Fix Now or Replace Soon on an Aging System: Decision Tree
Why this matters
An old unit breaks and the customer asks the question you cannot dodge: fix it, or is it time to replace? Answer "just fix it" on a system that is three repairs from the grave and you become the tech who kept patching a lost cause. Push replacement on a unit with real life left and you are the upsell artist. The honest answer is a calculation the customer cannot make themselves, because they cannot see what you see. This tree is how you make that call defensibly and explain it straight.
Start here: is there a safety reason it should not run another day?
Before the economics, rule out the showstopper. If the aging unit has a genuine hazard, the decision is not "fix or replace," it is "do not run it as-is."
- A failed safety device, a cracked pressure or combustion boundary, compromised insulation, or a leak that mixes water and electricity: lead with the safety action. Shut it down, red-tag it, and the conversation becomes repair-to-safe or replace, not keep-running.
- No hazard: proceed to the economics below.
Never let an end-of-life economic argument talk a customer into running an unsafe unit one more season.
The repair-to-value test
The core question: does this repair cost a small or a large fraction of replacement, and how much life does it buy?
- Small fraction of replacement, buys years: fix it. This is the easy repair on a unit that otherwise has runway.
- Large fraction of replacement, buys one season: lean replace. Pouring a big share of a new unit into an old one that will need the next big repair soon is throwing good money after bad.
- Large fraction of replacement, buys years, and nothing else is failing: judgment call. It can still be worth it if the customer cannot replace now.
The trap is looking at this one repair in isolation. Read the whole unit.
Read the whole unit, not just the broken part
The failed part is one data point. Score the system around it.
- Age vs typical service life: past the normal lifespan for the type, every part is now a candidate.
- Repair history: this is the third major fix in two years tells you more than any single symptom. A unit nickel-and-diming the customer is a replace even when each repair is small.
- Condition of the rest: corrosion through the structure, worn-out adjacent components, efficiency loss. If the parts you are not fixing today are visibly tired, the next call is coming.
- Parts availability: when parts for the unit are getting scarce or slow, that alone shortens its remaining serviceable life.
A young unit with one failed part is a fix. An old unit with a failed part plus three tired neighbors and a repair history is a replace, even if today's repair is cheap.
When fix-now is right
- The unit is within its service life and this is its first real repair.
- The repair is a small fraction of replacement and buys multiple seasons.
- The customer plans to sell or move, and replacement value will not follow them.
- Budget rules out replacement and the fix is safe and disclosed as buying time.
When replace-soon is right
| Signal | Points to |
|---|---|
| Past typical service life | Replace |
| Repair is a large share of new cost | Replace |
| Third major repair in a short span | Replace |
| Adjacent components also failing | Replace |
| Parts scarce or back-ordered | Replace |
| Efficiency well below current standard | Replace |
When several rows light up, do the safe repair if they want to limp it, but tell them plainly: "I will fix this, and it will run, but this unit is at the end. Budget for replacement, because the next failure may not be worth fixing."
Give the customer the real choice
Frame it as buying time vs buying a fresh start, with honest numbers as ratios: "This repair is about a third of a new unit and might get you a season. A new unit costs more up front but resets the clock and cuts your running cost. On a unit this old, I would not put much more into it." Document the unit's age, the repair history you saw, and which path they chose.
The recap
- Rule out a safety reason it should not run. Safety action first if there is one.
- Weigh the repair as a fraction of replacement against the life it buys.
- Read the whole unit: age, repair history, adjacent condition, parts supply.
- Fix a young unit's first failure; lean replace when the signals stack.
- Give them the buy-time vs fresh-start choice and write down their decision.
References
- Manufacturer documentation on typical equipment service life and end-of-life indicators
- Trade-standard practice for repair-or-replace evaluation
- See related: When the Cheap Fix and the Right Fix Differ (decision tree); Reading Rust and Corrosion Patterns; The Fault That Comes Back Bigger (decision tree)