Just Serviced, Now a Different Fault: PM-Induced vs Coincidence Decision Tree

Why this matters

A customer calls back days after routine maintenance with a brand-new complaint. The unit you serviced is now doing something it never did before. Two stories fit the same facts: either your maintenance disturbed something and induced the fault, or an unrelated component was already on the edge of failure and happened to let go on its own schedule. Getting this call wrong damages trust either way. Claim coincidence when you caused it and the customer feels lied to. Eat the cost as PM-induced when it was genuinely a coincidence and you train yourself to absorb failures that were never yours. The decision is not about ego. It is about a disciplined chain of evidence that points to one origin or the other.

This logic is trade-agnostic. An HVAC tech who replaced a capacitor, a plumber who flushed a water heater, an electrician who torqued panel lugs, and an appliance tech who cleaned a dryer vent all face the identical question when the next fault appears.

Symptom presentation

The signature of this situation is timing plus newness. The new fault did not exist on your pre-service walkthrough (if you documented one), and it appeared inside a short window after you left. The fault is frequently adjacent to your work area but not always. A blower motor that fails three days after a coil cleaning, a supply line that weeps after a water heater swap, a circuit that trips after a panel torque, a control board that dies after a vent clean-out all share this profile.

Watch for the customer narrative drift. People compress time and over-attribute. "It worked perfectly until you touched it" is a memory, not a measurement. Equally, a tech who is defensive will over-attribute to coincidence. Neither narrative is evidence.

Quick checks

Run these before forming an opinion. They take minutes and they decide most cases.

• Is the new fault physically adjacent to what you touched? Adjacency raises but does not prove disturbance.
• Did your service procedure apply force, heat, vibration, or fluid to the failed area? A torque wrench near a brittle lug, a vacuum near aged wiring, a flush stirring sediment that lodges in a valve.
• Was there a pre-existing condition you noted or photographed? A bulging cap, a corroded fitting, a wear pattern that predicts the failure.
• Does the failed part show age consistent with end-of-life, or does it show fresh mechanical insult (a cracked tab, a nicked wire, a backed-out fastener)?
• Could a single root cause upstream explain both your original finding and the new fault?

Isolation tree

Start at the failed component and walk outward.

Step 1: Establish adjacency. If the new fault is in or directly downstream of your work zone, treat disturbance as the leading hypothesis until ruled out. If it is in a separate subsystem you never accessed, coincidence rises sharply.

Step 2: Inspect for mechanical insult. Disturbance leaves marks. Look for fresh scratches, broken tabs, partially seated connectors, a fastener at the wrong torque, a gasket pinched, a wire with a new nick. Age-related failure looks different: uniform corrosion, dried electrolyte, fatigue cracking, scale buildup. The physical evidence on the part is your strongest single signal.

Step 3: Test the disturbance pathway. Ask whether your specific action could mechanically reach the failure. Did you remove and reseat the connector that is now intermittent? Did your sediment flush mobilize debris that is now clogging a valve? Did your vacuum hose snag the wire that is now open? If there is no plausible physical pathway from your hands to the fault, disturbance loses.

Step 4: Test the coincidence hypothesis against base rates. Aging components fail on their own schedule. A part already at end-of-life will fail soon regardless of who was near it. If the failed part shows clear end-of-life wear and your procedure had no mechanical reach to it, coincidence is the honest call.

Step 5: Resolve ties by documentation. If pre-service photos or notes recorded the part as healthy, disturbance gains weight. If they recorded it as degraded, coincidence gains weight. Absence of documentation is itself a lesson for next time.

Confirming diagnosis

A PM-induced diagnosis is confirmed when you can name the specific action, the physical pathway, and the matching insult evidence on the part. "I reseated this connector, it is now intermittent, and the locking tab is cracked" is a confirmed chain.

A coincidence diagnosis is confirmed when the failed part shows independent end-of-life wear, sits outside any pathway your procedure could reach, and ideally was flagged as aging in your pre-service notes. ISO 14224 reliability data and ISO 13379-1 condition-monitoring logic both frame failure as the end of a wear trajectory, not a random event, which is why end-of-life evidence is decisive.

Next steps

If PM-induced: own it, correct it at no charge, and update your service procedure so the disturbance cannot recur. Document the corrective action.

If coincidence: explain the end-of-life evidence in plain language, show the part, and quote the repair as a separate job. Reference your pre-service note if you have one.

Going forward, adopt a documented pre-service walkthrough on every PM. A two-minute photo set of the components you will work near is the cheapest insurance against this exact ambiguity, and NFPA 70B's condition-based maintenance framework treats that baseline record as a core practice.

References

• ISO 14224, Petroleum, petrochemical and natural gas industries, collection and exchange of reliability and maintenance data for equipment.
• ISO 13379-1, Condition monitoring and diagnostics of machines, data interpretation and diagnostics techniques.
• NFPA 70B, Standard for Electrical Equipment Maintenance, condition-based maintenance and baseline documentation.
• ACCA quality maintenance guidance for documenting equipment condition during service visits.