When to Retire an Asset: The Calculation No One Runs

Key takeaways

See EAM vs CMMS for the system that tracks this.

• Most plants keep aging assets running long past the point where the math justifies it. The retirement decision gets postponed because the capital request feels disruptive, while the operational cost of keeping the asset accrues quietly in maintenance hours, parts spend, and unplanned-downtime impact.
• The retirement calculation has four inputs: annual operating cost (parts + labour + unplanned downtime cost), declining residual life, replacement-asset capability, and risk of a catastrophic failure. The first three are usually present in the CMMS; the fourth is a judgement call.
• The threshold rule is unromantic: when the annual cost to keep an asset running starts approaching a large fraction of what it would cost to replace it, the long-standing rule of thumb is the "50% rule", with some teams using a stricter trigger, the math has tipped. The decision should follow the math within two cycles, not wait for the failure that forces it.
• The biggest argument for not retiring early, "the asset still works", is the same argument that produced the late-decommissioning failure that just cost three days of production. Working is not the threshold. Total cost of ownership against a like-for-like replacement is.

Why the retirement decision gets postponed

An aging asset rarely fails in one dramatic event. It declines: a small mechanical issue this month, a slightly longer changeover next month, a 5% drop in throughput somewhere. None of these is enough to trigger a capital request on its own. The maintenance team adapts. The operators adapt. A pattern of accommodation builds up around the asset, and the cost of that accommodation becomes invisible.

Meanwhile, the capital request to replace the asset feels like a big deal. It needs a business case, a procurement cycle, an installation plan, training. The operational cost of keeping the asset feels like a series of small decisions; the replacement cost feels like one large decision. The math compares unfairly, and the asset stays.

The retirement calculation is the framework that makes the comparison fair. It is a one-page exercise that any maintenance manager or reliability engineer can run, and it is the single biggest unforced error in most plants' capital planning.

The four inputs

1. Annual operating cost

Three components, all from the CMMS:

• Parts spend, the last 12 months of parts consumed for this asset, including the cost of spares carried specifically because of this asset.
• Labour hours, work-order hours on this asset multiplied by the loaded labour rate. Include both reactive and preventive.
• Unplanned downtime cost, production minutes lost on this asset's line, attributed to this asset, multiplied by the plant's standard cost-per-production-minute.

This third item is the largest and the most often missed. The CMMS knows the labour and parts. The OEE event stream knows the production-minute impact. The article on the work order management system covers how to link the two so that the cost-per-failure number is calculable.

2. Declining residual life

The remaining design life adjusted for actual condition. A bearing rated for 20,000 hours with 18,500 on the clock has less than 8% of its design life left. An asset with a structural component (a chassis, a frame) at end-of-design-life may keep running but has no further life.

This is the input that turns a flat operating cost into a rising one, replacement parts will be needed more often, and at some point the OEM stops supporting the model.

3. Replacement-asset capability

Not just like-for-like cost. The new asset is usually better, faster cycle times, lower energy consumption, integrated sensors that feed the OEE system without a custom integration. The retirement calculation should credit the replacement asset for these improvements in production-minute and energy terms, not just match the old asset's specs.

4. Risk of catastrophic failure

The judgement call. The other three inputs are quantitative. This one is asking: if the asset fails in the worst plausible way, what does that cost?

For an auxiliary fan, the answer is "the cost of an emergency replacement." For a critical line asset with a 10-week OEM lead time, the answer can be a multi-week production outage. The retirement calculation should not run on point estimates without this judgement attached.

The threshold rule

The math:

• Annual operating cost (parts + labour + downtime) approaching ~50% of the replacement-asset cost → the math has tipped (the classic repair-vs-replace "50% rule").
• Plus declining residual life < 25% of design → urgency.
• Plus risk of catastrophic failure non-trivial → the decision becomes time-sensitive.

The 50% figure is a rule of thumb, not a precise number, some organisations retire earlier and some run assets harder.

The principle behind it: once annual running cost is that high relative to replacement cost, the total cost of ownership of a new asset (amortised capital plus its own lower running cost) usually beats continuing to feed the old one, and faster still if the new asset delivers any productivity gain.

Below the threshold, keep the asset. Above it, the next decision cycle should commit to retirement.

What "next decision cycle" means varies by plant. For most mid-market plants it is the annual capital plan. For plants with quarterly capex review, it is the next quarter. What it should not mean is "we'll look at it again in a year." A year above the threshold is a year of avoidable operating cost.

The accounting trick most plants fall into

Plants tend to capitalise the new asset and expense the old one's maintenance. The accounting separates them; the operating reality does not. The retirement calculation must compare like-for-like:

• Old asset: annual operating cost (above) summed over remaining residual life.
• New asset: capex amortised over expected useful life + annual operating cost of the new asset.

Done this way, the new asset often comes out ahead earlier than a simple operating-cost ratio suggests, once amortised capex and the new asset's lower running cost are both in the comparison. The article on manufacturing KPIs covers the cost-per-good-unit metric that makes the like-for-like comparison clean.

What to do with the calculation

Run it once a year on the top 20 most-failure-prone assets. The output is a short ranked list:

• Assets above the threshold with high catastrophic-failure risk → retire this cycle.
• Assets above the threshold with low risk → retire next cycle if still above.
• Assets near the threshold → flag for monitoring and re-run in 12 months.
• Assets well below the threshold → leave alone.

Most plants find one to three assets in the "retire this cycle" bucket in any given year. The decision is not dramatic; the discipline is. Without the calculation, the same one to three assets get postponed for 18-24 months until they force the issue with a major failure. The article on root cause analysis covers how the failure-mode pattern feeds the catastrophic-failure-risk judgement.

How Fabrico fits

The calculation can be run in any plant with a CMMS, an OEE system, and a finance function. Where a unified OEE + CMMS platform helps is that the operating-cost number, including the production-minute impact, is one query rather than a manual reconciliation across systems.

Fabrico is built so the retirement calculation is a quarterly report rather than an annual special project. The preventive maintenance schedule data on the asset also feeds the residual-life input. To see what the calculation would look like on your worst-offending assets, book a demo .

Frequently asked questions

What if our finance team will not approve the capex?

That is a different problem with the same fix. The retirement calculation produces the numbers that turn the conversation from "we want a new asset" to "the current asset is costing roughly half its replacement value annually, with the trend rising." Most finance teams respond to defensible numbers more than to operational complaints. Without the calculation, the conversation stays subjective.

How accurate does the operating-cost number need to be?

Within 15%. The threshold rule has enough margin that point-precise inputs are not needed. What matters more is that the same calculation is applied consistently across assets so the prioritisation is honest.

Should we include energy cost in the operating-cost number?

For energy-intensive assets, yes. Older motors and compressors often consume 10-25% more energy than current equivalents at the same throughput, and the energy delta over remaining residual life can be larger than the parts spend.

What about used or refurbished replacement assets?

Run the same calculation with the refurbished asset as the comparison instead of new. The threshold often shifts (a refurbished asset has shorter useful life and may have less reliability than new), but the framework holds.

What is the single biggest mistake in retirement decisions?

Waiting for the failure that forces the issue. By the time the asset fails in a way that demands replacement, the plant has usually paid 12-24 months of above-threshold operating cost and lost an additional production window to emergency procurement. The calculation done annually avoids both.