When one of Everest Customer Solutions’ pharmaceutical customers faced financial challenges bringing a novel, rare-disease product to market, they realized they needed a better model than their typical launch playbook. Marketing managers determined that the financial model for this product would not be viable if they followed their normal tactics for market readiness, entry, and support. As a rare-disease treatment, there were disparate factors influencing each geography. Everest built a reusable ML model to segment key markets and enable insights and scoring of future geographic markets, optimizing the investment required in each.

The Challenge

• Our customer had a suite of trusted but expensive marketing tactics that had been identified for this flagship rare-disease product launch
• The pricing, target audience, and government and payer variables tipped the equation to a potentially canceled launch due to financial infeasibility
• The marketing and launch team needed a way to pare down the mix of tactics per geographic region to optimize marketing spend for the launch
• Extensive survey data was collected for key markets and evaluated for the anticipated benefit of each tactic

AI to the Rescue

• Everest built and tuned an Artificial Intelligence Machine Learning model based on the survey and impact data to identify patterns and segment the geographic markets.
• The team was able to prioritize tactics through these segment views and identify where specific investments made sense.
• Once validated, the model was deployed in Microsoft’s cloud-based Azure Machine Learning service and accessed from a web-based user interface.
• With this tool, marketers and field-based teams could evaluate micro-markets to estimate the required investment for launch in a specific CBSA or territory unit. 

Benefits of the AI Model

• The AI tool found patterns in a dataset without utilizing traditional, expensive “big data” that many assume they might need.
• The results and related visualizations (see radar charts) told a story from the data that group debate just couldn’t uncover.
• A small investment in the model pre-launch saved millions in potential marketing budget for this unique product launch.

A Success Story

The customer successfully refined the launch strategy, marketing tactics, and marketing budget to enable this ground-breaking product to continue through its commercial journey with the mission of changing lives for an underserved and pain-filled community.

Conclusions

AI Models are more robust and better understood than ever, and the tools to enable them have improved dramatically in the last few years.  While Generative AI — Large Language Models (LLMs) and Generative Pre-trained Transformers (GPTs) — have gotten most of the buzz lately, huge opportunities exist for supervised and unsupervised Machine Learning AI models to help you make sense of multi-dimensional data.  And while training data is often required, you don’t need every human-written word to solve serious, real-world problems!  Contact us to discuss your AI strategies and visions and we’ll help you turn them into reality.