The Plant-to-Patient Matrix: How MSD Saved Three Million Vaccine Doses

When you are distributing life-saving vaccines and oncology treatments across the globe, “every dose matters” cannot just be a corporate slogan. It must be a mathematically enforceable reality.

MSD operates at the absolute forefront of global healthcare, but their commercial supply chain was fighting the exact same physical friction that plagues every major enterprise. The transportation and logistics network was highly variable. In-transit disruptions were causing cascading receipt delays. Most dangerously, they were operating within a disparate ecosystem of trading partners, legacy systems, and fragmented data that made it impossible to trace performance or diagnose root causes in real-time.

When your data is siloed, you cannot predict a failure. You can only react to a destroyed payload.

MSD refused to accept this vulnerability. They partnered with TransVoyant to deploy an intelligence architecture capable of fusing this chaotic ecosystem into a single, predictive, plant-to-patient control tower.

Abstracting Chaos: The Device-Agnostic Advantage

To achieve absolute product integrity, you must fuse order data, shipment tracking, customer requirements, business priorities, and product handling units into one continuous mathematical stream.

The legacy approach to this is attempting to force every trading partner onto a single proprietary tracking device or software system. That is an impossible, never-ending IT and process integration nightmare.

The TransVoyant Continuous Decision Intelligence (CDI™) platform completely bypassed this friction. We deployed a device-agnostic data ingestion and normalization layer. Through advanced APIs, the platform seamlessly captured data streams directly from MSD’s ERP systems, disparate trading partner clouds, and highly fragmented IoT temperature devices, fusing it all alongside live external global event data (weather, port congestion, flight delays, etc.).

This created an unbroken, highly auditable “single source of truth.”

The Execution: Three Million Doses

Visibility alone does not save vaccines. The platform had to weaponize this newly fused living data lake to trigger autonomic interdictions.

We engineered the platform to execute automated exception management based on three intersecting physical constraints:

1. Real-Time Location & Delay: Exactly where is the payload, and what is the precise predicted time remaining until delivery?
2. Product Stability Data: What is the exact thermal degradation curve of the specific SKU inside the container?
3. Time-Out-of-Range (TOR): How long has the payload been exposed to a thermal anomaly, and when exactly will it reach total failure?

By continuously calculating the physics of these three variables, the platform shifted MSD from reactive tracking to predictive defense.

The results were absolute. As David Komjati, MSD’s Director of CoE Transportation and Distribution, confirmed: “We’ve avoided discards of over three million vaccine doses through visibility, insights, and actions taken on our international shipments.”

But scaling this architecture across disparate global regions requires absolute data uniformity. Frederic Brut, Head of Supply Chain EEMEA & APAC at MSD, validated the platform’s ability to execute this across highly volatile international corridors, pointing to TransVoyant as the “key partner for us to drive data integration, return shipment visibility, lane risk assessment, and predictive analysis.”

The Autonomic Future of the Cold Chain

Three million saved doses are not the result of a better dashboard; it is the result of superior supply chain architecture.

By driving data integration, outbound and return shipment visibility, and executing predictive lane risk assessments, MSD fundamentally altered the physics of their network. They eliminated the variable costs of reactionary service levels and secured the uninterrupted supply of critical medicines to the patients who need them most.

In the modern pharmaceutical supply chain, you are either operating a predictive network, or you are writing off lost payloads and degrading patient outcomes. There is no middle ground.