Indirect and MRO Spend: The Half of Manufacturing Procurement Nobody Owns

Manufacturing procurement is really two functions wearing one name. Direct materials — the things that become the product — are forecast, contracted, expedited and reported on, by people whose job title says procurement. Indirect and MRO — bearings, lubricants, PPE, tooling, contract labour, facilities, the compressor that needs a service — are handled by whoever needs them, when they need them.

The asymmetry is not irrational. Direct materials sit in the bill of materials, move with production volume, and land in the gross margin line where the CFO can see them. Indirect spend is fragmented, unforecastable at the item level, sits in overhead, and is spread across every cost centre in the plant. One of those is easy to build a function around.

#Why the usual fix does not stick

The standard response is to extend the existing procurement team's remit to cover indirect. It rarely survives contact with a plant, for a reason that has nothing to do with capability.

A direct materials buyer is measured on continuity of supply to the production line. When a line stoppage and an indirect requisition compete for the same hour, the buyer makes the correct decision every single time, and the indirect requisition waits. Do that for a quarter and the plant learns that going through procurement is slow, so it stops.

#The four costs of an unowned category

Price, but less than you think

Uncontracted buying does cost more per item. On genuinely small purchases the absolute recovery is modest, and it is the number most often over-promised in a business case.

Duplicated stock across sites

Three plants each holding safety stock of the same components because none can see the others' inventory. Working capital tied up in triplicate, plus the obsolescence that follows.

Emergency premiums

A part not held, needed now, bought at spot price with expedited freight. Often the largest recoverable line, and usually recorded as maintenance cost rather than as procurement cost — which is why it is invisible.

Supervisor time

The most expensive and least visible. A maintenance supervisor spending several hours a week sourcing consumables is being paid to keep equipment running and is instead ringing suppliers.

Build the case on the last three. The first is real but it is the smallest of the four, and leading with it sets an expectation the programme will then be measured against.

#What actually works

Separate the queue before separating the category

Before any sourcing work, give indirect requests their own route and their own owner, with a committed turnaround. That single change addresses the reason people bypass the process. It is also the change that generates the data everything else depends on, because a routed request is a recorded one.

Catalogue the repeat items and nothing else

The long tail of indirect spend is genuinely long, and trying to catalogue all of it is how these projects die. A small set — the consumables reordered every month, the PPE, the standard fasteners and lubricants — covers a surprising share of transaction volume, which is what determines process cost.

Keep it current. A catalogue with stale prices or discontinued lines trains people to ignore it, and rebuilding that trust costs more than the original build.

Make cross-site stock visible before buying more

Multi-site manufacturers buy parts they already own more often than they realise. Full inventory integration is a large project; a shared list of high-value spares with holdings by site, refreshed weekly, captures most of the benefit and can be built in a fortnight.

The constraint is rarely technical. It is that someone has to maintain the list and answer the question when it is asked at 6am.

Attack the emergency premium at its cause

Every emergency purchase has a preceding cause: a stock-out, a reorder point set years ago against different consumption, a planned maintenance job whose parts were ordered late, or a failure nobody could have anticipated. Only the last is genuinely unavoidable.

Tagging emergency purchases against cause for one quarter tells you which of the four dominates, and each has a different and inexpensive fix. Without the tagging, all of them look like the same unavoidable operational cost.

#What to measure

• Share of indirect requests routed through the owned queue rather than raised directly. The leading indicator for everything else.
• Average time from request to confirmed order, for indirect. This is what determines whether the queue keeps being used.
• Emergency purchases per month, tagged by cause.
• Duplicate holdings across sites, by value, for the high-value spares list.
• Supervisor hours spent on purchasing, sampled rather than measured continuously.

The second metric deserves a target and visible reporting. Everything in this article depends on the compliant route staying faster than the workaround, and that is the number that says whether it is.

#Staffing it

The work described here — running the request queue, getting quotes, maintaining catalogue content, keeping the cross-site spares list current, tagging emergency purchases, chasing order confirmations — is a genuine full-time role or more at most mid-sized manufacturers. It has usually never been filled, because it is hard to justify a headcount for work that is currently invisible.

It also needs no plant-floor knowledge. The technical decision — which bearing, which lubricant, whether this pump can wait until Friday — stays with maintenance and engineering, where it belongs. What moves is the administration around that decision, which is the part currently consuming supervisor time.