You Cannot Budget for Equipment You Cannot See

Ask an operator when a given condenser will need replacing and the answer usually comes from a spreadsheet: the unit went in around 2013, expected life is about fifteen years, therefore budget for it around 2028.

The arithmetic is reasonable. The input is the problem.

A median is not a schedule

ASHRAE's published median service life for a residential single or split-package air conditioner is 15 years. That figure is a median drawn from a large population — which means, by definition, that half of all units did not reach it.

Planning from a median produces a forecast that is correct on average across a large portfolio and wrong about every individual asset. It cannot tell you which specific units are the early half. And the early half is precisely what wrecks a capital plan, because those are the failures that arrive unbudgeted, in summer, at emergency prices.

Age is a weak predictor and condition is a strong one

What actually determines whether a unit reaches its median is how it has been operated. The documented faults in installed equipment are not just efficiency losses — they are operating conditions outside the design envelope:

• Undercharge reduces the refrigerant flow that cools the compressor motor. The DOE review found undercharge in 29–78 % of systems depending on the study.
• Low airflow across the indoor coil — present "more than 50 % of the time in all regions studied" — depresses evaporator temperature.
• Overcharge raises head pressure; NIST measured 6 % greater total power at 30 % overcharge.

Two condensers installed the same week, in identical houses, can therefore be in entirely different mechanical condition after eight years — and nothing in your records distinguishes them, because both have simply been cooling houses without complaint.

Being honest about what is not known

We are not going to give you a figure for how many years a given fault removes from equipment life. We looked for a credible study quantifying that and did not find one — it would require instrumenting large numbers of systems and following them for a decade or more.

Anyone who quotes you a precise number for this is repeating something unsourced. The defensible statement is narrower: components run outside their design envelope wear faster than components run inside it, and the field data says most systems spend years outside it undetected.

What condition data would change

A capital plan built on installation dates answers "how old is everything." A capital plan informed by continuous condition data can start to answer "which of these is deteriorating" — which is the question the budget actually depends on.

That does not require predicting failure dates, and Vetralis does not claim to predict them. It requires knowing which units are drifting away from their own normal operating pattern, so that the ones heading for early replacement identify themselves before the summer they fail, rather than during it.