Vehicle assembly lines are stopping this month because of a component that costs a few dollars. Volkswagen warned about it before Christmas; in the first days of January the announcements have arrived in sequence, with plants cutting shifts and model variants being deprioritised, and the cause is a semiconductor shortage that nobody in the automotive industry saw as a serious risk twelve months ago. The interesting part is not the shortage itself. It is that almost no purchasing organisation can answer, quickly and from its own systems, the one question the shortage asks: where are we concentrated, and what stops if that concentration fails?
How a few dollars stops a factory
The mechanics are unglamorous. When demand collapsed in the spring, manufacturers cancelled component orders. Foundries reallocated that capacity to customers who were buying heavily, which in 2020 meant laptops, tablets, displays, networking equipment and games consoles. Automotive chips are largely made on older, cheaper process nodes, and older capacity is not where anybody has been investing. When vehicle demand recovered faster than expected in the autumn, the capacity had been sold, and the lead time to get it back is measured in quarters, not weeks. There is no meaningful spot market for a qualified automotive-grade part. So a supply chain that was engineered for cost efficiency has discovered that its most critical dependency was a category of part nobody senior had ever discussed.
Three questions your system probably cannot answer
Try these against your own environment and time the answers. Which finished items or projects stop if this one supplier stops? Most enterprise systems can produce a list of purchase orders by vendor. Far fewer can traverse the bill of materials upward and return the revenue that depends on a single item. What share of our volume on this item sits with one source? The vendor master knows who you pay. Concentration is a property of the item, not of the vendor, and in most implementations there is simply no field where it lives. What is the real lead time, as opposed to the number in the item master? The planning lead time was typically set during implementation, from whatever the supplier said at the time, and has been quietly wrong for years. Every safety-stock calculation downstream inherits that error. If those answers take a week of spreadsheet work, the shortage has already told you something more useful than any market forecast.
Why concentration hides so well
The vendor master is an accounting artefact. It exists to pay invoices, apply tax treatment and age balances. It answers "who did we buy from" with precision and "what do we depend on" not at all. Worse, concentration is usually invisible at the tier you can see. A category manager who has dutifully qualified two suppliers reports the item as dual-sourced. Both suppliers buy the same die from the same fab. The risk was never at tier one, and the only way to discover that is to ask the question explicitly, supplier by supplier, for the parts that matter. The second hiding place is qualification time. An approved alternate exists, but it lives in an engineering document, and moving production to it requires sample approval, testing, possibly a customer or regulatory sign-off, and between six weeks and nine months of calendar time. An alternate-supplier field without a requalification-time field beside it is worse than no field, because it creates the impression of an option that cannot be exercised inside the horizon of the crisis.
The three fields worth adding before you do anything else
This is a data exercise before it is a sourcing exercise, and it is a small one. A single-source flag at item level, maintained by the category owner, not derived from purchase history. Purchase history tells you where you bought last year, which is not the same as where you could buy. A qualified-alternate record with a requalification duration. Name the alternate and state honestly how many weeks until first good production. If the answer is "unknown", that is itself the finding. Plant-level origin, not supplier country. The vendor address is frequently a sales office in another jurisdiction entirely. What matters is which facility makes the thing, because facilities are what flood, burn, lose power and get allocated. These three fields, populated for the items that carry most of your revenue, would let you answer in an afternoon what is currently taking companies a month.
| Record | Evidence to connect | Decision it supports |
|---|---|---|
| Single-source status | Item-level capability, not only last year's purchases | Identify where an interruption has no ready alternative |
| Qualified alternate | Approval state and honest requalification duration | Check whether the option is usable in time |
| Plant-level origin | Actual production site and import records | See dependencies hidden behind vendor addresses |
| Observed lead time | Receipt history against the planning master | Revisit the assumptions used for stock planning |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
Allocation rewards the organisation with clean data
Here is the part that will surprise finance. When a supplier moves to allocation, it distributes scarce parts according to a mixture of relationship, volume history and confidence. The customer who arrives with a credible twelve-month forecast, a firm commitment and consistent historical accuracy gets served ahead of the customer who arrives with an urgent phone call and a revised number every fortnight. Forecast accuracy has always been discussed as a working-capital lever. In an allocated market it becomes an access lever. The organisations that will get parts this year are not necessarily the largest; they are the ones whose demand signal the supplier believes.
Practical Guidance for Supply Risk Data Review
- Rank items by revenue at risk, not by spend. Spend tells you where the money goes; revenue at risk tells you what stops. The cheap parts are the dangerous ones.
- Take the top fifty items and ask each supplier one question in writing: who makes the critical sub-component, and where? Record the answer in a field, not an email.
- Re-baseline lead times against the last six months of actual receipts. Then fix the master data rather than adding buffer on top of a wrong number.
- Record requalification time next to every named alternate. An unexercisable option is not a mitigation.
- Store plant-level origin, and reconcile it against your customs declarations, which usually hold the truth your item master does not.
- Give the demand plan to your critical suppliers and keep it stable. In allocation, forecast credibility buys availability.
- Decide now which customers or products get the parts if you must choose. That decision is easier to make in January than under pressure in March.
- Report concentration to the board as a number. The percentage of revenue dependent on single-sourced items is a governance metric, and almost nobody has it.
The Regional Angle
Three features of this market change the calculation. First, dual-sourcing is often not legally available here. Many product categories can be purchased in a Gulf state only through an appointed agent or exclusive distributor. The concentration is therefore structural rather than commercial: there is one authorised channel in the country, and going around it by importing through another jurisdiction can breach the agency arrangement, void the warranty, or leave you holding equipment that the local service network will not support. When your risk register says "single source", establish whether that is a procurement choice or a distribution right, because the mitigations are completely different. Where it is a distribution right, the lever is the supplier relationship and stock commitments, not a second vendor. Second, the origin data you need already exists in this region, just not in your system. Customs declarations and certificates of origin state where goods were actually manufactured, and the clearing agent holds that history. Meanwhile the item master records the supplier's regional sales entity in a European or Singaporean address. Anyone trying to map plant-level exposure quickly should start with the last twelve months of import declarations rather than with a supplier questionnaire. Third, this region's instinctive hedge is buffer stock, and it is about to look like foresight. Family-owned groups here have carried inventory for decades as a matter of commercial philosophy, absorbing the carrying cost that Western procurement doctrine spent twenty years eliminating. The problem is that nobody can see it. The buffer sits across multiple entities, several warehouses, a free-zone facility and a mainland store, in systems that do not consolidate, so the group holds a hedge it cannot count. Before ordering more, count what you already own at group level. In a number of cases the answer will change the order. A fourth point deserves a sentence. Where group procurement has standardised on one global vendor to obtain a discount, the in-region concentration is not only in parts but in the single local partner able to install, commission and service the equipment. When that partner has a staffing problem, you have an availability problem, and no amount of component supply fixes it.
The objection worth taking seriously
The honest objection is that concentration is usually the right answer. Single-sourcing lowers unit cost, stabilises quality, reduces engineering and qualification effort, simplifies logistics and strengthens the relationship that gets you help when you need it. Dual-sourcing every significant part, maintained continuously, would have destroyed far more value over the past decade than this shortage is going to cost most companies. And no dataset would have prevented what happened: visibility does not manufacture wafers. Both points are correct, and any response that treats resilience as an unpriced good deserves the scepticism it will receive from finance. But the claim here is narrower. Better concentration data does not create supply; it changes the order in which you act and the honesty of what you promise. It tells you which items to commit to early, which customers to warn in January rather than April, which redesign to fund now, and where a modest amount of inventory is genuinely worth its carrying cost. It converts a general anxiety into a ranked list. The cost of three fields and fifty phone calls is trivial against the cost of discovering your dependency when the line stops, and the organisations that will look competent this year are the ones that did the ranking before the shortage made it urgent.
Common Questions
Is this only an automotive problem?
No. Automotive is where it surfaced first because of long qualification cycles and lean ordering, but anything containing a controller, a display, a sensor or a communications module draws on the same capacity. Industrial equipment, appliances, medical devices, networking hardware and anything described as connected are exposed.
Should we place long-term orders now?
For genuinely critical items, non-cancellable commitments are becoming the price of supply. Make that a deliberate decision with a named owner and a stated view on demand, rather than a panic reaction, and accept that some of it will turn out to be excess.
Where should a small operation start, without a data programme?
One spreadsheet, fifty items, four columns: what it is, who makes it, where, and what stops without it. The exercise is valuable precisely because it is crude and finishable.
What should we expect over the next twelve months?
Expect the shortage to persist well past the first quarter, because adding trailing-node capacity takes longer than the current commentary implies and the consumer electronics demand that displaced automotive orders has not gone away. Expect allocation to spread from semiconductors into resins, packaging, freight capacity and anything else where 2020 order cancellations collided with a faster recovery. Expect customers and insurers to start asking for multi-tier origin disclosure as a condition of business. And expect at least one board in your industry to ask, this quarter, for a single number describing single-source exposure, which is why it is worth being able to produce it.
Supply Risk Data Review — we turn your item master, receipt history and import records into a ranked concentration picture, so the next shortage finds you with a list instead of a spreadsheet exercise.