Parts: What to Hold and What Not To

Holding inventory ties up money and space, and restaurant groups have neither to spare. The correct answer for most parts is to hold none and rely on the contractor.

A short list of exceptions is worth stocking, and it is almost never the list operators expect.

What to hold

Door gaskets for the principal refrigeration units. Cheap, fail predictably, and a failed gasket runs a unit continuously until replaced. Holding one per critical unit is inexpensive insurance.

Filters. HVAC filters, water filters. Consumable, scheduled, and a filter change deferred because none is on site is a change that does not happen.

Fuses and contactors for critical equipment, if there is anyone competent to fit them.

Fluorescent tubes and lamps, including hood and walk-in lighting, because a dark walk-in is a safety issue.

Hose washers, tap seals and basic plumbing consumables.

Thermometers. Spare calibrated units for temperature verification. Cheap, and a site without a working thermometer cannot demonstrate compliance.

Test strips for sanitiser concentration.

Fusible links where the suppression contractor supplies them and the site is remote.

That is most of it. Perhaps a few hundred dollars per site.

What not to hold

Compressors, motors, control boards. Expensive, model-specific, and the contractor sources them faster than you can.

Anything requiring a licensed technician to fit. Holding a part you cannot legally install saves nothing.

Parts for equipment you have one of. The economics only work across multiple identical units.

Anything with a shelf life — belts and hoses degrade in storage.

Parts bought "because it was on offer". Restaurant storerooms accumulate these and they are never the part that fails.

The standardisation argument

The real parts strategy is upstream of inventory.

Specify the same models across sites. An operator with one cooler model across thirty sites has a meaningful parts position, predictable failure patterns, technicians who know the equipment, and negotiating leverage.

An operator with fifteen models across thirty sites has none of that, and each failure is a fresh diagnosis and a special order.

This is a procurement decision made years before the maintenance consequence appears, and it is rarely made with maintenance in mind.

Where standardisation is not possible, at least record the models and consolidate over time as equipment is replaced.

Lead times and criticality

The question that determines whether to hold something is not cost, it is what happens while you wait.

Map the critical assets against parts lead time. A walk-in condensing unit with a six-week lead time is a different risk from one available next day.

For long-lead critical items, the options are: a supplier commitment to hold stock, a contractual response arrangement, redundancy in the equipment, or a plan for rented replacement.

Ask contractors what they hold. A vendor with local stock of the components your equipment uses is worth more than a marginally cheaper rate.

Working with contractor stock

Most operators should push inventory to the vendor, and negotiate for it.

Ask, in the service agreement: what parts do you stock locally for our equipment, what are your typical sourcing times, and what happens when a part is not available.

A vendor who cannot answer is a vendor whose response commitment covers arrival and not resolution.

Consider a consignment arrangement at larger scale, where the vendor holds stock for your estate and you pay on use.

The record that matters

Log parts used on the work order, with part numbers.

This accumulates into the useful thing: which parts fail, on which assets, how often. After a year that history tells you what is worth stocking far better than any general list, including this one.

It also reveals repeat failures. The same part replaced three times on one unit is a diagnosis nobody made.

The standardisation conversation

The parts strategy is decided at procurement, years before maintenance feels the consequence, and maintenance is rarely consulted.

Ask to be involved in equipment specification. The maintenance function should have a view on model selection, and usually does not have a seat.

The argument to make: an estate with three cooler models has predictable failures, a viable parts position, technicians who know the equipment and real negotiating leverage. An estate with fifteen has none of that, and every failure is a fresh diagnosis and a special order.

Bring evidence. Cost per asset by model, from the register, shows which specification actually performs. This is the most valuable output of the whole measurement effort and it takes a few years of data to produce.

Accept incremental progress. Nobody replaces an estate to standardise it. Specifying consistently as equipment is replaced converges over a decade, which is the realistic timeframe.