On a Tuesday morning in March, an operations director for a mid-sized consumer goods company gets a message: her primary contract manufacturer outside Ho Chi Minh City has gone dark for 72 hours after a regional power failure. Five years ago, this would have meant a war room, an all-hands call with the CEO, and a scramble to find capacity anywhere that would take the order. Instead, she opens a dashboard, sees that a qualified backup supplier in Querétaro already carries 30% of that SKU's volume, reroutes the order in eleven minutes, and gets back to her actual job. The customer never learns anything happened. Margin barely moves.
That is not luck. It is design. And it points to the gap between how most companies talk about supply chain resilience and how the operators who actually deliver it think about the problem. The talk is about redundancy — more suppliers, more inventory, more optionality, everywhere, all the time. The design is about selectivity — knowing exactly where disruption would actually hurt, and building real optionality only there, so the rest of the network can stay lean. Resilience that is not selective is not resilience. It is just cost with a good excuse.
The Three Places Resilience Actually Pays for Itself
Most supply chain strategies fail the unit-economics test because they treat every SKU, every supplier, and every lane as equally worth protecting. They are not. The operators who get this right concentrate their investment in three specific places, and leave the rest of the network alone.
1. Sourcing, but only for the critical few
Dual-sourcing the entire catalog sounds prudent and is almost always a mistake — it doubles supplier management overhead and qualification cost for SKUs that were never at real risk. The better move is to rank every SKU on two axes: how much margin or revenue it actually carries, and how exposed its current supply base is to disruption. The top-right quadrant — high impact, high exposure — is usually 10 to 15% of the catalog. That is where a second qualified source earns its cost. Everywhere else, single-sourcing at the best price is still the right call. In a recent engagement, narrowing dual-sourcing to that critical tier cut new qualification spend by more than half while covering over 80% of the revenue that was actually at risk.
2. Inventory, positioned by exposure, not by history
Most safety stock policies are still set off trailing sales history, which protects against ordinary demand noise but does almost nothing for a real disruption. Real resilience means repositioning a portion of that inventory dollar-for-dollar toward the chokepoints in the network — the single-source component, the sole-port lane, the supplier operating at capacity — regardless of how fast that specific item historically turned. The total inventory investment does not need to grow. It needs to move to where a stockout would actually cost something.
3. Visibility, built as a trigger, not a dashboard
Most companies now have some version of a control tower — a screen showing where every container, truck, and PO stands. Very few have translated that visibility into a pre-approved response. A dashboard that shows a disruption forming but requires a meeting to decide what to do about it has not bought any resilience; it has bought a faster way to watch a problem happen. The operators who move fast have already written the playbook: if this supplier goes dark, this backup activates, this person has the authority to approve it, and this is the cost ceiling before it needs escalation. The value is not seeing the disruption. It is not needing a meeting to respond to it.
The Trap Most Leaders Fall Into
The most common mistake I see in supply chain strategy work is treating resilience as insurance you buy uniformly across the network — diversify every supplier, hold buffer everywhere, add redundancy as a blanket policy. It feels responsible in a board meeting. It is usually the fastest way to erode unit economics without meaningfully reducing risk, because the investment is spread evenly across a network where the actual risk is not evenly distributed. A handful of nodes carry almost all the exposure. Protecting everything equally means underprotecting the parts that matter and overpaying for the parts that do not.
The second trap follows from the first: treating resilience as a one-time project instead of an operating discipline. A network gets re-mapped for risk once, a few backup suppliers get qualified, and the initiative is declared complete. Twelve months later the supplier base has shifted, the SKU mix has changed, and the map no longer reflects where the exposure actually sits. Resilience has to be re-scored on the same cadence as the rest of the operating review, or it quietly goes stale.
I have seen both traps in the same company. A network gets diversified everywhere after one bad disruption, unit economics take the hit, and eighteen months later the team quietly lets the redundancy lapse because no one can point to what it is actually protecting — so the next disruption lands just as hard, on a network that is now both more expensive and no better defended. The fix in both cases is the same: fewer nodes protected, more deliberately, reviewed on a fixed schedule instead of after the next fire drill.
What Leaders Should Do in the Next 90 Days
If you are running a company, a division, or a portfolio company, the next 90 days should go toward precision, not toward a redundancy budget.
- Score the network. Rank every SKU or product family on impact and exposure, and be honest about where the top-right quadrant actually sits. Most leadership teams are surprised by how small it is — and by which items land there.
- Qualify a second source only for that top tier. Resist the pressure to expand the list because a supplier relationship "feels shaky." Feelings are not a segmentation criterion; impact and exposure are.
- Move inventory dollars, don't just add them. Reposition buffer stock toward the chokepoints identified in the scoring exercise, funded by trimming safety stock on items that scored low on both axes. The finance team should see this as cash-neutral, not as a new ask.
- Write the playbook before the disruption, not during it. For every node in the top-right quadrant, document the trigger, the response, the decision owner, and the cost ceiling that requires escalation. If your team cannot execute that plan without a meeting, it is not a playbook yet — it is a slide.
What I Am Seeing Now
In the work I am doing now — with a PE-backed industrial manufacturer, a global consumer distributor running 120-plus sites, and a founder-led platform scaling toward an exit — the companies pulling ahead are not the ones with the most suppliers or the deepest inventory. They are the ones that know exactly which 10% of their network actually determines whether a bad week becomes a bad quarter, and that have already decided what happens the moment it does.
None of this requires a bigger budget. It requires a narrower, more honest conversation about where the real exposure sits, and the discipline to leave the rest of the network alone. Boards do not reward companies for the redundancy they built. They reward them for the disruption that never became a headline. Resilience, done well, does not show up as a line item. It shows up as a Tuesday that never becomes a crisis.