Data Quality – Cancellation Predictions

Leonardo Tocci

Posted: 1 September, 2021

Challenge

Divide customers into segments in order to feed a prediction model for cancellations and to be able to efficiently focus marketing efforts.

What was our approach?

Following problems that La Nación had had with the segmentation of their database, we moved forward first with a data quality process so as to enrich the customer database. Secondly, we used the enriched data to do adequate segmentation.

In our work on data quality, we built inference models to complete information that was lacking and defined metrics to constantly monitor it.

Benefits

• Significant improvement in data quality
• Single customer detection and database cleaning
• Improved customer segmentations.

My role

Service Manager: I assisted the project team with the adoption of the necessary methodology for running the service and provided support for technical needs in the data team. I also collaborated with the technician in charge of the service on matters related to communication between the client and the rest of the teams.