Context

Netquest is a company specialising in data collection for the market research industry. They offer a 360-degree view of consumers by collecting and analysing behavioural and declarative data from a single source: their panellists, when they browse marketplaces.

They currently identify the different pages of the purchase funnel (Product, Cart, check-out…) by rules, which hinders scalability and generalisability.

Objectives

• Improve the efficiency, accuracy and scalability in the identification and classification of the different pages of the funnel of the buying process of users in marketplaces.
• Optimise the classification process to facilitate the extraction of information at later stages, so that only relevant information is passed on to the next stages of analysis. This allows you to provide your clients with more detailed insights, contributing to a better understanding of consumer behaviour and more informed decision making.

How have we done it?

• NLP
• DocumentClassification
• TDiDF
• ML Classifier
• Robustness
• Scalability

Results

• Reduction of the development time of a model for a new channel from 8h to 30 min.
• Improved accuracy of content classification.
• Improved scalability of the classification process.