Procurement Transformation: The Definitive 2026 Roadmap

A nine-stage maturity model for European procurement functions, with the test that proves you have reached each stage and the trap that stops most functions from reaching the next one.

TL;DR: Procurement transformation runs through nine stages in three phases: control, value, and advantage. Most large European functions sit between stages three and five in 2026, and the hardest gap on the model is influencing what the business asks for rather than how it buys. Each phase takes 12 to 18 months, and the stages cannot be skipped.

In this guide:

• How to use this maturity model
• Phase 1. Control
• Phase 2. Value
• Phase 3. Advantage
• Where most European functions actually sit
• Four traps that stall a transformation
• What the board needs to see

How to use this maturity model

The model has nine stages grouped into three phases, and its purpose is diagnostic rather than aspirational. Each stage has a test: a specific question you either can or cannot answer with data today. If you cannot answer the test for a stage, you have not reached it, regardless of what the operating model diagram says.

The stages are sequential because each depends on data or behaviour the previous one establishes. Should-cost modelling at stage five requires category strategies from stage three, which require the spend visibility of stage one. Functions that attempt a later stage without the earlier ones do not fail loudly. They produce a capability that works in a pilot category and cannot be repeated, which is a more expensive outcome because it consumes the credibility that funds the next attempt.

One caution on pace. Each phase realistically takes 12 to 18 months in a large organisation, so a function starting at stage one should plan on three to four years. That is longer than most transformation programmes are funded for, which is why the reporting habit at the end of this guide matters as much as the model itself.

Phase 1. Control

Stage 1. Spend visibility

The test: Can you produce, this week and without a manual exercise, total third-party spend by category, supplier, and business unit for the last 12 months?

Most functions believe they have passed this stage and have not. The common failure is a spend cube that requires two analysts and three weeks to refresh, which means the answer exists but cannot inform a decision at the speed decisions get made. Classification coverage is the usual gap: a long tail of spend sitting in miscellaneous categories that nobody has mapped because the effort is dull and the benefit is indirect.

The trap: Buying a spend analytics platform before agreeing a category taxonomy. The tool will classify against whatever structure you give it, and a poor taxonomy produces confident, precise, useless output.

Stage 2. Channel control and compliance

The test: What proportion of spend goes through a compliant channel with a purchase order, a contracted supplier, and an agreed price?

Channel control is where procurement stops being advisory. Until spend flows through channels the function controls, category strategy is a recommendation the business can decline, and every negotiated rate leaks through maverick buying. The number to watch is the share of addressable spend under contract, tracked by business unit rather than in aggregate, because the aggregate hides the units where compliance has collapsed.

The trap: Enforcing compliance through policy alone. Where the compliant channel is slower than the workaround, the workaround wins. Fixing the channel experience usually moves compliance further than escalating breaches does.

Stage 3. Category strategy

The test: For your top 20 categories, does a documented strategy exist that states the supply market structure, the sourcing approach, the target supplier base, and who owns it?

Category strategy is the point where the function starts making choices rather than running events. A real strategy explains why a category is consolidated or diversified, why it is contracted for three years or one, and what would trigger a change. Twenty categories is the practical threshold: enough to cover the majority of addressable spend, few enough to maintain properly.

The trap: Writing strategies that no sourcing decision references. If a category strategy has not changed a sourcing outcome in 12 months, it is documentation rather than strategy.

Phase 2. Value

Stage 4. Supplier segmentation and performance management

The test: Are suppliers segmented by strategic importance rather than spend volume, with a different management approach and review cadence for each tier?

Spend volume is the wrong primary axis, because a low-spend supplier holding a single-source qualification for a critical component carries more risk than a high-spend commodity supplier with five alternatives. Segmentation by criticality, substitutability, and innovation potential gives a different and more useful map. Performance management then follows the segment: quarterly business reviews and joint roadmaps for the strategic tier, scorecards and exception management for the rest.

Sarah Ricketts, Senior Vice President of Procurement at Rolls-Royce, centralised a procurement function of more than 1,200 people alongside a data and digital transformation, which is the scale at which segmentation stops being a spreadsheet exercise and starts requiring a system of record that the whole function actually uses.

The trap: Segmenting once and never revisiting. Supplier criticality changes with product portfolio and geopolitics, and a three-year-old segmentation is describing a supply base you no longer have.

Stage 5. Should-cost and total cost capability

The test: For your strategic categories, can you build a should-cost model from input costs, conversion, and margin, and does it inform the negotiation position?

Should-cost capability changes the nature of a negotiation from splitting the difference on a quoted price to discussing the components of a cost structure both sides can see. It requires genuine category expertise, access to commodity and labour indices, and buyers with the analytical training to build the model. That combination is why this stage separates functions more sharply than any other.

Total cost of ownership is the same discipline applied to the decision rather than the price. A supplier that quotes 4 percent lower and requires a second inspection stage, longer lead time, and more working capital is more expensive, and the function that cannot demonstrate that with numbers loses the argument to the one that shows the quoted price.

The trap: Treating should-cost as a negotiation weapon rather than a shared model. Suppliers who see a should-cost model used only to extract price stop engaging with it, and the capability degrades into an internal spreadsheet.

Stage 6. Demand management with the business

The test: Has procurement changed what the business specifies or how much of it it consumes, not just what it pays per unit?

This is the widest gap on the model. Every stage up to five improves how the organisation buys. Stage six changes what it asks for, which requires influence over engineering specifications, marketing briefs, and IT architecture decisions that were never procurement's to make. The value available here typically exceeds everything in the previous stages combined, and capturing it depends almost entirely on relationships rather than process.

Angelique Van Der Burg, Executive Vice President and Chief Procurement Officer at Infineon Technologies, leads procurement across materials, machinery, R&D services, and logistics, a scope that reaches directly into what the business specifies rather than only what it pays. Infineon reports a 56.8 percent reduction in its carbon footprint since 2019 alongside that mandate, which is the kind of outcome that requires demand-side influence rather than sourcing alone.

The trap: Attempting demand management before stage five. Without should-cost credibility, a procurement challenge to a specification reads as cost-cutting by a function that does not understand the requirement, and the door closes.

Phase 3. Advantage

Stage 7. Supplier-enabled innovation

The test: Can you name a product or process improvement in the last 12 months that originated with a supplier and reached implementation?

Supplier-enabled innovation is the return on strategic segmentation. It requires suppliers to bring ideas before they are commercially obliged to, which happens only where the relationship has enough continuity and enough upside for them. The measurable version is a pipeline: ideas submitted, ideas evaluated, ideas implemented, and the value attributed.

The trap: Running an innovation programme with suppliers whose contracts are re-tendered annually on price. The commercial model has to reward the behaviour you are asking for.

Stage 8. Regulatory and sustainability criteria inside the sourcing decision

The test: Do supplier emissions, human rights due diligence, and regulatory exposure appear as weighted criteria in the award decision, or in a report produced afterwards?

The distinction is the whole stage. Reporting Scope 3 emissions is a compliance activity. Weighting supplier emissions in a sourcing award is a procurement capability, and it changes the supply base over time in a way reporting never does. The same applies to due diligence obligations, which are becoming a sourcing constraint rather than a disclosure exercise for European buyers.

Thomas Udesen, Chief Procurement Officer at Bayer, co-founded the Sustainable Procurement Pledge in 2019, which has grown to more than 15,000 procurement professionals across 140 countries. The premise of that work is that the sourcing decision is where procurement's environmental influence actually sits, which is the argument for pulling these criteria forward into the award rather than reporting them after it. Our guide to Scope 3 for CSCOs covers the measurement side in detail.

The trap: Setting supplier sustainability requirements without a way to verify them. Self-declared supplier data that nobody audits produces a compliant-looking supply base and no actual change.

Stage 9. Agent-supported execution

The test: Do agents handle a defined class of transactional sourcing decisions inside written guard-rails, with a named human owner and an audit log?

Ard Verboon, Chief Procurement Officer at Schneider Electric, told PASA in 2025 that his team has been running autonomous negotiation bots on transactional and commoditised spend, covering the RFQ, pricing, and award cycle within defined limits. Verboon oversees roughly €18 billion in annual procurement spend, and the categories the bots handle are deliberately the ones where a wrong answer is recoverable.

This stage sits last for a reason. An agent needs clean supplier master data from stage one, a compliant channel from stage two, a category strategy that defines what a good outcome is from stage three, and written decision rights. Deployed before those exist, it automates a process nobody has fixed. The build sequence for agentic supply chains applies here in full.

The trap: Treating agent deployment as the transformation rather than its final stage. The technology is the easiest part of stage nine, and the organisations that reach it did not get there by starting with it.

Where most European functions actually sit

Most large European procurement functions sit between stages three and five in 2026. Category strategies exist and supplier segmentation has been done, but should-cost capability is concentrated in a few categories and demand management with the business remains aspirational. That clustering is consistent with what CPOs report when they compare notes honestly, and it explains why so many transformation programmes plateau at the same point.

The reason the plateau falls there is that stages one to five can be delivered by procurement alone. Stage six cannot. It requires engineering, marketing, and IT to accept procurement's influence over what they specify, and no amount of internal capability building produces that consent. It is earned through the credibility built in stage five and through relationships that a reorganisation cannot create.

Functions that break through tend to do it category by category rather than through a programme. One engineering category where a should-cost model changed a specification and saved real money becomes the reference case that opens the next conversation. This connects directly to how the CSCO mandate itself has widened, since in many European organisations the CPO now reports to the CSCO and the two agendas are planned together.

Four traps that stall a transformation

Technology before visibility. A source-to-pay platform implemented over an unresolved taxonomy produces faster access to unreliable data. Fix stage one first, even though it is the least interesting work in the programme.

Savings the CFO does not recognise. Procurement savings claims are widely distrusted at board level because the definitions vary between functions and years. Agree the methodology with finance in advance and report only what they will validate. A smaller number that survives scrutiny is worth more than a larger one that does not.

Structure without capability. Centralising a function moves reporting lines and changes very little on its own. Rolls-Royce's centralisation worked alongside a capability and data programme, and it is the second half that produces the outcome.

Compliance as reporting. Treating sustainability and due diligence as disclosure rather than as award criteria produces a well-documented supply base that has not changed. Stage eight exists to make this distinction concrete.

What the board needs to see

A board wants three things from a procurement transformation update: which stage the function occupies, what capability is being built this year, and what value that capability released, in numbers finance has agreed. The stage model is useful in the boardroom precisely because it gives a multi-year programme a progress measure that is not an annual savings figure.

The reporting habit that separates funded transformations from audited ones is validation. A CPO whose savings number carries the CFO's agreement gets the next tranche of investment. One whose number is self-certified gets a review, and the review consumes the year. Agreeing the methodology before the results exist is unglamorous and is the highest-return governance decision available at the start of a programme.

Berlin this December runs several sessions on exactly this ground, including next-generation procurement from spend automation to strategic sourcing, a multipolar sourcing playbook for regions with diverging rules, and a session on why human judgement still decides outcomes in AI-driven supply networks. The TFEST26 agenda is published in full, and the CPO conversations in those rooms tend to be considerably more candid about what stalled than any conference stage usually allows.

Compare transformation notes with European CPOs and CSCOs at TFEST26 in Berlin, December 1 and 2, 2026

---

The procurement mandate keeps widening, and the pressure from tariffs, due diligence obligations, and AI deployment is arriving faster than most functions can build capability for. The stages themselves have not changed much in a decade. What has changed is how expensive it now is to be stuck at stage four. We update this roadmap as European CPOs tell us what moved them forward.

— TFEST26 Editorial Team