, while elsewhere we personalise search query relevance while not losing sight of margin. More concretely, I will discuss how we have set up ranking in an explainable and interpretable way that allows for a balance between relevance, profit and any other business-based concerns there might be. In addition, I will briefly discuss three algorithms that we have developed and implemented, and how these are combined to optimise the customer experience:
- prediction of rebuying probabilities through detecting personal shopping habits
- construction of unbiased search term-article relevances through structural position bias corrections
- personalisation of search results while taking profitability into account
This talk will provide the application-minded Data Scientist with an inside view into the deliberations that inform our ranking algorithms and setup.
No previous knowledge expected
Bas is a data scientist at Picnic Technologies. In his role, he functions as the Merchandising chapter lead, where he is responsible for AI projects that affect the Picnic app and what customers will see there, such as promotion forecasting, ranking & personalization, and recommendations. He has a background in physics, holding a Ph.D. from the University of Groningen and previously working at the Massachusetts Institute of Technology, MPI-PKS and Ocado Technology.