The replace-or-keep question is usually framed as a threshold: once cost per hour passes some line, the machine goes. It is a tidy idea and it produces bad decisions, because cost per hour is a lagging average that hides the two things that actually make an old machine expensive — unpredictability and the cost of the work it prevented. A machine averaging reasonable numbers while failing without warning during critical weeks is costing far more than its ledger shows. The useful approach is reading several signals together and knowing which one is driving the case. You can work through your own fleet's position in a free 30-minute session.
Knowing When a Machine Has Passed Its Point
Six signals that together make the replacement case, and why cost per hour on its own will mislead you in both directions.
Why a Cost Threshold Alone Fails
Cost per hour is an average over a period, and averages conceal variance. Two machines with identical annual maintenance cost are not equivalent if one incurred it through predictable scheduled work and the other through three unplanned failures at the worst possible moments.
It also misses everything outside the maintenance ledger. A machine that is unavailable when it is needed costs whatever the delayed work cost, and that figure appears in a project's numbers rather than the fleet's. The machine looks acceptable on the report that gets reviewed and expensive on the one that does not.
And it cuts the other way. A machine with a high cost per hour because of one major rebuild may now be in better condition than a cheaper-looking unit approaching the same work. Judged on the figure alone, you would replace the wrong one.
Six Signals Worth Reading Together
The direction and steepness over several periods matters more than the current figure. A rising trend on one machine while comparable units hold steady is the clearest single indicator.
How predictable the machine is. Unplanned stoppages per period, and whether they are becoming more frequent, matters more to a project than total maintenance spend does.
An older machine whose parts take weeks to source has a downtime profile that has nothing to do with its mechanical condition. This often decides the case before cost does.
Emission norms, client requirements, and site standards can make a perfectly serviceable machine unusable on the work you want to win. A cost analysis will never surface this.
What the machine is worth now against what it will be worth after another year of the same work. Waiting has a cost that does not appear in maintenance figures.
A machine occupying bay time and technician attention is displacing other work. On a constrained workshop this is a real cost that nothing in the ledger captures.
Build the History the Decision Needs
Cost, downtime, and failure frequency accumulating per machine from day one, so the question is answerable when it is asked.
What to Have Ready Before Proposing Replacement
Cost per productive hour for this machine over several periods, alongside comparable units doing similar work.
Unplanned stoppage count and trend, which is usually more persuasive to a project director than cost is.
Downtime hours and what they delayed, expressed in the terms the project uses.
Major work already done, since a recent rebuild changes the argument considerably.
Realistic current residual value, and what it is expected to be in twelve months.
The alternative honestly costed, including what the replacement will cost to run rather than only to buy.
Frequently Asked Questions
Is there a cost-per-hour figure that means replace?
Not a universal one. It depends on what a replacement costs, what the machine still does well, and what your utilisation looks like. Comparison against similar machines in your own fleet is far more informative than any published threshold. You can work through the comparison on a call with our team.
How much history do we need before this is answerable?
Enough to see a trend rather than a snapshot, which usually means a year or more. This is the argument for starting to capture now even if no decision is pending.
Should a machine with low utilisation be replaced or redeployed?
Often redeployed or disposed of rather than replaced. A poor cost per hour driven by low usage is a utilisation finding, and buying a newer machine to sit idle repeats the problem at a higher price.
Does a recent major repair mean we should keep it?
It changes the condition but not the decision logic. Money already spent is gone either way, and the question remains what the machine will cost from here.
Which signal usually decides it in practice?
Unpredictability, more often than cost. Projects tolerate expensive machines more readily than unreliable ones. Start with a free trial to begin tracking both.
Read the Trend, the Variance, and the Constraints
Track cost direction rather than level, count unplanned stoppages, watch parts lead times and site acceptability, and build the case with history that was captured as the work happened rather than reconstructed when the question arrived.
