August 2026

From One Vaccine Fridge to Five Billion Decisions

Written by

· 8 min read

From One Vaccine Fridge to Five Billion Decisions — Artwork by Dia Azzawi

Most people can picture a product, a clinic, and a forecast. What is harder to see is the planning surface created when thousands of products must be considered across hundreds or thousands of locations.

Every product–location pairing is a potential planning relationship. When active, it must reconcile demand, usable supply, lead times, variability, constraints, policies, capacity, and commitments.

We created the interactive visualization below to make that hidden scale tangible. It begins with a single vaccine refrigerator and expands through clinics, hospital networks, and national health systems. The figures show maximum planning coverage—not actual plans, calculations, or human decisions.

How much might the planning system need to cover?

The simple version is: everything you might supply, multiplied by every place you might supply it. That gives the largest planning-coverage envelope the system may need to govern.

Everything you might supply × every place you might supply it = every possible item-place relationship

For each relationship that is actually active, the system must work out what is needed, what is usable, when it can arrive, what might change, which rules apply, how much capacity exists, and what has already been promised.

The same idea, written as two equations

Maximum planning coverage: Cmax = |I| × |S|

For each active pair: Pi,s = F(Di,s, Ui,s, LTi,s, Vi,s, Ci,s, Πi,s, Ki,s, Mi,s)

I is the set of items in scope. S is the set of served or stocking locations. The planning function then evaluates demand, usable supply, lead times, variability, constraints, policies, capacity, and commitments for each active item-location pair across the planning horizon.

The coverage expands. Human workload does not have to.

The chart shows the maximum number of item-location relationships a system might need to cover. It does not represent actual stocked relationships, calculations, individual plans, or human workload.

Real supply networks are more selective and structured. Not every product is stocked at every location. Product tiers, segmentation, aggregation, and dependent replenishment reduce the number of active relationships.

But those relationships are also connected. They may share suppliers, substitutes, transfers, capacity, budgets, commitments, policies, and decision authority. These connections create the real planning complexity.

Governance does not mean asking people to approve every relationship. It means controlling the evidence, policies, calculations, authority, and exceptions used across the network.

How Vista makes this manageable

Scale the calculations. Concentrate human judgment.

Vista separates the scale of calculation from the scale of human attention. It can evaluate a large planning surface without requiring people to touch every relationship.

Define which products and locations are actually in scope.

Organize coverage into governed assortments, tiers, and planning hierarchies.

Apply shared evidence, policies, constraints, and authority across the network.

Use deterministic engines for calculations that must be reproducible.

Use AI to investigate changes, explain results, and identify meaningful exceptions.

Send consequential decisions and policy exceptions to the appropriate people.

Record evidence, calculations, recommendations, approvals, and released decisions in the Planning Record.

Monitor the released plan and surface material changes.

Some figures come from published operating data. Others are clearly identified modeling assumptions. The U.S. endpoint is a 2026 thought experiment, not a proposed system. It scales NHS England's maximum coverage envelope from approximately 59 million people to 342.3 million U.S. residents, producing about 5.4 billion possible relationships.