GrowthLoop brings AI decisioning to biopharmaceutical marketing
GrowthLoop and EVERSANA are expanding GrowthLoop’s Composable AI Decisioning solution into biopharmaceutical marketing, using continuous experimentation to guide when and how healthcare professionals are engaged. The companies say the approach connects customer context and causal evidence to subsequent campaign decisions.
- GrowthLoop and EVERSANA expanded Composable AI Decisioning into biopharmaceutical marketing.
- Morning sends produced a 13% higher weighted composite engagement score than afternoon sends in one campaign.
- Display advertising delivered a 52% higher score per user than email for peer-to-peer audiences.
- The system combines digital engagement signals with prescription activity.
- Real-time decisioning using new in-session behavior is planned for a later phase.
From engagement data to causal context
GrowthLoop said its expanded solution builds on the company’s composable CDP and AI Studio. It applies continuous experimentation to identify which marketing actions improve healthcare professional engagement, then uses those findings to inform later interactions.
The companies describe this work as addressing a challenge for biopharmaceutical marketers. First-party data can show past engagement metrics, but marketers also need evidence of which actions make a difference, for which audiences, and under what circumstances. GrowthLoop and partners including EVERSANA are building a Causal Context Graph for each organization, designed to capture healthcare professional context and learnings from previous interactions.
According to GrowthLoop, the graph provides feedback for AI decisioning that is reliable and verifiable. EVERSANA Chief Innovation Officer Faruk Capan said pharmaceutical clients need to understand which marketing decisions produce a measurable difference for the metrics that matter to their businesses.
Experiments changed timing and channel choices
In one campaign, GrowthLoop’s decisioning capabilities identified opportunities to improve send timing and channel selection. The results used a weighted composite engagement score combining digital engagement and prescription activity, allowing the companies to assess clinical and digital outcomes together.
Existing predictive models indicated that audiences would respond better to afternoon sends. However, GrowthLoop said its experimentation and AI decisioning found that morning sends produced a 13% higher weighted composite engagement score than afternoon sends.
The campaign also found differences linked to audience source. Digital display advertising, including banner ads on websites or apps, delivered a 52% higher weighted composite engagement score per user than email for audiences sourced through peer-to-peer programs. Other audience sources showed no difference, suggesting that the channel change could be targeted to peer-to-peer audiences rather than applied across every group.
Learning across longer campaign journeys
Customers can evaluate campaigns with a weighted score combining email views, email engagement, and prescription activity. Prescription activity receives the greatest weight because GrowthLoop identifies it as the primary business objective, although the data can be delayed by months and provide less frequent feedback.
To address that gap, the system also uses digital engagement signals while campaigns are running. With each healthcare professional interaction, learnings can inform subsequent choices such as send timing, the interval between touchpoints, and channel selection.
The companies said the capabilities are designed for regulated industries, where marketers must follow detailed requirements for messaging cadence and data security. GrowthLoop said its warehouse-native architecture and journey orchestration tool supported the team’s goals while maintaining industry compliance.
GrowthLoop Co-Founder and Co-CEO Anthony Rotio said marketing has historically lacked causal evidence of what engages customers or changes actions. The company’s roadmap includes real-time decisioning for known and unknown audiences using context already available in the data warehouse. Incorporating new in-session behavior is planned for a later phase.
