What are procurement KPIs, and which ones should you track?

What is a procurement KPI?

A procurement KPI (key performance indicator) is a measure that shows whether procurement meets one of its goals, such as lower costs, compliant buying or faster orders. Each KPI has a written formula, a data source, an owner and a target, and is tracked over time.

The difference between a KPI and a metric is the goal. The number of purchase orders sent last month is a metric: it describes activity. The share of invoices that had a purchase order raised before the invoice date is a KPI when the organisation has decided it wants buying to follow the process and has set a target for it. The same number can be a KPI for one team and a background metric for another.

This page lists the common procurement KPIs with their formulas, by goal and by role, and shows how to put them on a dashboard. Three other guides go deeper on parts of the subject: procurement management gives the six headline indicators of the function, procurement process improvements uses measures to find where a process loses time, and supplier performance management covers the KPIs and scorecards for individual suppliers.

Which procurement KPIs are most common? Examples with formulas

The KPIs below are the ones procurement teams use most, grouped by the goal each one measures. Formulas follow published definitions where one exists, from APQC’s procurement benchmarking measures, the CIPS glossary and public guidance; adapt the scope to your organisation and write it down.

Supplier on-time delivery is listed here for completeness. How to measure each supplier, combine KPIs into a scorecard and run reviews is covered in supplier performance management. Spend under management and the spend KPIs depend on clean, classified spend data, which is what a spend analysis produces.

How do you measure procurement savings without overstating them?

Savings are the most quoted procurement KPI and the easiest to overstate. Three rules keep them honest: write down the baseline before the negotiation, count savings on the quantity actually bought rather than the forecast, and agree the method with finance so the figure can be found in the budget.

APQC’s procurement measures keep two kinds of savings apart, and so should a KPI report. Cost takeout (sometimes called hard savings) is a lower price than the one paid before, which reduces spend. Cost avoidance is a price increase that procurement prevented or reduced, such as a supplier asking for 8% and accepting 3%: it protects the budget but does not lower spend. Reporting the two as one number is the most common reason finance does not recognise procurement’s savings.

A worked example, with fictional figures. A company buys about 40,000 cartons a year at 1.20 each. After a tender, the new contract price is 1.08. The baseline is written down before the tender: 1.20. In the first year the company actually buys 36,000 cartons. Savings delivered are (1.20 − 1.08) × 36,000 = 4,320, not the 4,800 the forecast volume would give. Meanwhile the budget assumed a standard price of 1.10, so purchase price variance is (1.08 − 1.10) × 36,000 = −720, a favourable variance. The two KPIs answer different questions: savings compare with what the company paid before, price variance compares with what the budget planned.

The cost of the procurement function itself is a KPI too. APQC defines it per 1,000 of purchases and per purchase order: the total cost of people, systems, overhead and outsourced work, divided by the value bought or the number of orders. It is useful for comparing your own years. Comparisons with other organisations need the same definitions on both sides.

Which KPIs show whether people follow the purchasing process?

Control KPIs show whether purchases go through the agreed route: an approved request, a purchase order to an approved supplier under a contract, then an invoice that matches. They usually come first, because savings and supplier terms only work when buying follows that route.

• Purchase order compliance. The share of invoice value with a PO raised before the invoice. Measure it on value and on count: a few large invoices can hide many small ones bought outside the process.
• Retrospective purchase orders. Orders whose PO date comes after the invoice date. The City of York Council, for example, defines them this way in its reporting on “No Purchase Order No Pay” and monitors them by department. A rising share means people order first and do the paperwork later.
• Maverick spend. CIPS defines it as “unauthorised spend often with non-approved suppliers”. Count it as spend with suppliers outside the approved list, or outside the channel the policy sets for that category.
• Non-PO invoices outside the listed exceptions. Rent, utilities and taxes have no PO by design. Anything else without a PO is a purchase that skipped the process. The difference between the two documents is explained in purchase order vs invoice.
• Catalogue use. The share of purchase order lines placed through catalogues or punchout, where the price and supplier are already agreed. APQC measures it as PO lines transacted with e-procurement catalogue suppliers.

Where a control KPI is poor, the fix is usually upstream: contracts and catalogues that cover what people actually buy, a simple way to raise a request, and a procurement policy that lists the exceptions. How to find and fix the cause is covered in procurement process improvements.

Which KPIs measure the speed of procurement?

Speed KPIs measure the time between two dated events. The most useful three cover the three main waits a requester or supplier feels: getting an order out, getting a contract signed and getting paid.

Requisition-to-PO cycle time

APQC defines it as the days, weekends included, from receipt of the purchase requisition to release of the purchase order to the supplier. Report the median and split catalogue items from free-text requests: they behave very differently.

Contract cycle time

Days from an approved need to a signed contract. APQC measures the “average cycle time in days to establish a contract with a supplier”. Split it by route: a request for quotation and a full RFP are not comparable.

Procure-to-pay cycle time

Days from the requisition to the payment, measured separately for goods and for services in APQC’s measures. It shows how the whole chain performs, not one team.

On-time payment

The share of invoices paid by their due date. For public buyers it is also a legal KPI: EU public authorities must, as a rule, pay within 30 days of receiving the invoice, UK public contracts imply 30 days from a valid invoice, and US federal agencies owe interest when they pay after the required payment date.

Speed KPIs are easy to improve in the wrong way: approving faster by checking less. Read them next to a control KPI, such as PO compliance or the share of invoices blocked at matching, so a faster process is not a looser one.

Which KPIs do public-sector buyers use?

Public buyers track the same cost, control and speed KPIs, plus indicators of competition, transparency and access for small firms, because the law asks them to spend public money openly. Several of these indicators are defined in public documents, with thresholds set by the body that publishes them.

The Commission’s scoreboard measures countries, not single buyers: it marks each Member State green, yellow or red against these thresholds. A public body can compute the same ratios on its own tenders to see where it stands. The thresholds are the same in the 2019 edition and in the 2024 reporting period; check the scoreboard before quoting them.

Contract KPIs are a different thing. In the UK, the Procurement Act 2023 requires a contracting authority to set at least three KPIs for most public contracts above £5 million and to publish the supplier’s performance against them at least once a year. Those KPIs measure a supplier on one contract; how to set and review them is covered in supplier performance management.

Which KPIs should a procurement manager track?

A procurement manager should track the few KPIs that match what the role controls: the function’s cost and control KPIs, one speed KPI and the team’s savings. Each role sees the same definitions at a different level of detail, so the figures add up from a buyer’s orders to the head of procurement’s report.

Accounts payable KPIs are explained in invoice processing. The goals the head of procurement’s KPIs measure come from the procurement strategy: a KPI should show progress on one of its objectives, or it is reporting for its own sake.

How do you choose and define procurement KPIs?

1. Start from the goals. List the three to five objectives in the procurement strategy, such as cutting indirect costs, bringing spend under contract or paying suppliers on time.
2. Pick one or two KPIs per goal. Prefer KPIs that use data you already have; a KPI that needs a manual count every month will not last.
3. Write a definition sheet for each KPI: formula, scope, data source, owner, frequency and target. Note what is excluded, such as taxes, intercompany charges or framework call-offs.
4. Measure a baseline over at least one full period before setting the target. Set the target from that baseline, not from a figure found in a vendor’s blog.
5. Agree the definitions with finance before publishing, so procurement and finance report the same number.
6. Review the set once a year with the strategy. Drop KPIs that nobody acts on, and keep the definitions of the others stable.

A definition sheet can be one block of text per KPI. An example, ready to copy and adapt:

KPI: Purchase order compliance (%)
Goal: Purchases follow the procurement process
Formula: Invoice value with a PO dated on or before the invoice date / total invoice value x 100
Scope: All supplier invoices; excludes rent, utilities, taxes and the other exceptions listed in the procurement policy
Data source: Accounts payable invoice register and purchase order table (ERP)
Owner: Procurement manager
Frequency: Monthly, by department
Baseline: [value and period]
Target: [value and date]
Action when below target: Department review of retrospective orders; reminder of the no PO, no pay rule

On benchmarks: many published lists give target values without saying where they come from, and benchmarking bodies such as APQC publish their measure definitions openly but their values mostly to members. Different organisations also count spend and invoices in different ways. Use your own trend as the main reference, and compare with others only where the definitions are the same.

What should a procurement KPI dashboard show?

A procurement KPI dashboard shows the chosen KPIs on one screen, each with its target, its trend and a way to see the transactions behind it. It answers three questions in order: are we on target, where are we off target, and which orders, invoices or suppliers cause it.

Level 1: the function

Six to eight KPIs for the whole organisation, each with the current value, the target, the last 12 periods and a red, amber or green status. This is the page the head of procurement and finance look at.

Level 2: by category or department

The same KPIs split by category, department, site or buyer, to see where a figure comes from. A PO compliance of 88% can hide one department at 60%.

Level 3: the transactions

The list of orders, invoices or suppliers behind a figure: the retrospective POs of last month, the invoices blocked at matching, the spend with suppliers outside the approved list. This is where people act.

A dashboard is only as good as its data. Most procurement KPIs need four sources joined on the same supplier and category codes: the purchase orders, the goods receipts, the invoices and the contract list. Spend KPIs also need the spend classified by category, which is the work of a spend analysis. Clean the supplier master data first, or the same supplier under three names will spoil every supplier KPI.

An example of level 1, with fictional figures for a company buying about 40 million a year:

The tool matters less than the definitions. A spreadsheet works for a first dashboard with a monthly extract from the ERP. A business intelligence tool suits larger data and several levels of detail. Most procure-to-pay and source-to-pay suites include standard procurement reports; check that their formulas match your definition sheets before you rely on them. Which software covers which steps is explained in procure-to-pay software.

What are the common mistakes with procurement KPIs?

• Too many KPIs. A dashboard with twenty-five KPIs is read by no one. Keep six to eight at level 1 and leave the rest as background metrics.
• Savings counted on forecasts. Savings reported at contract signature, on forecast volumes, rarely match the budget. Count them on what was actually bought.
• Mixing cost avoidance and cost takeout. Report them on separate lines, or finance will discount the whole figure.
• Changing formulas. Changing the scope of spend under management each quarter makes the trend meaningless. Change a definition only once a year, and restate the past figures.
• Targets copied from elsewhere. A target taken from a vendor’s list or another company’s report may count things differently. Set targets from your own baseline.
• Speed without control. A faster requisition-to-PO time can mean fewer checks. Read speed KPIs next to compliance KPIs.
• No one owns the figure. Every KPI needs one person who can act on it and reports what was done when it is off target.

A KPI set that avoids these mistakes is part of running procurement as a function, which is covered in what is procurement management.