Personalized Pharma Campaigns Powered by ML

Challenge

Pharmaceutical companies possess vast amounts of data, yet often there are challenges to leverage it effectively in their advertising campaigns for drugs. Campaign decisions by  advertising companies are primarily made based on prior experiences and intuition rather than data-driven insights. The challenge lies in harnessing data about doctors’ historical engagement, demographics, and affinity to create more targeted and effective campaigns.

Solution

The SmartCat team tackled this challenge by developing the NBE (Next Best Experience) model, which provides data-driven insights into activating doctors through best possible channels. This approach allows clients to prioritize channels that are most likely to engage doctors, thereby boosting conversion rates. By ranking each channel based on the probability of positive engagement, SmartCat enabled clients to make informed decisions about resource allocation. Collaborating closely with clients, SmartCat identified and utilized a unique combination of features during model training and inference, optimizing the efficacy of the solution.

Results

The implementation of the NBE model led to significant improvements in campaign performance. Moreover, a majority of doctors from target lists for multiple campaigns were successfully assigned to suitable channels, demonstrating the effectiveness of the data-driven approach.

Smart Tip:

When facing similar challenges, thoroughly analyze available data to identify valuable insights. Utilize this data to create meaningful features that enable machine learning models to discern patterns in behaviors effectively.

Smart Fact:

Studies have shown that leveraging data-driven insights in pharmaceutical advertising can lead to a 30% increase in campaign effectiveness and a 50% reduction in marketing costs.

Technologies Used:

The NBE model leveraged the XGBoost algorithm as its foundation. Development and deployment were facilitated using AWS SageMaker, with features and pipelines stored in the Feature Store. Results were stored in Snowflake, with metadata and pipelines stored on AWS cloud infrastructure.

About the Client:

A leading global provider of life sciences services, employing over 6000 professionals across multiple sectors, including pharmaceutical advertising, partnered with SmartCat to develop the AI component of their omnichannel solution, aiming to enhance the effectiveness of pharmaceutical advertising campaigns.