The party that causes an enterprise's loss is usually not a party it can sue, and because recovery through a supply chain runs in series, the probability of full recovery multiplies down while each intervening cap truncates what is left.
Here is an observation most executives find uncomfortable once they check it: the organisation most likely to cause this enterprise's next serious loss is one it has never contracted with, never assessed, and quite possibly cannot name.
That is not a failure of diligence but the ordinary geometry of supply. An enterprise contracts with a first tier it chose; that tier contracts with a second tier it chose. Dependency runs the whole length of the chain instantly and invisibly, while commercial reach stops at the first signature. The enterprise is exposed to twenty parties and can recover from one.
The general commercial position, which counsel will confirm for a given jurisdiction, is that a contract binds those who signed it. A party harmed by a failure two tiers down must recover from its own counterparty, who must recover from theirs. Some jurisdictions have modified this and many contracts modify it deliberately, but the default shape holds: recovery is not a claim against the cause but a sequence of claims, each dependent on the last.
Two consequences follow, and they compound. Each link is a chance to fail — a contract that does not cover this failure, a counterparty that will not pursue its own supplier, a supplier that no longer exists — so the probability of full recovery is a product of the probabilities, not an average. And every link that holds applies its own cap and exclusions, so the money surviving each stage is a fraction of what entered it.
The contract discipline teaches each doctrine carefully and separately: why the client cannot reach the subcontractor, why the prime remains liable regardless of which supplier failed, why the subcontractor's exposure runs only to the prime. What it never does is compose them into a number an executive could act on. The composition is left to nobody, which is why almost no enterprise has it.
The Strategic Context
Consider a hypothetical modular timber building manufacturer producing volumetric modules for apartment and school projects. Its factory is the visible asset. Its actual position sits in a chain it mostly does not see.
Engineered timber panels arrive through a local distributor from an offshore mill, which buys structural adhesive from a chemical manufacturer. A proprietary connector system comes from a single overseas supplier. Testing and certification sit with an independent body whose terms disclaim nearly everything.
Now suppose a run of panels carries a defective adhesive bond, undetected at manufacture, discovered eighteen months later when modules are installed and occupied. The remediation cost is many multiples of the panel purchase price, and the manufacturer will meet a claim from the head contractor long before it recovers anything.
Its recovery path is four links long: to the distributor, whose terms cap liability near the value of the goods and exclude consequential loss; to the offshore mill, on its own terms and in its own jurisdiction; to the adhesive maker, whose relationship is with the mill and not with anyone in this country; and, if the defect is traced to a process rather than a batch, nowhere at all. Article 27 in this collection takes up allocation across a single contract boundary. The concern here begins where that one ends: what happens when there is more than one boundary between the loss and its cause.
Where Leaders Misread the Chain
The first misreading is that the prime carries it. In one sense it does — the first-tier supplier stays liable for its subcontractors' performance, and that is precisely why enterprises accept subcontracting without much argument. But liability is not capacity. A first-tier supplier remains liable up to its own cap, out of its own balance sheet, for a failure originating in a firm three times its size and offshore. Liability that exceeds the liable party's means is a formality.
The second is that assessing tier one assesses the chain. Supplier assessment is nearly always a survey of counterparties, weighted by spend. The failures that matter are weighted by dependency, and dependency does not decay with distance. A twelve-thousand-dollar connector, sole-sourced from one plant, can stop a factory as effectively as the loss of the factory.
The third is that a written chain is a reachable chain. Enterprises accumulate collateral warranties, parent guarantees, step-in rights and flow-down clauses and file them as evidence the gap is closed. Few test what those instruments cover, whether the giver could pay, or whether anyone knows how to operate them under pressure.
The fourth is that the risk register holds this. Supplier entries name counterparties, because counterparties are what the enterprise has records of. Whether a register can be truthful at all while it doubles as an input to individual appraisal is argued in [Related article: Honest Instrument, or Performance Input?] and is not the question here. The narrower point is that the register's supplier entries stop exactly where the enterprise's contracts stop, and the cause of the loss usually sits past that line.
Reframing the Issue
Recovery is not a legal question that arises after a failure. It is a system property, fixed long before, by the shape of a chain nobody designed as a whole.
Systems in series are governed by their weakest element, not their average, and adding steps always reduces reliability. Four links at eighty per cent each do not deliver eighty per cent; they deliver about forty. Unlike an engineered series system, this one also shrinks the payload at every stage, because each link applies its cap and exclusions to whatever survived the last.
Two conclusions follow that most enterprises have not drawn. The first is that beyond two intermediate parties, full recovery should be treated as improbable regardless of how good the paperwork looks. The second is that a loss which cannot realistically be recovered has already been retained, whether or not anyone has decided to retain it. It should therefore be funded, insured, or engineered out — and choosing between those three is a real decision with a real cost, not a filing exercise.
The reframe also changes what direct reach is worth. Buying a right of direct action against a party two tiers away is not an administrative refinement. It converts a serial recovery into a parallel one, and that is the only structural change available to a buyer who cannot shorten the chain.
How Recovery Decays Along a Chain
The Links Multiply, They Do Not Add
Each link contributes two independent factors: whether a claim passes through it at all, and how much money survives if it does. Both are less than one. Applying even crude values — a strong link at 0.8, a doubtful one at 0.4, an offshore or insolvent one at 0.1 — produces a defensible expected recovery within an hour, and the answer is usually a small fraction of the loss.
Enterprises resist the arithmetic because the inputs are estimates. That objection has the logic backwards. The current position is an implicit estimate of one hundred per cent, held by nobody, tested never.
The Incentive at the Next Link Is Not Yours
The most underestimated factor is not legal at all. Your counterparty must be willing to sue its own supplier — a firm it depends on, may be tied to for years, and may value more highly than it values you.
Its rational move is often a commercial settlement that preserves that relationship at the expense of your recovery, and it does not need your consent to make it. Where the same chain also carries the enterprise's cash, that pressure runs harder still. The point generalises: at every link, the party holding your claim is optimising its own position, and yours is an input to that calculation rather than the object of it.
What the Bridging Instruments Actually Buy
Collateral warranties, direct agreements, parent guarantees and step-in rights all do one thing: they create a contractual relationship where the chain provides none. Their value is not in existing. It is in three properties, each of which can be checked in advance — what the instrument covers, whether the giver can pay, and whether the enterprise can operate it in the days when it matters.
Each of these is a purchased control with a price, usually paid in commercial terms conceded elsewhere. What such controls cost, and why the instrument authorising them contains no field for it, is the subject of [Related article: The Only Spend With No Business Case]; this article stays with reach rather than price, and asks only how far up the chain the control actually extends.
There is a further instability worth naming. The chain mapped this year is not a fixed object: a link can be sold, and the enterprise's rights sit against a legal person rather than the business it currently contains. What follows when a critical link passes to an owner the enterprise would have rejected is argued in [Related article: You Contracted With a Company, Not a Capability]. Here the chain is treated as static, which is a simplification the map should be re-run against annually.
Decision Framework
The recovery chain map is built for failures, not for suppliers, and it produces a single output per failure: the name of the last party from which money will realistically come.
Start with the three to five failures that would materially damage the enterprise. For each, trace the causal path from the origin of the failure to the enterprise, naming every legal person in between. Physical tracing, not spend analysis: follow the component, the data or the custody.
For every link, record six things: whether a contract exists at that link and covers this failure; the cap and the excluded classes; the notification window against the realistic detection time; the counterparty's ability to pay a claim of this size; the jurisdiction and whether a judgement there is worth having; and whether that party has any incentive to pursue the next link.
Then apply three tests.
The stopping-point test. Reading from the origin toward the enterprise, identify the first link that fails on any of the six. That is where recovery stops. Write the name down. If the name is your own enterprise, the loss is retained.
The retention test. Multiply the pass-through factors and apply the surviving cap. If expected recovery is below a threshold the board has set — half the loss is a defensible starting line — the shortfall is a retained exposure and must be funded, insured or designed out within the year, with the choice recorded.
The parallel-reach test. For each link scoring low on pass-through, decide explicitly whether to buy direct reach past it or to accept the retention. Where direct reach is bought, test it: name the person who would invoke it, and the week they would do so.
The map is owned by the executive accountable for the operational dependency, not by legal or procurement, and it is re-run whenever the chain changes.
From Strategy to Execution
Immediate, within ninety days. Map one failure — the one that keeps the executive team awake — end to end, and produce the stopping point. One completed map does more to change behaviour than a programme of supplier questionnaires, because it produces a name.
Medium term, two to four quarters. Map the remaining priority failures. Require first-tier suppliers to disclose the sub-tier sources on which critical items depend, and treat refusal as information about the exposure rather than a procedural obstacle. Audit the bridging instruments already held against the three properties above. Reconcile the retained shortfalls with the insurance programme in one meeting.
Long term. Design chains for reachability, not only for cost and lead time. Where two sources are comparable and one is two links closer, that proximity is worth paying for and should appear in the evaluation. Over time the enterprise should be able to state, for each material dependency, who pays when it fails.
Signals to Monitor
Critical components whose sub-tier origin nobody in the enterprise can name. First-tier suppliers that decline to disclose their sources. Collateral warranties and guarantees held for years without review, or given by entities whose accounts nobody has read. Claims your counterparty settles quickly and quietly for a small fraction, without explaining the basis. A widening gap between the time a defect is created and the time it is detected. Concentration of several distinct dependencies in one plant, one region or one jurisdiction.
Questions for the Leadership Team
- For our three worst credible supplier failures, at which named party does recovery stop?
- Which parties can materially damage us without having any contract with us at all?
- What proportion of a major loss would we realistically recover, and who calculated it?
- Where our first-tier supplier would have to sue its own supplier to make us whole, what is its commercial incentive to do so?
- Which of our bridging instruments have been tested against what they actually cover and who could pay under them?
- Which retained shortfalls have we knowingly accepted, and where is that decision recorded?
Closing Perspective
The instinct after a serious supplier failure is to look for someone to hold responsible. The chain usually supplies a defendant, and it is usually the wrong one: a first-tier firm that did not cause the loss, cannot fund it, and will spend two years passing the claim upstream while the enterprise carries the consequences.
An enterprise cannot make its chain shorter by wishing. It can know where recovery stops, and it can decide in advance what to do about each stopping point — buy reach past it, insure the gap, hold a second source, or accept the exposure with open eyes and a funded position.
What it cannot honestly do is keep treating recovery as the plan. That assumption is being made today, in every dependency this enterprise has never mapped, by nobody in particular. Naming the stopping point converts an assumption into a decision, and the decision belongs to the executive who owns the dependency, not to the lawyers who will be called after it fails.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.