Procurement won the argument

Kraljic told purchasing to become strategic in 1983, and it worked — procurement owns more of the business than it ever has. The question nobody planned for is which parts it actually decided to take.

A CPO Outlook 2026 discussion piece · EBG | Network

Write down everything your procurement function is expected to cover today. Then mark the items it argued its way into.

For most functions the marks cover a good share of the page. Supplier risk. Sustainability. Innovation through the supply base. The seat in the capital expenditure meeting. Procurement wanted that ground and it went and got it, over three decades, against real resistance. This is a profession that won its argument.

Now mark the last thing that came off the list. That column tends to stay empty — not because nothing was automated or moved, plenty was, but because the expectation stays even when the task leaves. The invoice work moved to a hub and procurement still owns the outcome. The risk screening got a tool and procurement still owns the answer.

Winning ground and never conceding any is a strategy with a known endpoint. That is the shape worth looking at before anyone reaches for a target operating model.

The gap is measured, and it is widening this year

The grouping above is a reading. The imbalance underneath it is not. The Hackett Group’s 2026 Procurement Agenda and Key Issues Study projects procurement workloads rising 8% in 2026 while headcount falls 0.9% and operating budgets fall 0.4%. Technology spending, meanwhile, rises 6.1%.

Eight per cent more work against slightly fewer people is not a crisis in a single year. Repeated across a decade, it is the arithmetic behind the answer procurement leaders actually give when asked what stops them — and it explains why the technology line rising faster than the people line does not, by itself, close anything.

What the won ground costs to hold

Ask Nordic procurement leaders what stops them acting when they can already see the opportunity, and the answer is not what the sector usually talks about. In EBG’s Xperience 2026 survey — the latest in a series of Nordic procurement surveys — 62% named competing demands — capacity already absorbed. Stakeholder alignment followed at 53%, and unclear mandate at 36%. Data confidence, the thing most often blamed, came fourth at 29%, and consequence culture last at 18%.

That is a small Nordic sample — 45 responses to that question, from a survey of 48 senior professionals in 48 companies — so it is a signal rather than a proof. But it points the same way as much larger work. Deloitte’s 2025 Global CPO Survey, covering more than 250 CPOs across 40 countries, put siloed operations (57%) and competing priorities diluting focus (46%) as the top two barriers to delivering value, ahead of capability gaps (40%) and talent shortage (34%).

So the ground has a cost to hold, and it is not primarily an information problem. Which makes the next question more interesting than the usual one.

Not “what should we stop” — but “how did we get it?”

The reflex when a function is stretched is to run a prioritisation exercise. That treats the accumulation as an accident. It plainly wasn’t: each domain arrived through a mechanism, and the mechanism — not the item — is what decides whether it can ever be handed back.

There is no research that settles this for procurement specifically. There are, however, at least four readings worth putting to a room of peers, each with something behind it.

1. The profession asked for it

The oldest reading is also the proudest. When Peter Kraljic wrote “Purchasing Must Become Supply Management” in Harvard Business Review in September 1983, the argument was explicitly that the function should take on more: more risk analysis, more market intelligence, more strategic weight. Forty years of professional development has been built on that invitation. Much of what sits in the upper bands of the chart above is the profession collecting on that argument. Nobody handed it over; procurement made the case for decades and the case was accepted.

2. Nobody else was going to

A second reading is structural rather than ambitious. Some work lands in procurement because procurement is simply where the supplier relationships are, and there is nowhere else for it to go. European due diligence law makes this concrete: the final CSDDD agreed in December 2025 scopes the risk analysis to business partners across the chain of activities — which puts the work where the contracts and the supplier data already live.

Henkel’s Bertrand Conqueret put the same point plainly in a World Economic Forum piece in January 2025: “Procurement is leading sustainability into the upstream part of the value chain. No one else is doing it.” Read generously, that is a description of a function absorbing orphaned work because it is the only place the work fits.

3. It was claimed — the seat at the table

The third reading is the uncomfortable one, and it deserves to be held as a hypothesis rather than an accusation. Procurement has spent decades arguing for relevance, and taking on visible, board-level agendas is one way a function earns standing. Organisational theory has a name for the pattern, if not the proof: DiMaggio and Powell’s 1983 work on institutional isomorphism describes how organisations adopt practices that peers have adopted, in pursuit of legitimacy, rather than because those practices have been shown to work better. If that mechanism is at play, some of this ground was neither delegated nor orphaned. It was volunteered — which, usefully, means it can also be given back.

4. It was never actually decided

And the fourth reading is the least dramatic and possibly the most common: nobody decided anything. Work appeared, someone competent picked it up, and it stayed. The Xperience finding that 36% name an unclear mandate — who has authority? — as a reason action slows is at least consistent with a function holding work that was never formally granted to it. You cannot hand back what you were never given.

Which is exactly why the question is worth an hour of a room’s attention rather than a slide. Most functions will find all four mechanisms present across different parts of their portfolio — and will find that they have been applying one remedy to all of them.

The part that shows up in your team, not your org chart

There is one place where the research is firmer, and it is about people rather than functions. Role overload — being asked to do more than time allows — has been studied for decades. Tang and Vandenberghe, writing in Frontiers in Psychology in 2021, tested the mechanism across two employee samples (N=212 and N=191) and found that role overload undermines performance by way of psychological strain: the load produces strain, and the strain — not the load directly — is what degrades the work.

Their more useful finding is what buffers it. High-quality leader–member exchange — a working relationship of trust and reciprocity between a person and their manager — substantially reduced, and in places suppressed, the effect of overload on strain. Role ambiguity and role conflict did not respond the same way; overload did.

Two caveats, stated plainly: those samples were customer-service employees in Canada, not procurement professionals in the Nordics, and one study is not a settled literature. But if it travels even partly, the implication cuts against the usual instinct. When a team is carrying too much, the first-order intervention is not another tool and not another restructuring. It is the quality of the relationship between the people carrying the load and the person leading them.

Can the function keep taking ground?

That is the honest form of the forward question, and it is not rhetorical. A function can keep taking on more indefinitely, provided capacity genuinely arrives with it. The current wave’s promise is that AI supplies that capacity.

The evidence so far suggests the capacity is real but does not aggregate on its own. Gartner, surveying 101 chief procurement officers in January and February 2026, found only 36% very confident in their ability to redesign roles and processes for AI, and made the sharper point that productivity gains remain confined to the individual level unless the operating model changes around them. EBG’s own numbers describe the same expectation curve: roughly 29% of respondents save four or more hours a week with AI today, and about 67% expect to within two years.

The reflex, then, is to redesign. That is worth a caution — and the caution is about people, not plans. Gartner research reported in Harvard Business Review found that the average employee went through ten planned enterprise changes in 2022, up from two in 2016, while willingness to support enterprise change fell from 74% to 43% across the same period.

Note the shape of that: it is the chart at the top of this piece again, in a different unit. Changes accumulated fivefold; the appetite to absorb them roughly halved. Reorganising a function whose people have already been reorganised repeatedly is not a neutral act, which argues for being unusually clear about what you are redesigning and why, rather than reaching for the org chart because the workload grew.

Worth noting what that framing does not do. It does not treat the answer as a matter of finding the right tool, and it does not treat the scope as something to be endured. It treats the criteria as the hard part — and criteria are the one thing a room of peers can genuinely stress-test against each other’s reality.

The question, then

Procurement is being asked to run the function it has and build the function it needs, at the same time, with the same people. The chart at the top of this piece is why that feels impossible: for fifteen years the answer to every new demand has been to take it on, and taking on has a ceiling.

Whether the profession keeps building on the same foundation, or changes what the foundation is for, is not something an article can settle — and anyone claiming a general answer is overreaching. What can be done is to ask the question properly: how did we get this, what genuinely stays, what changes, and what should honestly go back. A function that argued its way into this much ground is not short of agency. It is short of a moment to decide what to do with it. One considered view comes from the stage in October; the more useful answers usually come from the people sitting next to you.