Published 14 August 2026 · 6 min read
Procurement savings numbers have a credibility problem. Finance often does not believe them, and frequently for good reason — the method is rarely written down and almost never agreed in advance.
The fix is boring and effective: agree the baseline before you negotiate, and separate the types.
The three types, kept separate
- Hard savings. Spend that falls in absolute terms against a like-for-like baseline. Same item, same volume, lower price. This is the only type finance will recognise in a budget.
- Cost avoidance. A supplier proposed a 9% increase; you settled at 3%. Real, defensible, and not a budget reduction. Report it separately or you will lose the argument.
- Value gains. Better payment terms, longer warranty, consignment stock. Genuine, hard to monetise. Describe them, do not force a number.
Getting the baseline right
Most disputes are baseline disputes. Options, in descending order of defensibility:
- Prior-period actual price for the same item and comparable volume. Strongest.
- Weighted average paid across the last 12 months, if the price varied.
- Current contract rate, where one exists.
- Market benchmark. Weakest, and only where you genuinely have no history.
Whichever you pick, write it down and get finance to agree it before the negotiation, not after.
The volume trap
If you negotiate a 10% unit-price reduction and volume then rises 25%, total spend went up. That is not a failed negotiation, but reporting it as a saving without noting the volume change destroys trust.
Always report on constant volume: what would we have paid at last year's volume, versus what we now pay at that same volume. Show the volume movement as a separate line.
A worked example
Last year: 10,000 units at GBP 12.00 = GBP 120,000. New rate GBP 10.80. This year's volume: 12,500 units.
- Constant-volume saving: 10,000 × GBP 1.20 = GBP 12,000 hard saving
- Actual spend: 12,500 × GBP 10.80 = GBP 135,000, up GBP 15,000
- Reported as: GBP 12,000 saved, GBP 27,000 of volume growth absorbed
That version survives scrutiny. “We saved GBP 12,000” on its own does not, once someone opens the ledger.
Where classification comes in
None of this works without knowing what you bought. Category-level baselines require category-level classification, and if your transactions are not classified consistently between periods the comparison is meaningless. That consistency — the same taxonomy applied the same way each time — matters more than the sophistication of the analysis on top.
Further reading
- Maverick spend — often the largest recoverable pool.
- Spend cubes — the structure baselines come from.
- Procurement KPIs — what else to report.