01How to use this list
Group the KPIs into four questions and look at one or two from each group, rather than all twelve at once:
| Question | KPIs |
|---|---|
| Is spending under control? | 1. Spend under management, 2. PO compliance rate, 3. Budget variance |
| Is the process fast? | 4. Requisition-to-order cycle time, 5. Approval cycle time, 6. Invoice cycle time |
| Is it efficient? | 7. First-time match rate, 8. Cost per invoice, 9. Invoices per AP FTE |
| Are suppliers paid correctly? | 10. On-time payment rate, 11. Discount capture rate, 12. Duplicate and erroneous payment rate |
Define each KPI once, write the definition down, and do not change it between months. A KPI whose definition moves is a story, not a measurement.
02Control: is spending under control?
1. Spend under management. Share of total spend that went through your purchasing process (approved request, PO or contract). Formula: spend through the process ÷ total supplier spend. Example: $3.1M of $4.0M paid to suppliers in the year went through approved POs or contracts, so 77.5%. The other 22.5% is where surprises come from.
2. PO compliance rate. Share of invoices whose PO was issued before the invoice date. Formula: invoices with a PO dated before the invoice ÷ total invoices. Example: 640 of 800 invoices, so 80%. Look at the 160 by department: they are rarely spread evenly.
3. Budget variance. Gap between actual and budget, by budget line. Formula: (actual + committed - budget) ÷ budget. Example: the IT budget is €120,000, actual €78,000, committed €51,000. Variance: (78,000 + 51,000 - 120,000) ÷ 120,000 = +7.5%. Counting commitments is what makes this KPI useful before the year ends rather than after.
03Speed: is the process fast?
4. Requisition-to-order cycle time. Days from requisition submission to PO sent. Formula: median of (PO sent date - requisition submitted date). Use the median rather than the average: one purchase stuck for 60 days distorts an average and hides the typical experience.
5. Approval cycle time. Time spent waiting for approvals, per step and per approver. Example: median 0.6 days for budget owners, 2.8 days for one department head. That one number usually explains most of the complaints about "slow purchasing". Our guide to the purchase approval workflow covers deadlines and escalation.
6. Invoice cycle time. Days from invoice receipt to payment, or to approval for payment. Formula: median of (payment date - invoice received date). APQC reports that top-performing organizations take half as many days as bottom performers from receipt of the invoice to payment.
04Efficiency: is it efficient?
7. First-time match rate. Share of PO invoices that pass three-way matching with no human intervention. Formula: invoices matched automatically ÷ invoices with a PO. Example: 410 of 640 PO invoices, so 64%. Read the exceptions by cause (price, quantity, missing receipt): each cause has a different fix.
8. Cost per invoice. Total cost of the accounts payable process divided by invoices processed. Formula: (AP staff cost + systems + outsourcing) ÷ invoices processed. Example: $14,500 a month ÷ 800 invoices = about $18 per invoice. APQC reports that top performers spend about one fifth as much per invoice as bottom-quartile organizations, so the spread is large; compare yourself with your own previous quarters first.
9. Invoices per AP FTE. Invoices processed per full-time equivalent per month or year. Example: 800 invoices a month with 2.5 FTE, so 320 per FTE per month. This KPI rises with matching and PO coverage, not with working faster.
05Payment: are suppliers paid correctly?
10. On-time payment rate. Share of invoices paid on or before their due date (and not much earlier, unless a discount justifies it). Formula: invoices paid by due date ÷ invoices paid. Example: 712 of 800, so 89%. Late payment costs fees, goodwill and, in some countries, statutory interest.
11. Early-payment discount capture rate. Discounts taken ÷ discounts offered. Example: suppliers offered $9,400 of early-payment discounts this quarter; you took $2,100, so 22%. A "2/10 net 30" term is worth about 37% a year in annualised terms (2/98 × 365/20), more than most uses of cash.
12. Duplicate and erroneous payment rate. Value of payments later reversed, recovered or credited because they were duplicated or wrong, divided by total payments. Example: $11,200 recovered out of $3.9M paid, so 0.29%. Track the count too: a few large errors and many small ones call for different fixes.
06Building the dashboard
- Start with four. One per question: spend under management, approval cycle time, first-time match rate, on-time payment rate.
- Monthly, with a 12-month trend. A single month says little; the direction says a lot.
- One owner per KPI. Purchasing owns 1, 2 and 4; budget owners own 3 and 5; accounts payable owns 6 to 12.
- Drill down to the transaction. A KPI that cannot be traced back to the requests and invoices behind it will be argued about, not acted on.
Our free Monthly Procurement Dashboard template tracks six of these indicators month by month, and the Budget vs Actual Tracker computes budget variance with commitments. When collecting the data by hand takes longer than reading it, see spend analytics in Procura.
07A word on benchmarks
Benchmarks from APQC, from procurement software vendors and from consultancies are useful for orders of magnitude, but they are built on different definitions, company sizes and years. Compare yourself first with your own previous periods, then with a benchmark only when you are sure the definition matches. Improving your own first-time match rate from 50% to 70% is worth more than knowing where you sit against a median whose definition you cannot check.
Two traps are worth naming. The first is optimising one KPI at the expense of another: paying every invoice on day one lifts the on-time payment rate and ruins your cash position, while widening match tolerances lifts the first-time match rate and lets overbilling through. Read the KPIs in pairs (cycle time with error rate, match rate with tolerance changes) so that a gain on one side is not a loss hidden on the other. The second is measuring only what the system records. Purchases made outside the process, on personal cards or by invoice without a PO, do not appear in cycle times at all, which is why spend under management comes first in the list.
08P2P KPI checklist
- Each KPI has a written definition that does not change month to month.
- Cycle times use the median, not the average.
- Budget variance includes commitments, not only actuals.
- Match exceptions are analysed by cause.
- Each KPI has a named owner.
- The dashboard shows a 12-month trend.
- Every figure can be traced back to its transactions.
- Benchmarks are compared only when definitions match.