Repair or Replace Equipment? A Four-Factor Decision Framework

Something comes back broken. Somebody has to decide whether to fix it or bin it, and in most places that call gets made in under a minute, on instinct, by whoever happens to be holding it.

I used to make that call badly for years. Not because I was careless — because the four things that should have decided it were spread across a supplier inbox, a pile of invoices, a finance spreadsheet and my own memory. None of which were in the room.

The framework in one table

1. Is it still covered?

The single most expensive thing not to know. Warranty claims are missed for one reason: nobody knew the cover was live at the moment the decision was taken. By the time someone thinks to check, the item has been repaired at your expense or replaced.

This extends well past manufacturer warranty. Extended cover, annual maintenance contracts (AMC) and comprehensive maintenance contracts (CMC) all change the arithmetic, and each has a policy number and a claim route that needs to be to hand rather than in a drawer.

Holding cover on the asset record — with the provider, the policy number, what the cover cost, and who to call — turns this from something a person has to remember into something the system answers at the moment the question is asked.

*Cover held against the asset: warranty, AMC and CMC contracts with providers, renewal dates and claim details.*

2. What has it already cost you?

A first repair is a repair. A fourth repair on the same item in eighteen months is a replacement you've been paying for in instalments without noticing, because each individual decision looked reasonable in isolation.

Accumulated repair cost against an asset is the number that exposes that pattern, and it only exists if every ticket recorded what the work actually came to rather than what it was expected to. Estimate is a planning figure; actual is the one that tells you the truth about an item.

A rule of thumb worth adopting: if cumulative repair spend has passed roughly half of replacement cost, stop repairing and put the case for replacement. The exact threshold is yours to set — having the number at all is the point.

3. What is it worth now?

Not what it cost. Current book value — purchase price depreciated over its useful life — is what makes the comparison meaningful.

Spending 60% of replacement cost repairing something with eighteen months of life left is a different decision from spending it on something nearly new, and the two look identical if all you've is the original invoice.

4. What has happened to it before?

The service history, and the custody history alongside it.

An item that has failed the same way three times has a design or handling problem rather than a repair problem. An item that comes back damaged from the same context repeatedly is telling you something about the context — a particular job, a particular vehicle, a particular way it gets packed.

Neither pattern is visible in an individual ticket. Both are obvious in a history, which is why a repair log that lives in email is worth so much less than one that lives on the asset.

The thresholds worth writing down

Judgement is fine until three people are exercising it differently. A written threshold turns a recurring argument into a rule somebody can apply without you.

Set these once, per category rather than per item, and revisit them annually. The exact percentages matter far less than having them written down.

The failure modes this framework doesn't catch

Two cases where the numbers point one way and the right answer is the other.

The item that keeps stopping work

A cheap item that's repeatedly unavailable can cost far more in delay than its replacement price. The framework values the asset; it doesn't value the work waiting on it. If something has caused three job delays, replace it whatever the book value says.

The last unit of its kind

Sometimes the economic call is to scrap and the operational call is to repair, because it's the only one you've and a replacement is twelve weeks out. Note that reason on the ticket — otherwise the decision looks irrational to whoever reviews the spend later.

Measuring whether your decisions are getting better

• Repeat repair rate — items repaired more than twice in twelve months. Falling means the thresholds are working.
• Claim capture rate — repairs on covered assets that were actually claimed. Should approach 100%; most organisations start well below.
• Estimate versus actual variance — consistently over means your estimates are optimistic, which distorts every decision made on them.
• Average age at replacement — drifting up suggests you are running assets past the point where repairs are sensible.

Making the four facts available at the moment of decision

None of these are hard to capture. Each is a field or a record created in the ordinary course of work: the cover recorded when the asset was acquired, the cost entered when a repair closed, the depreciation method set once, the history accumulating on its own.

What is hard is having all four in the same place when the decision is being made — which is standing next to a broken item, not sitting at a desk with four tabs open.

• Record cover at acquisition, not when you need it. Provider, kind, policy number, claim contact, expiry.
• Close every maintenance ticket with an actual cost, even a rough one. A blank is worth nothing.
• Set a depreciation method on anything above a value threshold you choose, so book value exists without anyone calculating it.
• Keep the repair history on the asset rather than in a mailbox, so patterns are visible rather than remembered.

If your answer to any of the four is “somebody would have to look that up”, the decision is being made without it. Decisions made without them are, on average, expensive in a way that never appears as a line item.