Sanjib Das: One Global Spare vs Three Local: When to Pool Inventory (Part 4)

(Part 4 of our series with reliability engineer Sanjib Das)

In Part 1, Reliability Expert Sanjib Das showed how a transparent scoring model removes emotion from stocking.
In Part 2, we looked at why “just in case” inventory quietly drains money through shelf-life limits and preservation work.
In Part 3, we saw how the right spares shorten the consequence window of both planned and unplanned failures.

Now, in Part 4, we zoom out.

What happens when a company operates multiple sites, sometimes in different countries, sometimes right across the border, each holding their own version of “critical” spares?

Sanjib has seen something that’s both obvious and underused: You don’t always need the same spare in five places. Sometimes you only need one.

The logic behind global (or regional) spare pools

Sanjib describes it clearly, drawing from his experience in Singapore, Malaysia and Thailand:

“If you operate multiple sites, you need to determine which spares should be stocked locally and which can be managed globally.”

The principle is simple:

• High-criticality = stock locally: If a failure has major safety or production impact, local storage is essential.
• Medium-criticality = stock globally or regionally: If the consequence is moderate and the lead time predictable, you can pool.

The outcome is dramatic:

“If you take the global approach your cost will go down further.”

This is because instead of stocking three or five identical parts across sites, you hold one shared stock, sized using the same scoring model introduced in Part 1.

Why pooling works: the savings mechanism

Sanjib breaks the logic down into three main efficiency levers.

1. Medium-critical spares aren’t consumed often

If the spare is:

• expensive,
• has long lead time,
• and the underlying failure mode is not frequent,

then the probability of needing it simultaneously in multiple locations is extremely low.

“If a spare is classified as medium-critical, you should not expect the failure to occur frequently. If it does, then the assessment was wrong.”

This is a key insight: Pooling works because the assessment is correct, not because you gamble.

2. Logistics between sites are predictable

In Southeast Asia, Sanjib has seen regional pooling work between Singapore, Malaysia and Thailand.

According to his experience, distance matters less than:

• customs time
• transportation reliability
• the failure’s acceptable consequence window

If a site can absorb a few days of downtime risk without major consequence, then a shared pool makes sense.

3. Corporate-led optimisation is far more efficient

Sanjib describes the typical scenario:

“Each site should hold its high-critical spares locally. Medium-critical items, however, don’t need to be stocked at every individual site; they can be managed as a global pool across five or ten sites, and the overall cost drops significantly.”

This is not just inventory cost. It also reduces:

• preservation work
• warehouse footprint
• asset management complexity
• procurement duplication

Across five or more plants, the savings compound rapidly.

What needs to be true for pooling to work

Sanjib is explicit that pooling only works when certain conditions are met.

1. A common criticality assessment

All sites must classify parts using the same scoring logic:

“You need to understand… this is medium, that’s why get the buy-in from the management.”

If sites disagree on what is “medium”, the model collapses.

2. Management alignment

Pooling changes how responsibility is distributed. Sites need confidence that:

• global supply is monitored
• reorder points are respected
• transport times are understood
• they will not be left exposed

3. Preservation standards for stored spares

A single global part must be preserved correctly:

• dry air
• humidity control
• rotation for rotating equipment
• periodic checks

(As covered in Part 2: big motors need rotation; vessels may need nitrogen blankets.)

4. Clear rules on when to draw from the pool

Sanjib emphasises that this is about removing emotion:

• no just-in-case withdrawals
• no one-site “stockpiling”
• no hidden caches

Stocking is driven by risk, not fear.

How SPARROW enables global pooling

Effective pooling depends on two things: shared, harmonised data and consistent, cross-site scoring.
This is exactly where SPARROW’s architecture fits.

• SPARROW.Clean harmonises material masters across all plants and maps them to a master catalogue of more than 40 million industrial parts in nine languages. A maintenance manager in China and another in the Czech Republic can work on the same master data in their own language, enabling one global source of truth and revealing pooling opportunities that would otherwise stay hidden.
• SPARROW.Plan applies a consistent scoring model to every site. Criticality, lead time, failure behaviour, consequence and commonality are evaluated the same way everywhere, allowing cross-site usage, updated lead times and new failure events to automatically adjust stocking recommendations.
• SPARROW.Pool then operationalises the pooling strategy. It supports internal pooling across a company’s own plants, and provides access to an external network of verified spare parts suppliers, expanding the pool of available inventory without inflating internal stock.

Together, these tools allow organisations to implement the same disciplined approach Sanjib developed, continuously, automatically and at the scale of five, ten or twenty plants at once.

Want a real-world example? Discover how Royal Avebe uses SPARROW to consolidate spare parts across three sites and unlock pooling benefits without increasing risk.

Where we go next

We have now covered:

• Part 1 – How to remove emotion from stocking decisions
• Part 2 – The real cost of carrying the wrong inventory
• Part 3 – How good stocking shortens downtime, planned or unplanned
• Part 4 – how pooling creates cross-site efficiency

In Part 5, we take the final step:

From static spreadsheets to dynamic stocking rules that update themselves

Sanjib ends the interview by describing how he began converting his Excel formulas into code that pulls live data from SAP and Maximo:

“AI and machine learning makes our life easier… it is dynamic.”

In the final article, we explore:

• why static spreadsheets fail
• how ERP data can automatically refresh scores
• what “dynamic stocking rules” look like
• and how AI helps plants stay reliable at lower cost

👉 Part 5 — Excel Is Static. Failures Aren’t: Make Spare Decisions Dynamic.