Specification governance: procurement's missing lever

Procurement owns the price. Engineering owns the spec. Finance owns the budget. Nobody owns the gap between them.

Unchecked specifications are the most expensive cost problem most organizations never address. When engineers default to a specific grade of stainless steel, a particular tolerance class, or a single-source component — and nobody in procurement has the authority or the data to challenge that choice — the cost is baked in before a single RFQ goes out.

The spec gets written before procurement sees it

In most manufacturing organizations, the specification is created during product development — by design engineers, R&D teams, or external consultants. Procurement joins the conversation after the spec is frozen. At that point, the only lever remaining is price negotiation against a fully defined requirement.

A manufacturing spend analysis from aPriori identifies product cost outliers as parts where the should-cost estimate differs from the price paid by more than 20%. These gaps routinely trace back to over-engineered specifications, not supplier margin. The tolerance class was tighter than necessary. The material grade was specified for a legacy application that no longer applies. The component was designed around a single supplier's catalog — and nobody noticed.

Performance specs versus conformance specs: the structural choice

The Chartered Institute of Procurement and Supply (CIPS) draws a sharp distinction between two types of specifications. Performance specifications define what the product must do — load capacity, operating temperature range, power consumption. Conformance specifications define how it must be made — exact materials, processes, dimensions, and tolerances.

Most organizations default to conformance specs because they feel safer. The engineer who specifies a particular grade of aluminum alloy and a particular machining process faces zero career risk if the part works. The same engineer faces real risk if a performance spec leads to a supplier innovation that fails. Procurement's job is to shift this calculus — by providing the cost data that makes over-specification visible and the sourcing alternatives that make performance specs viable.

Specification creep is invisible on a P&L

Specification management systems track two key performance indicators: change rates after approval and deviations at goods receipt. When either number trends up, specification creep is active. Each late change triggers engineering change orders (ECOs), supplier re-quoting, expedited charges, and potential inventory write-offs. None of this shows up as a line item on a procurement savings report.

"Technically, there might be nothing broken in your procurement, yet budgets creep higher, orders duplicate, and suppliers get frustrated. Without proactive change management, hidden costs quietly mount."
— Precoro

Organizations without formal specification governance see this play out as fragmented and maverick spend. Off-contract purchases dominate. Supplier counts grow unchecked. Duplicate and rogue buying create leakage without oversight. All symptoms trace back to the same root cause: nobody owns the spec end-to-end.

What a specification governance model looks like in practice

The components are not complicated. A central specification database with quality checks for completeness, clarity, and feasibility. A formal policy that binds specifiers and buyers under the same governance framework. Continuous spend analysis to detect non-standard items, duplicate suppliers, and uncontrolled variants. And most importantly: early procurement involvement in product development, before the spec is frozen.

The firms that do this well run design-to-cost gates where procurement and cost engineering jointly review specifications for cost outliers — parts where the should-cost estimate diverges from the price paid by 20% or more. These gates are not about blocking engineering decisions. They are about making the cost consequences of those decisions visible before the invoice arrives.

Gartner reports that 74% of procurement leaders say their data is not AI-ready. Specification governance depends on clean, classified spend data — without it, there is no way to systematically identify which specs are driving excess cost. The data foundation is the prerequisite. Everything else builds on it.

What this means in practice

• Build a specification database before you build a challenge process. You cannot challenge what you cannot see. A central repository with quality checks for completeness, clarity, and feasibility is the prerequisite for everything that follows.
• Run a should-cost analysis on your top 20 spend items. Flag every part where the estimated cost differs from the price paid by more than 20%. These are your specification outliers — the parts where the spec, not the supplier, is driving excess cost.
• Insert procurement into the design review, not the sourcing review. The spec gets frozen during product development. If procurement joins after that gate closes, the only remaining lever is price. Early involvement shifts the conversation from "can we get this cheaper" to "do we need this at all."
• Default to performance specs for non-critical components. If the part is not safety-critical, not brand-defining, and not subject to regulatory certification, write a performance spec. Let suppliers compete on how to achieve the outcome. Reserve conformance specs for where exact replication matters.
• Track change rates and deviations as leading indicators. Rising change rates after approval signal specification creep. Rising deviations at goods receipt signal specs that cannot be met. Both are cheaper to fix before the ECO than after.