Remove the Workaround Before Diagnosing, or Work Around It, Decision Tree

Why this matters

Once you've spotted a customer-built workaround (see the related reference article), you face an immediate choice: pull it out and diagnose the raw fault, or leave it in place and work your diagnosis around it. Pull it too fast on the wrong system and you can trip a hazard, strand the customer without a working system mid-visit, or destroy evidence you needed to see. Leave it in place too long and you diagnose a compensated system that never shows you the actual fault, burning the whole visit on a state that isn't the real problem. The right call depends on what kind of workaround it is and what it's protecting against.

Start here: is the workaround protecting against a hazard

Before anything else, identify whether the workaround exists because it defeats or bypasses a safety function, or because it merely masks a performance or comfort symptom.

If the workaround bypasses a safety device or interlock (a jumpered limit switch, a wedged safety door, a disabled shutoff), stop and treat this as the priority finding, not a side note. Do not simply remove the bypass and walk away if the underlying condition it was defeating is still active and unaddressed; you could re-enable a device that immediately trips again, or worse, leave the system in a state where the customer re-installs their own bypass after you leave because the "real" fix wasn't obvious to them. Diagnose what condition the safety was tripping on first, with the bypass still defeating it if that's the only safe way to observe the system running, then remove the bypass as part of restoring correct operation, not as your first move.

If the workaround is purely compensating for a performance, comfort, or convenience symptom (a backed-off setting, a manual reset habit, a sequencing habit) with no safety function involved, proceed to the next branch.

If removing it is quick, reversible, and low-consequence

For most performance-level workarounds, removing them is simply undoing a setting or habit, at no cost and no risk: restoring a thermostat to its intended setpoint, removing a physical prop, running the sequence the customer avoids. If removal is quick and reversible, remove it first and observe the raw fault before doing anything else. This gives you the cleanest possible view of the real symptom and is almost always worth the few minutes it costs, because diagnosing the compensated state risks missing the fault entirely.

Watch what happens once removed. If the original symptom reappears clearly, you have successfully reproduced the complaint and can proceed with normal diagnosis. If nothing changes, either the workaround was not actually addressing this fault, or the underlying condition has resolved on its own, both of which are useful findings worth reporting to the customer.

If removing it would strand the customer or destroy evidence

Some workarounds cannot be undone without consequence: removing a manual bypass that keeps essential equipment (heat in winter, water supply, a working refrigeration unit) running, when you cannot complete the repair in the same visit. In this case, work around it deliberately rather than removing it blind:

• Diagnose everything you can with the workaround still in place: readings, visual inspection, history from the customer.
• Where possible, briefly and safely disable the workaround only long enough to take the one or two readings you actually need, then restore it immediately if the repair isn't ready.
• Explain to the customer, in plain terms, what you're doing and why the workaround needs to stay in place until the real fix is done, so they don't feel like you left the problem exactly as you found it for no reason.

This is the right call whenever full removal would leave the customer without a working system and you cannot complete the permanent fix in the same visit.

If the workaround is intermittent or conditional

Some workarounds only kick in under specific conditions the customer manages by habit, like avoiding running two things at once. These are harder to "remove" because there's no single toggle to flip. Ask the customer to deliberately trigger the condition they normally avoid, under your supervision, so you can observe the fault directly instead of inferring it from their description. This is the fastest way to confirm a shared-capacity or interference fault that a customer has been quietly managing around for months.

If you're not sure which kind of workaround it is

When you genuinely can't tell whether removing something risks a hazard versus a mere inconvenience, default to the cautious path: treat it as though it might be safety-related until you've confirmed otherwise. Ask the customer directly why they started doing it, and look for any sign the workaround followed an incident (a trip, a spark, a smell, a near-miss) rather than a pure comfort complaint. A workaround installed after a scare is far more likely to be masking a real hazard than one installed out of simple annoyance.

Recap

1. Identify whether the workaround defeats a safety function first, before anything else.
2. If it's safety-related, diagnose the tripping condition with the bypass in place if needed, then restore correct operation, not just remove the bypass.
3. If it's a low-consequence performance workaround, remove it and observe the raw fault before diagnosing further.
4. If removal would strand the customer, work around it deliberately, disabling it only briefly for specific readings.
5. For conditional workarounds, have the customer trigger the avoided condition under your supervision.
6. When unsure which kind it is, default to the cautious, safety-first assumption until proven otherwise.

References

• Trade-standard practice for safety-device inspection before defeat or bypass removal
• See related: The Customer Built a Workaround That Hides the Real Fault; Is This Safe to Leave Running