The Suppliers That Quietly Decide Your Year: A Dependency Review for UK Owner-Managers

Every risk register I have been shown in the past few years shares an awkward characteristic: the names on it are the ones somebody remembered to mention. The largest spenders, the longest contracts, the suppliers whose account manager sends a Christmas card. The dependencies that actually hold the operation together rarely appear, because nobody was ever asked the right question about them.

That is not a failure of diligence. It is a failure of framing. A register built from categories and spend thresholds will reliably surface the visible, and reliably miss the quietly indispensable: the calibration laboratory that turns work around in five days, the toolmaker who still holds the drawings, the single haulier who knows how to load the awkward pallet, the software reseller whose renewal arrives by email and whose absence would take the quoting system down.

Single source, or single knowledge?

The first useful distinction is between a supplier that is the only one available and a supplier that is the only one we know how to use. In owner-managed firms the second is far more common than the first. Somebody in the business has a relationship, a mobile number and twenty years of goodwill, and the arrangement works so smoothly that it has stopped being visible as a decision.

The test is straightforward and slightly uncomfortable. Ask who, other than the incumbent, could place this order on a Monday morning. If the honest answer is that only one person knows the part number, the specification or the login, then the exposure is internal as much as external, and a dual-sourcing exercise will not fix it.

Four questions that expose the real dependency

I use a short set of questions with clients, and I try to keep the answers factual rather than reassuring. Quoted lead times are marketing; the last twelve delivery dates are evidence.

What is the actual lead time? Pull the delivery history and measure from purchase order to goods received. Note the two worst months, not the average. A supplier quoting four weeks who has twice taken eleven is a supplier with a four-week story and an eleven-week tail.

What does exit actually require? Notice periods, minimum terms, auto-renewal clauses, tooling ownership, return of drawings and the treatment of work in progress. A ninety-day notice period on a component with a six-week lead time leaves a window in which there is no lawful route to an alternative. That window belongs on the register.

Has an alternative been tested, or merely assumed? A plausible second source that has never been asked to quote is not a second source. The exercise is to pick up the telephone, request a quotation and a sample, and discover what the re-qualification would genuinely involve. Most firms find the alternative exists but takes three months longer than they hoped.

What does a day of stoppage cost? Not the lost margin on the affected order alone. Add the idle labour, the missed delivery penalties, the overtime to recover, the goodwill with the customer who was promised a date, and the management time diverted for a fortnight afterwards. That daily figure is the budget against which any buffer or dual-sourcing premium should be judged.

The arithmetic of the buffer

Holding stock is not free insurance; it is working capital converted into shelves. The discipline is to size the buffer against the measured lead-time variability rather than against a feeling. If the worst observed delay was five weeks beyond plan, then five weeks of cover is the number the conversation should start from, and the cash cost of that cover should be compared directly with the daily stoppage figure above.

Some items will fail that test, and correctly so. For a low-value consumable the answer is a stocked shelf. For a high-value, slow-moving component the answer may be a contractual commitment from the supplier, a consignment arrangement, or a frank conversation about priority in a shortage. All three are legitimate; drifting along without choosing is not.

The cost that is usually underestimated

The purchase price difference between two suppliers is the least interesting number in a switching decision. The real cost sits in re-qualification: first-article inspection, retooling, revised drawings, customer or regulatory approval, insurance notification and the management attention consumed throughout. In regulated or safety-critical sectors this can run to several months before the first conforming delivery arrives.

That is why switching cost and switching time must both be recorded. A dependency with a low switching cost and a short re-qualification is a minor concern. A dependency with a modest price and an eight-month approval cycle is a strategic risk wearing ordinary clothes.

UK frictions that lengthen every timeline

Domestic supply chains in 2026 are not insulated from border mechanics, and any dependency review that ignores this is incomplete. Goods arriving through Dover, Felixstowe and the major air freight hubs can absorb days of delay through groupage consolidation, customs declarations on the Customs Declaration Service, and the practicalities of import controls on particular categories. A supplier who imports on your behalf is a supplier with a second dependency behind them, and that second dependency deserves a line of its own.

Certification adds a further layer. Where products require UKCA or CE marking, or where the supplier holds the certification that allows you to trade — ISO 9001, a safety scheme accreditation, a sector registration — the switching timeline must include the time taken for a new supplier to demonstrate the same. Sector schemes are rarely transferable and are often the genuine reason a single source persists.

A one-page review

A workable review fits on a single sheet with one row per supplier and columns for: the part or service supplied; the proportion of revenue that depends on it; measured lead time and worst observed delay; contractual notice period; days of cover held; whether an alternative has been contacted and quoted; estimated re-qualification time; and the daily cost of stoppage. Score each row simply and sort by the product of likelihood and daily cost. The top five rows are the agenda.

The findings tend to divide into three responses. Dual-source where the item is standard and the switching cost is low. Buffer where the item is irreplaceable but storable. Contract where the item is critical and neither of the first two is practical — explicit priority, agreed notice, documented tooling ownership and a named escalation route.

None of this requires a new system or a consultant's framework. It requires an afternoon, a delivery history export and a willingness to telephone a supplier you have never used. The firms that do it rarely discover a crisis. They discover two or three dependencies they had been carrying without noticing, and they find them at a moment of their own choosing rather than the moment the supplier chooses for them.