The number procurement is reporting upward
Ardent Partners' 20th annual State of Procurement survey, fielded between January and March 2025 across 326 CPOs and procurement executives, put average spend under management (SUM) at 70.8% of enterprise spend, the highest figure in the firm's history of tracking it (Ardent Partners, 2025). Best-in-class procurement organizations reported 91.7% under management, against 61.1% for everyone else (Ardent Partners, 2025).
That is a genuinely good trend for the function. It is also a number that gets repeated at the board level as if it settles the question of control. It does not.
What spend under management actually counts
Spend under management measures whether a purchase runs through a formal channel: an approved contract, a preferred supplier list, a purchase order with sign-off, or a card program with pre-spend limits. It answers "does procurement have a process this dollar passed through," not "did the buyer follow the process procurement built."
Those are different questions. A purchase can sit inside a negotiated contract with a preferred supplier and still be non-compliant: wrong SKU, wrong quantity break, wrong service tier, a rate that was never updated after the last renewal. SUM counts it as managed because it touched the contract. It says nothing about whether the terms of that contract were actually honored.
Where the leakage McKinsey found actually sits
McKinsey's April 2025 analysis of procurement value leakage names four recurring sources: maverick spend, mismatched purchase orders, uncollected rebates, and contracts that auto-renew past the point where anyone re-checks the terms (McKinsey & Company, 2025). Three of those four can occur entirely inside spend that already counts as "managed." A mismatched PO against an approved supplier is still managed spend by definition. An auto-renewed contract with stale pricing is still managed spend. The metric has no mechanism for catching either.
In public-sector procurement specifically, McKinsey found that enforcing compliance with preferred-supplier contracts, on top of simply having them, preserved 10% to 50% of value that would otherwise leak away (McKinsey & Company, 2025). That is a compliance gain layered on top of coverage that already existed. Coverage and compliance are not the same lever.
The maverick spend data is a mess of denominators, and that's the tell
This is where the category gets genuinely confusing, and worth being precise about rather than picking whichever number sounds best. A 2019 Hackett Group and Coupa study on user experience and maverick spending found organizations losing up to 16% of negotiated savings to purchases made outside approved channels, and pointed to a lack of self-service and guided-buying tools as a primary driver (The Hackett Group, 2019, in partnership with Coupa; foundational study, dated). A widely repeated CIPS estimate from 2014, still cited in vendor content a decade later, put maverick buying at up to 80% of all invoices, even at large organizations with procurement functions (CIPS, Supply Management, 2014; dated, cite with age noted).
Those two numbers are not in tension because they measure different things. Sixteen percent is a share of negotiated savings lost. Eighty percent is a share of invoice count, which skews heavily toward low-dollar, high-frequency purchases: exactly the transactions least likely to be worth building a contract around in the first place. Neither is a share of total dollar spend, which is the number a board question about maverick spend is usually actually asking for.
APQC's open standards benchmarking program tracks that dollar-spend version directly, as the annual value of unmanaged or maverick spend as a percentage of total spend, and treats it as its own distinct metric from SUM for exactly this reason (APQC, Open Standards Benchmarking). If your organization only tracks SUM, you don't have this number. You have a proxy for it, and a generous one, since SUM only requires that a dollar entered a channel, not that it stayed compliant once it got there.
What to check before a SUM figure goes in a board deck
Before treating a spend under management percentage as evidence of control, three questions separate the number from the reality:
Does the SUM calculation exclude spend categories procurement has decided not to manage, or does it include the whole addressable base? A high percentage against a narrowed denominator is not the same claim as a high percentage against total spend.
Is there a separate, tracked figure for contract compliance within the spend that already counts as managed, distinct from SUM itself? If the only control metric in the room is coverage, non-compliant spend inside covered categories is invisible by construction.
When was spend under management last reconciled against actual invoice-level activity, rather than against the existence of a contract? A contract being on file is not evidence that the last twelve months of purchases against it matched its terms.
None of this makes spend under management a bad metric. It is a reasonable measure of reach. It was never built to measure compliance, and reporting it as if it does is the gap most procurement dashboards leave open.