Why Year-on-Year Savings Keep Getting Harder to Find

Every procurement function is asked the same question each budget cycle: what did you save this year, and what will you save next year? The pressure hasn't eased — Deloitte's 2025 Global CPO Survey, based on responses from more than 260 chief procurement officers, found that improving margins through cost reduction (72%) and driving operational efficiency (68%) remain the top two priorities for responding to macroeconomic pressure. The stakes are proportionally large: procurement typically controls 50 to 80% of a company's cost base, and Deloitte notes that a 5% reduction in procurement costs can move the bottom line as much as a 20% increase in sales.

Yet the research from 2025 and into 2026 tells a more complicated story than "just find more savings." Three themes stand out: the easy savings are gone, a growing share of negotiated savings never reaches the P&L, and the tools now emerging to fix both problems are still immature in most organisations.

The comparison problem: less capacity, more workload

The Hackett Group's 2026 Procurement Agenda and Key Issues Study puts a number on something most procurement leaders already feel. Workloads are projected to rise 8% in 2026, while headcount falls 0.9% and operating budgets contract 0.4% — producing an estimated 8.9% productivity gap and an 8.4% efficiency gap. Against that backdrop, it's notable that only 45% of procurement teams now expect increasing savings in 2026, down from 55% the year before, with a larger share expecting savings to stay flat.

This matters for year-on-year comparisons specifically. Once the obvious levers — first-round supplier consolidation, low-hanging renegotiations, basic tail-spend cleanup — have been pulled in prior cycles, each subsequent year has to find opportunity in a smaller and more fragmented pool of spend, using fewer people. Identifying "new" savings increasingly means going deeper into categories that were previously deprioritised, rather than re-running the same playbook.

The bigger issue: savings that are found but never realised

A separate and arguably more urgent theme in the research is savings leakage — the gap between what procurement negotiates and what the organisation keeps. GEP's research puts the floor at 20% of negotiated savings lost, with some analyses suggesting procurement teams lose 30 to 60% of negotiated value between contract signature and invoice payment. A related industry estimate cited by PASA puts enterprise-wide contract leakage at around 11% of contract value. The causes are consistent across sources: maverick spend (buying outside negotiated agreements, which GEP estimates can reach 20% of indirect spend), inconsistent contract compliance, missed renewal windows, and poor integration between procurement and finance systems that makes it hard to see where negotiated terms broke down.

Illustrating the gap between good and average performers, GEP research finds that world-class procurement teams achieve roughly 74.9% contract-compliant spend, against 59.5% for the average team.

The practical implication for anyone building a year-on-year opportunity pipeline: identifying a saving and banking a saving are different exercises, and most of the research now treats leakage as a governance and tracking problem as much as a sourcing one — not something a savings report can gloss over.

Where indirect spend sits

Indirect spend is where the research consistently points for remaining opportunity. Oliver Wyman estimates organisations can reduce indirect spend by 10 to 15% over three years, notable given indirect categories represent roughly 45% of total spend on average — yet only 41% of organisations regularly track execution of indirect savings once they're identified. Spend analysis research more broadly suggests 5 to 15% savings potential through supplier consolidation, demand management, and rate benchmarking, and that 20 to 40% of enterprise spend still occurs off-contract, invisible to standard reporting until someone runs the analysis.

The pattern across these figures is that opportunity is shifting from price negotiation on well-managed direct categories toward total-cost and compliance work on less-managed indirect and tail spend — categories that require better data before they yield anything, not better negotiators.

AI is being positioned as the answer, but adoption is early

Every major 2025–2026 report — Deloitte, McKinsey, Hackett — points to AI and agentic tools as the mechanism for finding and holding onto savings faster. Deloitte found 92% of CPOs are planning or assessing generative AI capability, and industry reporting cites AI-driven sourcing delivering around 20% cost savings in some deployments, with one cited case unlocking $25.5 million in previously hidden savings from unmanaged tail spend using agentic AI. McKinsey frames this as a structural shift: procurement moving from a transaction-processing function to a strategic driver of growth and resilience as agentic AI takes over routine sourcing tasks.

The caveat, consistently, is readiness. Only 11% of procurement leaders say they are fully ready to leverage AI, and 73% cite poor or incomplete data quality as the main barrier. In other words, the tools that promise to solve both the "opportunity is getting harder to find" problem and the "savings leak before they're realised" problem depend on the same spend-data foundation most organisations haven't yet built.

A structured answer: the Annual Procurement Opportunity Assessment

The research above describes symptoms. What most procurement teams still lack is a repeatable process for turning "where's next year's saving" into an evidence-based annual plan rather than a re-run of last year's category list. That's the gap our Added Value Sourcing™ Annual Procurement Opportunity Assessment is built to close, and it maps onto each issue above with reasonable precision.

The capacity problem — a shrinking pool of obvious opportunity, chased by fewer people — is addressed by structure rather than headcount. Ten weighted dimensions and forty-seven factors replace an unstructured annual review, so the same systematic pass surfaces new opportunity each cycle instead of relying on memory of what worked last time. Demand Management, Supplier Incentives, and TCO/Cost Modelling carry the heaviest weighting deliberately, because independent research consistently finds the most value hiding there: 15 to 30% waste in unjustified purchasing, contracts with no incentive for suppliers to reduce total cost, and 60 to 70% of total cost going unmanaged because only unit price gets attention.

The leakage problem — savings negotiated but never realised — is addressed directly by the dimension scoring whether the contract mechanism itself rewards lower spend or just higher volume, and by a separate dimension that reads off-contract and maverick spend straight from the transaction data rather than estimating it. A third dimension catches the contract-term gaps — missing protections, stale benchmarking clauses — that research ties to leakage in the first place.

The indirect and tail-spend problem is where the transaction-data dimension earns its keep: supplier concentration, ABC classification, and the true off-contract rate, computed from a pasted spend extract rather than sampled or guessed at.

The AI-readiness problem is addressed by starting with the formula-driven version of the same three techniques AI platforms run at scale — price variance detection, supplier concentration, contract matching — done transparently in a spreadsheet first. That's also how a team builds the clean, consistent spend data that any future AI layer will need; skipping straight to AI without it is why so many CPOs report readiness gaps.

One further lever the general research above doesn't surface as clearly, but which shows up repeatedly in practice, is collaborative and aggregated buying power: joint purchasing with other sites, group purchasing organisations and consortia, supplier-led preferred-customer boards, and industry peer networks. Organisations using group purchasing arrangements typically report 10 to 25% savings against categories bought independently, and public-sector cooperative purchasing programs commonly cite 15 to 20% savings for participating members. It's a large enough lever that our assessment scores it as its own dimension rather than a footnote — it's often the fastest-payback opportunity available, because the mechanism (a cooperative contract, a piggyback agreement, a seat at a supplier's advisory board) frequently already exists and simply isn't being used.

One further lever the general research above doesn't surface as clearly, but which shows up repeatedly in practice, is collaborative and aggregated buying power: joint purchasing with other sites, group purchasing organisations and consortia, supplier-led preferred-customer boards, and industry peer networks. Organisations using group purchasing arrangements typically report 10 to 25% savings against categories bought independently, and public-sector cooperative purchasing programs commonly cite 15 to 20% savings for participating members. It's a large enough lever that our assessment scores it as its own dimension rather than a footnote — it's often the fastest-payback opportunity available, because the mechanism (a cooperative contract, a piggyback agreement, a seat at a supplier's advisory board) frequently already exists and simply isn't being used.

The takeaway for year-on-year planning

Taken together, the research suggests that identifying credible new savings each year now depends less on running the same category reviews and more on four things: extending analysis into indirect and tail spend that hasn't been touched, building the tracking discipline to convert negotiated savings into realised ones, investing in the data quality that any AI-assisted opportunity identification will need to work at all, and checking whether the opportunity is even yours to negotiate alone versus available faster through collaboration. Organisations treating "find next year's savings" as purely a bilateral sourcing exercise are, according to this research, increasingly missing where the value is. Running a structured assessment like this one annually, ahead of budget planning, is one concrete way to put that shift into practice rather than just naming it.