Dataiku > Case Studies > Predictive Content Management for PagesJaunes

Predictive Content Management for PagesJaunes

Customer Company Size

Large Corporate

Region

Europe

Country

France

Product

Data Science Studio (DSS)

Tech Stack

Hadoop
Machine Learning
Statistics

Implementation Scale

Enterprise-wide Deployment

Impact Metrics

Customer Satisfaction
Productivity Improvements

Technology Category

Analytics & Modeling - Machine Learning

Applicable Industries

Telecommunications

Applicable Functions

Sales & Marketing

Services

Data Science Services

About The Customer

PagesJaunes.fr is the French equivalent of the YellowPages. It is the French leader in local advertising and information on web, mobile, and print, with annual revenue of over $1Billion. More than 80% of the French rely on PagesJaunes.fr to get information or promote their activity. The company is one of Dataiku’s core clients. PagesJaunes is constantly looking to improve customer satisfaction while also improving Category Manager’s productivity.

The Challenge

PagesJaunes.fr, the French equivalent of the YellowPages, is a leader in local advertising and information on web, mobile, and print, generating hundreds of millions of queries each year. The quality and relevance of results is a top priority for PagesJaunes. Category managers are responsible for maintaining the quality and relevance of the directory by creating the pertinent associations between terms and categories. The challenge was to improve user experience without increasing workload. The client wanted a solution that would help them measure and improve customer satisfaction, help Category Managers automatically detect and correct problematic queries, and optimize the quality of results to improve customer satisfaction.

The Solution

With Data Science Studio (DSS), PagesJaunes built an app that scores customer satisfaction. The app gathers search engine data and analyses them to isolate unsuccessful queries. An algorithm was created to compute a score for every query and is then used to predict which queries will provide unsatisfactory results for users. The algorithm is fed with qualified and enriched usage data. The technology eventually targets the engine’s failures enabling category managers to focus their operations on those failures. The solution also involved training more than ten PagesJaunes’ collaborators in Big Data technologies such as Hadoop, machine learning, and statistics with DSS.

Operational Impact

Closely monitoring and managing unsuccessful searches
Automatically detecting the most critical signals & applying the most relevant rules when interpreting a query
Targeting and fixing false query results

Quantitative Benefit