How to size MRO inventory for critical items

The problem is not the inventory, it is the criterion

Maintenance, repair and operations inventory usually grows by accumulation rather than by decision. Something is bought under emergency, material is left over from a shutdown, the item enters the warehouse with no replenishment policy, and a few years later the company has capital tied up in parts that will never move and is missing exactly the item that stops the unit. Sizing MRO inventory therefore does not start with a consumption spreadsheet, it starts with criticality classification.

The question that organises everything is easy to state and hard to answer honestly, namely what happens to production if this item is missing when it is needed. The answer splits the inventory into families with completely different sizing logic.

Classify before calculating

Consequence criticality

The first axis is the effect of the shortage, ranging from total unit shutdown, through partial loss of capacity and safety or environmental risk, down to simple operating inconvenience. High consequence items justify a service level close to the maximum even with very low turnover, because the cost of unavailability dominates any inventory calculation.

Demand predictability

The second axis separates regular consumption, which accepts classical statistical models, from intermittent consumption, which is the rule in critical items and makes the average useless. An item that moves twice in five years has no meaningful average, so sizing comes to depend on estimated failure rate, installed population and risk policy rather than on issue history.

Replenishment lead time and reliability

The third axis is the real replenishment lead time, which includes manufacturing, testing, documentation, international transport, customs clearance and receipt inspection. It is common for the actual lead time to be much longer than the one quoted on the order, and it is the variability of that lead time, more than its average, that determines safety stock.

How the calculation changes by family

For items of regular consumption and stable lead time, the traditional reorder point works well, with safety stock proportional to the variability of consumption and lead time within a chosen service level. For critical items of intermittent consumption the reasoning inverts and becomes one of availability, namely how many spare assemblies are needed so that the probability of a shortage when demand arises stays below the accepted limit, given installed population and failure rate.

Add to that the decision of where the stock sits, since not every item needs to be inside the unit. Part of the risk can be transferred to the supplier through consignment stock, a service level agreement with contracted lead time or reserved manufacturing capacity, arrangements that reduce tied capital without handing the risk back to operations.

The four mistakes that cost the most

The first is using turnover as the sole criterion, which systematically eliminates the low rotation critical item, which is exactly the one that cannot be missing. The second is ignoring obsolescence, since a part sitting in the warehouse for years can lose traceability, elastomer shelf life or compatibility with the current equipment revision. The third is sizing item by item without looking at the installed population, which leads to stocking three assemblies of equipment that exists in duplicate and none of another that exists in dozens. The fourth is treating cataloguing as bureaucracy, when inconsistent descriptions and duplicate codes are the most banal and most frequent cause of buying what is already in stock.

What good sizing delivers

The result of proper work is not a list of minimums and maximums, it is a policy, with criticality assigned per item, service level defined by class, replenishment lead time measured rather than assumed, a periodic review rule and clear accountability for approving exceptions. Because of that, the gain shows up on two fronts at once, namely less idle capital and fewer unplanned shutdowns.

The inventory policy speaks directly to the repair or replace decision, covered in when to repair and when to replace a safety valve, and to the monitoring regime adopted, covered in what predictive maintenance detects that preventive does not. The corresponding fronts are described in Solutions.