Gousto Case Study

Gousto Is Using Graph Technology to Personalise Its Ingredient Lists

The Challenge

A central Gousto value is constantly evolving its product and service to better serve needs and deliver maximum convenience to its customers. In the past year, it has grown its menu to offer even more choice – 30 recipes weekly, across a variety of new ranges (including gluten-free and plant-based). But this presents a challenge: the more choice, the more difficult its menus may be to navigate.

Hence the need for personalisation, said Gousto Data Scientist Irene Iriarte Carretero.

“The sweet spot between convenience and choice is personalisation, so we’re committed to using technology to make our customer interactions relevant – making it easier for shoppers to find the dishes they want and offering them an excellent experience at the same time,” Iriarte Carretero said.

The Solution

Gousto uses two different data sources: first, a subscriber’s previous interactions with the menu, and second, the information it holds on upcoming recipes. This allows the firm to create a recipe similarity measure, providing its menu designers with a good indication of which recipes they think each customer will enjoy most.

Neo4j’s graph database technology is particularly adept at capturing subtle interconnections in data and has been adopted as the way to model this less-than-obvious network of associations and connections in food items.

As a result, Iriarte Carretero said the team now has a hybrid recommendations system that combines the best of both collaborative- and content-filtering approaches to finding similarities in ingredients and dishes.

Why Neo4j?

“We researched our options and Neo4j came to our attention, and we loved its functionality and interface,” said Iriarte Carretero. “Cypher was also really easy to pick up to let us start playing with the data.”

Download Case Study