Leveraging data and machine learning to support Marketing & Risk transformation

Our mission was to support a leader in trade credit insurance in its data transformation journey, specifically seeking to:

• Reveal the most relevant and impactful use cases across the company, identifying opportunities for different departments. This had to be in line with short and longer term goals‍
• Structure the data science approach, making sure we bridge cutting edge data science ’methodology with business line objectives. Working with business lines and maximizing impact for risk and marketing department in particular was crucial to overall success
• Driven by an ROI analysis, identify and develop the right use cases for both the risk & marketing function

We leveraged a proprietary method called a “flash diagnosis”, to gather the right stakeholders around the table and uncover the relevant use case for each business function. This enabled us to select the most emblematic and ROI driven use cases, and formalize a true data roadmap for both the risk management department and the sales and marketing teams.We then deep dived in the development of these initial use cases. From this the biggest opportunities to deliver impact were identified in fraud detection, risk model optimization, client acquisition and retention.To progress these initiatives we worked to identify the prerequisites for initial execution and further industrialization, evaluating the existing data ecosystem (IT legacy, existing technological assets, data maturity among the teams, etc.) and performing a gap analysis.While we always prefer to leverage and evolve existing tools and platforms, we still needed to build new internal capabilities to ensure we had a proper foundation for optimizing these capabilities at scale. Specifically, we had to set up a Datalab for the whole group to achieve greater efficiency in their operations.

Data and analytics capabilities can be seen as a prerequisite for data projects, but also as a lasting asset aiming to optimize efficiency and profit across the business. In the context of a Data transformation project for an Insurance company, we pushed and proactively organized members of the Executive committee to collaborate on the structure of the Group’s data capabilities. This integrated a strategic viewpoint that  had considerable impact on ensuring the current data roadmap was aligned with the company’s long term goals.There were a lot of inefficiencies in the lead management flow due to lack of data literacy, unadapted processes and a lack of visibility and anticipation on customer behaviors etc. To address these weaknesses, we first delivered a solution that improved lead & relationship management with clients, before refocusing efforts on the retention of the riskiest clients (the ones with the biggest propensity to churn). These solutions changed the way the insurance teams worked, adding a layer of data driven insights to their daily routine. This has enabled a shift from a “one size fits all” approach, to a more relevant and segmented approach that highlights most risky clients. Teams can therefore adapt their efforts in order to maximize impact.To make sure this data-driven approach was democratized among business lines and insurance experts, we deployed a churn detection and prevention tool based on sequence modeling. This has simplified interpretation, readability and access to information, drastically improving adoption across the business.Thanks to our holistic approach to performance, which seeks to embed in-depth insurance expertise, clients’ processes, customer knowledge and a better understanding of behaviors, we have succeeded in increasing accuracy of the risk model by +25%.Our business first approach to data solutions provides us a unique ability to dialogue with all stakeholders, from top management to business end-users and data analysts.