A requisition arrives on a Tuesday. It names a supplier, quotes a price, attaches a proposal dated three weeks ago, and asks for a purchase order by Friday because the project kicks off Monday. The business has already run a selection process of its own: two conversations, one demo, and a decision made before anyone in procurement knew the requirement existed.
Procurement can still do things at this point. It can check the terms, push on payment schedule, ask for a discount, maybe trim a few percent. What it cannot do is change the specification, test the market, or walk away. Those options expired weeks ago, and nobody noticed them expiring.
This is the most expensive pattern in procurement, and it almost never shows up in a savings report. Late involvement does not produce a visible loss. It produces a smaller set of choices, and the cost of the choices that were never available is invisible by definition.
The decisions that set the price happen before the negotiation
Price is the last variable to be settled and the least important one to control. Most of what an organization will pay is determined by decisions made well before a price is discussed:
- The specification. Requirements written by a technical team, in good faith, frequently describe the incumbent’s product rather than the underlying need. When the requirement lists features only one supplier offers, the competitive field is closed before the first bid arrives.
- The scope. Whether a service is bought as a bundled program or as separable components determines whether anyone can price the parts. Bundles are harder to compare and harder to exit.
- The quantity and the duration. Term length, volume commitments and renewal structure are usually inherited from whatever was signed last time.
- The timeline. A deadline set by the business becomes a constraint on the supplier’s side of the table. Every week of runway that disappears is leverage transferred to the incumbent.
None of these are procurement decisions in most organizations, and none of them should be made without procurement in the room. A negotiator handed a single-source specification and a two-week deadline is being asked to recover value that was given away upstream.
A concrete mechanism: the specification that eliminates competition
Consider a company that needs a new maintenance management system. The facilities team, which will live with the tool, writes the requirements. They have used one product for eight years, so the requirement document describes how that product works: its module structure, its naming conventions, its particular approach to work order routing.
The document goes out to four suppliers. Three of them can meet the underlying business need, but each would meet it differently, and each has to answer “partially compliant” against a third of the line items. The incumbent scores highest because the test was built from its own answer sheet. The evaluation is honest, the process is documented, and the outcome was fixed before a single supplier saw the document.
Involving procurement early does not mean procurement writes the requirements. It means someone asks a different question about each line: is this a business outcome, or a description of how our current supplier delivers it? Rewriting the document in outcome terms typically brings all four bids into a comparable range, and comparability is what makes competition real. It also makes a should-cost model possible, because you cannot build up the cost of a requirement written as a product feature list.
What the runway is worth
The same dynamic runs on a calendar rather than a document in enterprise software renewals. Renewal terms usually include a notice window: cancel or renegotiate by a fixed date, or the agreement extends automatically on the supplier’s stated terms.
For example, take an agreement worth $2 million a year with a 90-day notice period and an uplift of 8% proposed for the coming term. If procurement learns about it 60 days before the notice date, there is no time to qualify an alternative, run a migration assessment, or build a credible case for changing suppliers. The conversation is about how much of the 8% can be talked down, and the answer is usually some of it.
If the same renewal surfaces 12 months out, the work changes entirely. Usage data can be pulled and matched to what is licensed. Unused entitlements can be identified and removed from the renewal base. A competitive alternative can actually be evaluated rather than named. The supplier knows all of this is happening, which changes what it proposes in the first place. The numbers in this example are illustrative, but the structure is not: the leverage is in the runway, and runway is only available to people who know the date.
Why it keeps happening
The usual explanation is cultural: the business does not see the value of procurement, so it goes around it. That is sometimes true, and it is rarely the whole story.
The more common cause is that nobody has defined the trigger. Involving procurement “early” is an aspiration, not a process. There is no threshold that fires, no field in the workflow that routes a requirement before a supplier is chosen, and no stage gate in the project lifecycle where commercial input is required to proceed. In the absence of a trigger, involvement depends on individual relationships, and relationships do not scale.
Capacity makes this worse. The Hackett Group’s 2026 Procurement Key Issues research reports an 8% workload increase for procurement in 2026, set against declining head count and operating budgets. A team under that kind of load has to triage, and triage selects for whatever is already urgent. Requirements that are still forming, which is exactly where the value is, look less urgent than the purchase order due Friday. So the function spends its time at the end of the process, which is where its time is worth least.
The third cause is measurement. Savings reports record what happened after procurement was involved. Nothing in the reporting captures the cost of arriving late, so the pattern generates no evidence against itself and never becomes a priority.
What early involvement actually looks like
Early involvement is a design problem, not an awareness campaign. Five things make it real:
- Define the trigger in the workflow. A threshold, a category, or a project stage gate that routes the requirement to procurement before a supplier is selected. Something in the system has to fire without anyone remembering to make it fire.
- Build the renewal calendar. Every agreement above a set value, with its notice date, auto-renewal terms and owner. Work backwards from each notice date to when the work has to start. This is the cheapest thing on the list and the fastest to pay for itself.
- Tier the response. Early involvement across every purchase is not achievable and not worth it. Define what full sourcing support, a light commercial review and a self-service channel each mean, and which spend goes to each. A team that promises everything to everyone gets routed around.
- Publish the intake path. Requesters go around procurement when they do not know how to go through it. One entry point, a stated response time, and an owner who answers.
- Measure the engagement point. Record, for each significant purchase, whether procurement was involved before the specification was written, before supplier selection, or only at contracting. Track the mix over time. It is the single most useful metric for this problem, and almost nobody collects it.
Ask four questions of your own function: What percentage of last year’s top 20 purchases reached us before a supplier was chosen? Do we have a renewal calendar with notice dates, and who checks it? Is there a stage in the capital project or IT intake process where our sign-off is required to proceed? When a requester wants to involve us early, do they know how?
The construction side of this is instructive because the timing is so visible. Working with EMA Project Management, we rebuilt the preconstruction process: standard bid templates by trade, formal scope leveling, unit price comparison and a structured buyout. The result was roughly 8% off project cost, about 10% fewer change orders, and around five days cut from bid to contract. None of that came from harder negotiation. It came from doing commercial work at the point where the scope was still being decided.
The S2V position
Late involvement is a structural problem and it responds to structural fixes: a defined trigger, a renewal calendar, a tiered service model and a metric that makes the engagement point visible. Our procurement assessments look at when procurement enters the process, not only at what it negotiated once it got there, because the value lost upstream is usually larger than the value available downstream.
The goal is not for procurement to touch more transactions. It is for procurement to be present when the decisions that set the cost are actually made, and to be genuinely useful when it gets there.