The internal request that precedes a purchase order — and the only point at which spend is genuinely still a decision.
A purchase requisition is an internal request to buy something, raised by whoever needs it and approved according to an authority matrix. It is not an order and creates no obligation to a supplier. The purchase order that follows is the external commitment. The requisition exists so that the decision to spend is made before anyone is committed.
Organisations that skip requisitions describe them as bureaucracy, and they are right that a slow requisition process is worse than none. But without one, the approval conversation happens after the supplier has been committed, at which point refusing creates a commercial problem. Approval after commitment is not approval.
The requester states what is needed, why, when and against which cost centre or project.
Value thresholds determine who approves. The matrix should be configured, not remembered.
Available budget verified before approval, so the decision is informed rather than retrospective.
Multiple requisitions for the same item can combine into one order, which is where purchasing earns its margin.
Purchasing selects the supplier, applies negotiated terms and raises the commitment.
A business sets approval limits by seniority: supervisors to 25,000, managers to 2,00,000, directors above that. It seems obviously sensible, and it produces a specific distortion.
A supervisor ordering routine consumables they buy weekly — low risk, well understood, pre-negotiated price — needs approval at 25,000. A manager commissioning a one-off service from a new supplier with no contract in place can approve 2,00,000 alone. The control effort is proportional to value and inversely proportional to actual risk.
A better structure varies the route by category as well as value: catalogue items with negotiated prices approve quickly at higher limits because the commercial risk was handled at contracting, while new suppliers, free-text requests and service engagements route through procurement regardless of value. The same total control effort is applied where it actually reduces risk, and the routine high-volume path gets faster, which is what stops people routing around it.
| Term | What it means |
|---|---|
| Requester | The person identifying a need, who is usually not the person who buys. |
| Approval limit | The value up to which a given role may authorise a request. |
| Conversion | Turning an approved requisition into a purchase order. |
| Catalogue item | A pre-negotiated item that can be requested without sourcing. |
| Free text requisition | A request for something not in a catalogue, which needs sourcing effort. |
If it takes a week to requisition something needed tomorrow, people will buy it another way.
Thresholds set years ago route trivial spend to senior people and material spend to nobody.
Two requisitions just below the limit. Detectable, and worth monitoring.
Specifying need is the requester's job; choosing the supplier is purchasing's. Conflating them loses negotiating leverage.
| Metric | What it tells you | What distorts it |
|---|---|---|
| Catalogue vs free text requisitions | How much spend is pre-negotiated. Free text is where sourcing effort goes. | |
| Approval cycle time by value band | Whether low-value requests are being delayed by disproportionate control. | |
| Requisitions rejected or amended | Request quality. High rates suggest requesters lack information or guidance. | |
| Requisition-to-receipt lead time | What the requester actually experiences, which drives workaround behaviour. |
We build and support the systems this page describes. Start with a free working proof of concept — no cost, no commitment.