A tier-1 automotive supplier ships stamped brackets to its customer at a Cpk of 1.67. Internal scrap is under one percent. By every conventional metric, the process is capable. Yet the customer's assembly line records a twelve-percent rejection rate on those same brackets. The discrepancy is not a measurement error. It is a system constraint that lives on the boundary between two companies.
Across two decades implementing ISO 9001, IATF 16949, and AS9100 systems in automotive and aerospace, I have audited this pattern repeatedly. Supplier scorecards show green. PPAP packages are complete. Incoming inspection passes. But the further the part travels into the customer's value chain, the more the hidden constraint costs emerge: tolerance stacks that fail at final assembly, batch variability that disrupts takt time, and emergency re-sorting that erodes both companies' margins.
The Theory of Constraints provides a diagnostic lens that most quality manuals skip. When Goldratt's five focusing steps are applied across organisational boundaries rather than within a single plant, the constraint is rarely where either party thinks it is. It is in the gap between what the supplier optimises and what the customer's system actually needs.
Where Supplier Metrics and Customer Reality Diverge
Suppliers optimise for what they are measured on. If the PPAP establishes acceptance criteria at incoming inspection, the supplier tunes its process to pass that gate. The problem is that incoming inspection criteria are almost always a subset of the functional requirements the customer's product demands. A casting may pass dimensional layout at the supplier's dock yet fail leak-testing after machining because porosity distribution—never specified in the control plan—falls outside the functional envelope.
This divergence creates a false sense of security. The supplier's SPC charts confirm stability. The customer's incoming inspection confirms compliance. Neither captures the constraint: the point where the part's actual behaviour under assembly conditions diverges from the assumed capability. In TOC terms, the constraint is not the supplier's cycle time or defect rate. It is the undocumented tolerance gap that emerges only when the part enters the customer's process.
Identifying this boundary constraint requires joint process mapping. I have led cross-functional workshops where supplier and customer engineers map the entire value chain together—from the supplier's raw material through to the customer's final test. In every case, at least one critical parameter was controlled by neither party because it fell in the organisational gap. That uncontrolled parameter is the quality constraint.
Exploiting the Constraint Without Owning the Process
Goldratt's second step—exploit the constraint—becomes politically complex when the constraint sits in someone else's factory. You cannot unilaterally install inspection points or alter cycle times. What you can do is redefine the verification boundary. If the constraint is a supplier's heat-treatment process, exploitation means moving your incoming inspection from dimensional checks to hardness and microstructure verification, ensuring defective material never enters your constrained downstream operation.
At one aerospace supplier I worked with, a sub-tier vendor supplied titanium forgings with inconsistent grain flow. The tier-1 customer's machining operation—the system constraint—wasted eighteen percent of cycle time reworking forging anomalies. Exploitation meant installing a simple acid-etch inspection station at the tier-1's receiving dock. The cost was minimal. The protection it provided to the downstream bottleneck was immediate.
The discipline here is restraint. Do not demand that the supplier improve everything. Demand conformance only on the parameters that affect the constraint. Suppliers resist blanket quality demands because they interpret them as cost-shifting. A targeted request—backed by joint root-cause data—is far harder to refuse and far cheaper for both parties to implement.
Cross-Boundary Constraint Resolution Sequence
- 01Map jointlySupplier and customer engineers trace the full value chain to locate where functional divergence occurs.
- 02Verify at the boundaryMove incoming inspection to target the constraint-driving parameter, not the full dimensional layout.
- 03Subordinate purchasingRelease orders at a pace the constrained operation can absorb without batching or expediting.
- 04Elevate collaborativelyInvest jointly in process improvement only after exploitation and subordination are exhausted.
- 05Repeat with the next constraintUpdate both parties' control plans and re-map when the constraint migrates.
Subordinating Purchasing Decisions to the Constraint
Subordination is the step most organisations violate at the supplier interface. Procurement teams are incentivised on piece price and volume discounts. They order in batches that optimise freight cost and satisfy contract minimums. But if the constrained operation downstream can only process material at a controlled rate, large batches create inventory ageing, handling damage, and batch-level variability that the system cannot absorb.
Subordinating purchasing to the constraint means restructuring the order rhythm. If the constraint processes two hundred units per shift, do not receive two thousand units in a single drop. Negotiate staggered deliveries, consignment stock, or frozen-period scheduling with the supplier. The objective is not to reduce the supplier's efficiency but to protect the constraint from the variation that bulk delivery introduces.
This step requires breaking down the functional silo between quality and procurement. In plants I have audited, the supplier quality engineer knows which incoming batches cause downstream defects. The buyer knows which delivery pattern achieves the lowest landed cost. These two people almost never share data. Subordination demands that the delivery rhythm serve the constraint, not the freight optimisation model.
The cost of not subordinating is hidden in the overhead. Emergency sort teams, expedited freight for replacement material, and the labour cost of re-inspecting suspect stock are all consequences of purchasing decisions made without reference to the constraint. These costs rarely appear on the supplier scorecard, but they erode the total system economics more than any unit-price saving procurement achieves.
Drum-Buffer-Rope Across Company Walls
Goldratt's drum-buffer-rope logic maps directly onto supplier quality management, but it requires reinterpreting each element for a two-company system. The drum is not the supplier's production rate. It is the rate at which your constrained operation can consume the supplier's material without accumulating variation. The buffer is not just physical stock. It is a time-qualified buffer of verified-conforming material that insulates the constraint from supplier-side disruption.
Improving Cpk from 1.0 to 1.33 at the constraint yields more than improving ten other operations from 1.33 to 2.0.
The rope is the information mechanism. In a single plant, the rope is a production signal—a kanban card or an electronic schedule. Across a supplier boundary, the rope is the release of purchase orders synchronised to the constraint's actual consumption rate, not a forecast-driven MRP explosion. When the rope is too long—when orders are placed far in advance based on optimistic forecasts—the supplier produces to a plan that does not reflect the constraint's real needs.
I have seen aerospace supply chains where the rope was effectively broken. The customer ordered to a twelve-week forecast. The supplier built to that forecast. By the time material arrived, engineering changes had shifted the requirements, and the buffer was full of the wrong configuration. Re-establishing the rope meant moving to a frozen two-week window with a rolling six-week visibility—a compromise that gave the supplier stability and the customer relevance.
Supplier Quality Management: Conventional vs Constraint-Focused
Conventional supplier management
- Incoming inspection against full PPAP dimensional layout
- Supplier scorecard tracks PPM, on-time delivery, and response time
- Orders placed to MRP forecast with quantity discounts
- Corrective action requested after defect escapes reach production
Constraint-focused supplier management
- Incoming inspection targets only constraint-driving parameters
- Scorecard tracks constraint consumption rate and boundary variation
- Order rhythm subordinated to constrained operation's actual pace
- Joint process mapping identifies the constraint before escapes occur
Elevating When the Constraint Is Not Yours to Fix
Elevation—Goldratt's fourth step—means investing capital to break the constraint. When the constraint is internal, the decision is straightforward: justify the expenditure, buy the equipment, train the people. When the constraint sits in a supplier's process, elevation becomes a commercial and technical negotiation. You are asking another company to invest in a process change that benefits your system.
The leverage point is data. If you can demonstrate—through joint 8D analysis and shared SPC data—that the supplier's process variation is driving quantifiable losses in your constrained operation, the business case for elevation becomes shared. At an automotive supplier I worked with, a sub-tier stamper's tool wear was the documented root cause of a downstream welding defect that cost the tier-1 customer forty thousand euros per month in rework. The elevation investment—new tool steel and a revised preventive maintenance schedule—was split between both parties and recovered in eleven weeks.
The alternative to collaborative elevation is supplier replacement, which most organisations default to prematurely. Replacement is slow, expensive, and introduces a new qualification cycle under PPAP or first-article inspection. If the current supplier's process is fundamentally capable but constrained by a specific, addressable limitation, elevation through joint investment is almost always faster and cheaper than re-sourcing.
The discipline is to elevate only after exploitation and subordination are exhausted. Many companies jump straight to capital investment—or supplier replacement—without first verifying that the constraint is protected by adequate boundary inspection or that purchasing rhythms have been subordinated. Elevation is the most expensive step. It should also be the last resort.
Why Constraint Inertia Strikes the Supplier Interface Hardest
Goldratt's fifth step warns against inertia. When a constraint is resolved, the system shifts, and a new constraint emerges. At the supplier interface, inertia is compounded by contractual rigidity. Quality agreements, control plans, and inspection instructions are written around the old constraint. When the constraint migrates—either because elevation succeeded or because the product mix changed—these documents are rarely updated in time.
I have audited plants where the supplier quality agreement still mandated one hundred-percent incoming inspection on a parameter that the supplier had eliminated through process redesign eighteen months earlier. The inspection consumed three full-time inspectors testing a parameter that no longer varied. Meanwhile, the new constraint—a different supplier's material consistency—had no verification coverage at all.
The remedy is a scheduled constraint review. At least quarterly, supplier quality engineers and production planners should jointly re-examine where the current quality constraint sits. If it has moved, update the quality agreement, re-target incoming inspection, and reallocate SPC resources. This is not administrative housekeeping. It is the mechanism that prevents the supplier quality system from optimising around a constraint that no longer exists.
Key Thresholds for Cross-Boundary Constraint Management
Building the Cross-Organisational Discipline
The plants that manage supplier constraints effectively share a common trait: they treat the supplier interface as a process, not a transaction. The boundary between two companies has a flow rate, a defect rate, and a constraint—just like any internal operation. When quality engineers apply the same rigour to that boundary that they apply to their own machining centres, the hidden costs collapse.
This means abandoning the fiction that supplier quality is the supplier's problem. The constraint may originate in the supplier's process, but its consequences are felt in the customer's system. The customer owns the consequence and therefore owns the responsibility to identify, protect, and help resolve the constraint. Suppliers who recognise this shared ownership become extensions of the customer's quality system. Those who do not remain vendors, perpetually managed by scorecards that measure everything except the thing that matters.
TOC does not replace Lean or Six Sigma at the supplier interface. Lean eliminates the waste of excessive WIP and unnecessary handling. Six Sigma reduces the variation that drives boundary defects. TOC tells you where to point those tools—which supplier, which parameter, which delivery pattern—so that the improvement effort translates into system-level quality gains rather than local optimisation that changes nothing the customer feels.